Search

Usage of language models in an in-context learning environment has been adapted for a wide range of tasks. Recent works have showcased the impact of pretraining data on the in-context performance of language models. In this work, we experiment with numbers having high and low frequencies in the pretraining data to understand the impact of term frequencies on the model's performance. We also experiment with random and adversarial demonstrations to understand the pretraining bias present in the model. Through these experiments, we showcase the importance of pretraining frequencies of the numbers present in the demonstrations and explain how highly frequent terms can be used in the demonstrations to achieve better task performance. Moreover, we also show the impact of pretraining bias on the model's performance and explain how the model overcomes this bias with more demonstrations. Master of Science Recent works focus on understanding and improving the arithmetic capabilities of the state-of-the-art (SOTA) systems in the domain of Natural Language Processing (NLP). This work focuses on designing and performing novel experiments to analyze the impact of training data on the performance of such systems. Through these experiments, this work showcases interesting properties of the SOTA systems which will promote future research to understand them better as well as help in creating better downstream applications.

In today's world, online communication through social network services (SNSs) has become an essential aspect of people's daily lives. As social networking sites (SNSs) have become more sophisticated, cyber attackers have found ways to exploit them for harmful activities such as financial fraud, privacy violations, and sexual or labor exploitation. Thus, it is imperative to gain an understanding of these activities and develop effective countermeasures to build SNSs that can be trusted. The existing approaches have focused on discussing detection mechanisms for a particular type of online social deception (OSD) using various artificial intelligence (AI) techniques, including machine/deep learning (ML/DL) or text mining. However, fewer studies exist on the prevention and response (or mitigation) mechanisms for effective defense against OSD attacks. Further, there have been insufficient efforts to investigate the underlying intents and tactics of those OSD attackers through their in-depth understanding. This dissertation is motivated to take defense approaches to combat OSD attacks through the in-depth understanding of the psychological-social behaviors of attackers and potential victims, which can effectively guide us to take more proactive action against OSD attacks which can minimize potential damages to the potential victims as well as be cost-effective by minimizing or saving recovery cost. In this dissertation, we examine the OSD attacks mainly through two tasks, including understanding their causes and combating them in terms of prevention, detection, and mitigation. In the OSD understanding task, we investigate the intent and tactics of false informers (e.g., fake news spreaders) in propagating fake news or false information. We understand false informers' intent more accurately based on intent-related phrases from fake news contexts to decide on effective and efficient defenses (or interventions) against them. In the OSD combating task, we develop the defense systems following two sub-tasks: (1) The social capital-based friending recommendation system to guide OSN users to choose trustworthy users to defend against phishing attackers proactively; and (2) The defensive opinion update framework for OSN users to process their opinions by filtering out false information. The schemes proposed for combating OSD attacks contribute to the prevention, detection, and mitigation of OSD attacks. Doctor of Philosophy This Ph.D. dissertation explores the issue of online social deception (OSD) in the context of social networking services (SNSs). With the increasing sophistication of SNSs, cyber attackers have found ways to exploit them for harmful activities, such as financial fraud and privacy violations. While previous studies have focused on detection mechanisms using artificial intelligence (AI) techniques, this dissertation takes a defense approach by investigating the underlying psychological-social behaviors of attackers and potential victims. Through two tasks of understanding OSD causes and combating them through various AI approaches, this dissertation proposes a social capital-based friending recommendation system, a defensive opinion update framework, and a fake news spreaders' intent analysis framework to guide SNS users in choosing trustworthy users and filtering out phishing attackers or false information. The proposed schemes contribute to the prevention, detection, and mitigation of OSD attacks, potentially minimizing potential damages to potential victims and saving recovery costs.

In recent years, the large amount of labeled data available has helped tend machine learning (ML) research toward using purely data driven end-to-end pipelines, e.g., in deep neural network research. However, in many situations, data is limited and of poor quality. Traditional ML pipelines are known to be susceptible to various issues when trained on low volumes of non-representative, noisy datasets. We investigate the question of whether prior domain knowledge about the problem being modeled can be employed within the ML pipeline to improve model performance under data paucity and in noisy contexts? This report presents recent developments as well as details, novel contributions in the context of incorporating prior domain knowledge in various data-driven modeling (i.e., machine learning - ML) pipelines particularly geared towards scientific applications. Such domain knowledge exists in various forms and can be incorporated into the machine learning pipeline using different implicit and explicit methods (termed: science-guided machine learning (SGML)). All the novel techniques proposed in this report have been presented in the context of developing SGML to model fluid dynamics applications, but can be easily generalized to other applications. Specifically, we present SGML pipelines to (i) incorporate prior domain knowledge into the ML model architecture (ii) incorporate knowledge about the distribution of the target process as statistical priors for improved prediction performance (iii) leverage prior knowledge to quantify consistency of ML decisions with scientific principles (iv) explicitly incorporate known mathematical relationships of scientific phenomena to influence the ML pipeline (v) develop science-guided transfer learning to improve performance under data paucity. Each technique that is presented, has been designed with a focus on simplicity and minimal cost of implementation with a goal of yielding significant improvements in model performance especially under low data volumes or under noisy data conditions. In each application, we demonstrate through rigorous qualitative and quantitative experiments that our SGML pipelines achieve significant improvements in performance and interpretability over corresponding models that are purely data-driven and agnostic to scientific knowledge. Doctor of Philosophy In this work, we present techniques for incorporating scientific knowledge into machine learning (ML) pipelines. We demonstrate these techniques with ML models trained with low data volumes as well as with non-representative, noisy datasets. In both these cases, we demonstrate through rigorous experimentation that incorporating scientific domain knowledge into the ML pipeline using our proposed science guided machine learning (SGML) techniques, leads to significant performance improvement.

Social media has become an important place for modern people to conveniently share and exchange their ideas and opinions. However, not all content on the social media have positive impact. Hate speech is one kind of harmful content that people use abusive speech attacking or promoting hate towards a specific group or an individual. With online hate speech on the rise these day, people have explored ways to automatically recognize the hate speech, and among the ways people have studied, the Bert-based approach is promising and thus dominates SemEval-2019 Task 6, a hate speech detection competition. In this work, the method of fusion of sentiment features and Bert-based approach is proposed. The classic Bert architecture for hate speech detection is modified to fuse with additional sentiment features, provided by an extractor pre-trained on Sentiment140. The proposed model is compared with top-3 models in SemEval-2019 Task 6 Subtask A and achieves 83.1% F1 score that better than the models in the competition. Also, to see if additional sentiment features benefit the detectoin of hate speech, the features are fused with three kind of deep learning architectures respectively. The results show that the models with sentiment features perform better than those models without sentiment features. Master of Science Social media has become an important place for modern people to conveniently share and exchange their ideas and opinions. However, not all content on the social media have positive impact. Hate speech is one kind of harmful content that people use abusive speech attacking or promoting hate towards a specific group or an individual. With online hate speech on the rise these day, people have explored ways to automatically recognize the hate speech, and among the ways people have studied, Bert is one of promising approach for automatic hate speech recognition. Bert is a kind of deep learning model for natural language processing (NLP) that originated from Transformer developed by Google in 2017. The Bert has applied to many NLP tasks and achieved astonished results such as text classification, semantic similarity between pairs of sentences, question answering with given paragraph, and text summarization. So in this study, Bert will be adopted to learn the meaning of given text and distinguish the hate speech from tons of tweets automatically. In order to let Bert better capture hate speech, the approach in this work modifies Bert to take additional source of sentiment-related features for learning the pattern of hate speech, given that the emotion will be negative when people trying to put out abusive speech. For evaluation of the approach, our model is compared against those in SemEval-2019 Task 6, a famous hate speech detection competition, and the results show that the proposed model achieves 83.1\% F1 score better than the models in the competition. Also, to see if additional sentiment features benefit the detection of hate speech, the features are fused with three different kinds of deep learning architectures respectively, and the results show that the models with sentiment features perform better than those without sentiment features.

Satellite imagery research used to be an expensive research topic for companies and organizations due to the limited data and compute resources. As the computing power and storage capacity grows exponentially, a large amount of aerial and satellite images are generated and analyzed everyday for various applications. Current technological advancement and extensive data collection by numerous Internet of Things (IOT) devices and platforms have amplified labeled natural images. Such data availability catalyzed the development and performance of current state-of-the-art image classification and cross-modal models. Despite the abundance of publicly available remote sensing images, very few remote sensing (RS) images are labeled and even fewer are multi-captioned.These scarcities limit the scope of fine tuned state of the art models to at most 38 classes, based on the PatternNet data, one of the largest publicly available labeled RS data. Recent state-of-the art image-to-image retrieval and detection models in RS have shown great results. Because the text-to-image retrieval of RS images is still emerging, it still faces some challenges in the retrieval of those images.These challenges are based on the inaccurate retrieval of image categories that were not present in the training dataset and the retrieval of images from descriptive input. Motivated by those shortcomings in current cross-modal remote sensing image retrieval, we proposed CLIP-RS, a cross-modal remote sensing image retrieval platform. Our proposed framework CLIP-RS is a framework that combines a fine-tuned implementation of a recent state of the art cross-modal and text-based image retrieval model, Contrastive Language Image Pre-training (CLIP) and FAISS (Facebook AI similarity search), a library for efficient similarity search. Our implementation is deployed on a Web App for inference task on text-to-image and image-to-image retrieval of RS images collected via the Mapbox GL JS API. We used the free tier option of the Mapbox GL JS API and took advantage of its raster tiles option to locate the retrieved results on a local map, a combination of the downloaded raster tiles. Other options offered on our platform are: image similarity search, locating an image in the map, view images' geocoordinates and addresses.In this work we also proposed two remote sensing fine-tuned models and conducted a comparative analysis of our proposed models with a different fine-tuned model as well as the zeroshot CLIP model on remote sensing data. Master of Science Satellite imagery research used to be an expensive research topic for companies and organizations due to the limited data and compute resources. As the computing power and storage capacity grows exponentially, a large amount of aerial and satellite images are generated and analyzed everyday for various applications. Current technological advancement and extensive data collection by numerous Internet of Things (IOT) devices and platforms have amplified labeled natural images. Such data availability catalyzed the devel- opment and performance of current state-of-the-art image classification and cross-modal models. Despite the abundance of publicly available remote sens- ing images, very few remote sensing (RS) images are labeled and even fewer are multi-captioned.These scarcities limit the scope of fine tuned state of the art models to at most 38 classes, based on the PatternNet data,one of the largest publicly avail- able labeled RS data.Recent state-of-the art image-to-image retrieval and detection models in RS have shown great results. Because the text-to-image retrieval of RS images is still emerging, it still faces some challenges in the re- trieval of those images.These challenges are based on the inaccurate retrieval of image categories that were not present in the training dataset and the re- trieval of images from descriptive input. Motivated by those shortcomings in current cross-modal remote sensing image retrieval, we proposed CLIP-RS, a cross-modal remote sensing image retrieval platform.Cross-modal retrieval focuses on data retrieval across different modalities and in the context of this work, we focus on textual and imagery modalities.Our proposed frame- work CLIP-RS is a framework that combines a fine-tuned implementation of a recent state of the art cross-modal and text-based image retrieval model, Contrastive Language Image Pre-training (CLIP) and FAISS (Facebook AI similarity search), a library for efficient similarity search. In deep learning, the concept of fine tuning consists of using weights from a different model or algorithm into a similar model with different domain specific application. Our implementation is deployed on a Web Application for inference tasks on text-to-image and image-to-image retrieval of RS images collected via the Mapbox GL JS API. We used the free tier option of the Mapbox GL JS API and took advantage of its raster tiles option to locate the retrieved results on a local map, a combination of the downloaded raster tiles. Other options offered on our platform are: image similarity search, locating an image in the map, view images' geocoordinates and addresses.In this work we also pro- posed two remote sensing fine-tuned models and conducted a comparative analysis of our proposed models with a different fine-tuned model as well as the zeroshot CLIP model on remote sensing data. Detection models in RS have shown great results. Because the text-to-image retrieval of RS images is still emerging, it still faces some challenges in the re- trieval of those images.These challenges are based on the inaccurate retrieval of image categories that were not present in the training dataset and the re- trieval of images from descriptive input. Motivated by those shortcomings in current cross-modal remote sensing image retrieval, we proposed CLIP-RS, a cross-modal remote sensing image retrieval platform.Cross-modal retrieval focuses on data retrieval across different modalities and in the context of this work, we focus on textual and imagery modalities.Our proposed frame- work CLIP-RS is a framework that combines a fine-tuned implementation of a recent state of the art cross-modal and text-based image retrieval model, Contrastive Language Image Pre-training (CLIP) and FAISS (Facebook AI similarity search), a library for efficient similarity search. In deep learning, the concept of fine tuning consists of using weights from a different model or algorithm into a similar model with different domain specific application. Our implementation is deployed on a Web Application for inference tasks on text-to-image and image-to-image retrieval of RS images collected via the Mapbox GL JS API. We used the free tier option of the Mapbox GL JS API and took advantage of its raster tiles option to locate the retrieved results on a local map, a combination of the downloaded raster tiles. Other options offered on our platform are: image similarity search, locating an image in the map, view images' geocoordinates and addresses.In this work we also pro- posed two remote sensing fine-tuned models and conducted a comparative analysis of our proposed models with a different fine-tuned model as well as the zeroshot CLIP model on remote sensing data.

Anomaly detection is a ubiquitous and challenging task relevant across many disciplines. With the vital role communication networks play in our daily lives, the security of these networks is imperative for smooth functioning of society. This thesis proposes a novel self-supervised deep learning framework CAAD for anomaly detection in wireless communication systems. Specifically, CAAD employs powerful adversarial learning and contrastive learning techniques to learn effective representations of normal and anomalous behavior in wireless networks. Rigorous performance comparisons of CAAD with several state-of-the-art anomaly detection techniques has been conducted and verified that CAAD yields a mean performance improvement of 92.84%. Additionally, CAAD is augmented with the ability to systematically incorporate expert feedback through a novel contrastive learning feedback loop to improve the learned representations and thereby reduce prediction uncertainty (CAAD-EF). CAAD-EF is a novel, holistic and widely applicable solution to anomaly detection. Master of Science Anomaly detection is a technique that can be used to detect if there is any abnormal behavior in data. It is a ubiquitous and a challenging task relevant across many disciplines. With the vital role communication networks play in our daily lives, the security of these networks is imperative for smooth functioning of society. Anomaly detection in such communication networks is essential in ensuring security. This thesis proposes a novel framework CAAD for anomaly detection in wireless communication systems. Rigorous performance comparisons of CAAD with several state-of-the-art anomaly detection techniques has been conducted and verified that CAAD yields a mean performance improvement of 92.84% over state-of-the-art anomaly detection models. Additionally, CAAD is augmented with the ability to incorporate feedback from experts about whether a sample is normal or anomalous through a novel feedback loop (CAAD-EF). CAAD-EF is a novel, holistic and a widely applicable solution to anomaly detection.

Recently it is obvious that climate change has became a critical topic for human society. As climate change becomes more severe, natural disasters caused by climate change have increasingly impacted humans. Most recently, Hurricane Ida killed 43 people across four states. Hurricane Ida's damage could top $95 billion, and many meteorologists predict that climate change is making storms wetter and wider. Thus, there is an urgent need to predict how much damage the flood will cause and prepare for possible destruction. Most current flood damage estimation system did not apply social media data. The theme of this thesis was to evaluate the feasibility of using machine learning models to predict hurricane damage and the input data are social media and satellite imagery. This work involves developing Data Mining approach and a couple of different Machine Learning models that further extract the feature from the data. Satellite imagery is used to identify changes in building structures as well as landscapes, and Twitter data is used to identify damaged locations and the severity of the damage. The features of Twitter posts and satellite imagery were extracted through pre-trained GloVe, ResNet, and VGG models separately. The embedding features were then fed to MLP models for damage level estimation. The models were trained and evaluated on the data. Finally, a case study was performed on the test dataset for hints on improving the models. Master of Science Natural disasters affect Millions of people's lives each year and it is becoming even more severe because of global warming. To make rescue more efficient when the roads and bridges are cut, social media and satellite imagery are effective data sources to help estimating flood damage. With the growth of social media, it is obvious that the post and information from people on the Internet are powerful. Also, with image processing technology improves, the information extracted from satellite images is crucial. In this work we have developed a data mining approach along with different combinations of pre-trained models using neural networks, satellite imagery and archived data from Twitter to estimate flood damage. The data mining approach leverages keywords to identify the event in the history posts in the Twitter, more specifically, we attain the geo-location, time, language information from Twitter, also using pre-event and post-event images which satellite took to generate vectors and thus effectively acquire very useful embedding features. With vectored information from Twitter and satellite imagery, we use pre-trained models and generate damage level prediction. The final results suggest that the proposed approach has potential to create more accurate prediction by using multiple data as input. Furthermore, the estimate result by using only satellite images even outperformed the result using Twitter information, which is an unexpected result comparing to previous studies.

In countries like the US, public school systems function through school districts, which are geographical areas where schools share the same administrative structure and are often coterminous with the boundary of a city or a county. School districts play an important role in the functioning of society. In a well-run school district with safe and well-functioning schools, graduating enough students can enhance the quality of life in its area. Conversely, a poorly run district may cause growth in the area to be far less than surrounding areas, or even a decline in population over time. To promote the efficient functioning of the school district, the boundaries of public schools are redrawn from time to time by the school board/planning officials. In the majority of the cases, this process of redrawing the school boundaries, also called school redistricting or school boundary formation, is done manually by the planners and involves hand-drawn maps. Given the rapid advancements in GIS made in the last decade and the availability of high-quality geospatial data, we opine that an objective treatment of the school redistricting problem by a data-driven model can assist the school board/ decision-makers by providing them with automated plans. These automated plans may serve as possible suggestions to the planners, who can adapt them to prepare their own plans in the way they see fit based on their subjective knowledge and expertise. In this dissertation, we propose algorithmic techniques for solving the problem of (school) redistricting, which is an NP-hard problem. We primarily investigate optimization-based algorithms for solving the problem. Our approaches include (i) clustering, (ii) local search, and (iii) memetic algorithms. We also propose ways of solving the problem using exact methods and fair redistricting techniques based on ethical considerations. The techniques developed here are generic enough to be applied to other redistricting problems with some degree of modification in the objective function and constraint-handling techniques. The source code and corresponding datasets are available at https://github.com/subhodipbiswas/schoolredistricting. Doctor of Philosophy In many countries, public school systems function through school districts, which are geographical areas where schools share the same administrative structure and are often coterminous with the boundary of a city or a county. To promote efficient functioning of the school district, the boundaries of public schools are redrawn from time to time by the school board/planning officials. In the majority of the cases, this process of redrawing the school boundaries, also called school redistricting, is done manually by the planners and involves hand-drawn maps. Given the rapid advancements in GIS made in the last decade and the availability of high-quality geospatial data, we opine that an objective treatment of the school redistricting problem by a data-driven model can assist the school board/ decision-makers by providing them with automated plans. In this presentation, we propose algorithmic techniques for solving the school redistricting problem. Our approaches include (i) clustering, (ii) local search, and (iii) memetic algorithms. We also show that MCMC-based techniques can aid in enabling exact methods to operate on this problem. Lastly, we briefly highlight ethical considerations involved in the process of school redistricting and throw light on some ways to devise more ethically-aware strategies for doing school redistricting. The results indicate that the proposed methods could be a valuable decision-making tool for school officials.

This thesis focuses on specific crypto algorithms, Message Authentication Codes (MACs), running on ultra-low SWaP devices. The type of MACs we used is hash-based message authentication codes (HMAC) and cipher-block-chaining message authentication code (CBC-MAC). The most important thing about ultra-low SWaP devices is their energy usage. This thesis measures different implementations' execution times on ultra-low SWaP devices. We could understand which implementation is suitable for a specific device. In order to understand the crypto algorithm we used, this thesis briefly introduces the concept of hash-based message authentication codes (HMAC) and cipher-block-chaining message authentication code (CBC-MAC) from a high level, including their usage and advantage. The research method is empirical research. This thesis determines the execution times of different implementations. These two algorithms (HMAC and CBC-MAC) contain three implementations. The result comes from those implementations running on the devices we used. Master of Science The deployments of 5G cellular networks are now onboard. The demand increased due to consumers and the availability of more affordable devices. The amount of investment in 5G technology and infrastructure increases market interest in IoT. The 5G network security is essential. How to secure user privacy and their sensitive data while they use 5g network has become a big issue and needs to be solved. However, not all popular crypto algorithms are suited to all devices, especially in those resource-limited microcontrollers. In this thesis, we will deal with Message Authentication Codes that provide the data integrity check. With resource limit devices, energy usage is an important issue. We will identify which implementations have better energy usage depending on the device features. This thesis will use three implementations for each algorithm. The result of our experiment provide a straightforward way that helps people understand which implementation can run more efficiently on specific ultra-low devices.

Illegal logging and timber trade presents a persistent threat to global biodiversity and national security due to its ties with illicit financial flows, and causes revenue loss. The scale of global commerce in timber and associated products, combined with the complexity and geographical spread of the supply chain entities present a non-trivial challenge in detecting such transactions. International shipment records, specifically those containing bill of lading is a key source of data which can be used to detect, investigate and act upon such transactions. The comprehensive problem can be described as building a framework that can perform automated discovery and facilitate actionability on detected transactions. A data driven machine learning based approach is necessitated due to the volume, velocity and complexity of international shipping data. Such an automated framework can immensely benefit our targeted end-users---specifically the enforcement agencies. This overall problem comprises of multiple connected sub-problems with associated research questions. We incorporate crucial domain knowledge---in terms of data as well as modeling---through employing expertise of collaborating domain specialists from ecological conservationist agencies. The collaborators provide formal and informal inputs spanning across the stages---from requirement specification to the design. Following the paradigm of similar problems such as fraud detection explored in prior literature, we formulate the core problem of discovering suspicious transactions as an anomaly detection task. The first sub-problem is to build a system that can be used find suspicious transactions in shipment data pertaining to imports and exports of multiple countries with different country specific schema. We present a novel anomaly detection approach---for multivariate categorical data, following constraints of data characteristics, combined with a data pipeline that incorporates domain knowledge. The focus of the second problem is U.S. specific imports, where data characteristics differ from the prior sub-problem---with heterogeneous attributes present. This problem is important since U.S. is a top consumer and there is scope of actionable enforcement. For this we present a contrastive learning based anomaly detection model for heterogeneous tabular data, with performance and scalability characteristics applicable to real world trade data. While the first two problems address the task of detecting suspicious trades through anomaly detection, a practical challenge with anomaly detection based systems is that of relevancy or scenario specific precision. The third sub-problem addresses this through a human-in-the-loop approach augmented by visual analytics, to re-rank anomalies in terms of relevance---providing explanations for cause of anomalies and soliciting feedback. The last sub-problem pertains to explainability and actionability towards suspicious records, through algorithmic recourse. Algorithmic recourse aims to provides meaningful alternatives towards flagged anomalous records, such that those counterfactual examples are not judged anomalous by the underlying anomaly detection system. This can help enforcement agencies advise verified trading entities in modifying their trading patterns to avoid false detection, thus streamlining the process. We present a novel formulation and metrics for this unexplored problem of algorithmic recourse in anomaly detection. and a deep learning based approach towards explaining anomalies and generating counterfactuals. Thus the overall research contributions presented in this dissertation addresses the requirements of the framework, and has general applicability in similar scenarios beyond the scope of this framework. Doctor of Philosophy Illegal timber trade presents multiple global challenges to ecological biodiversity, vulnerable ecosystems, national security and revenue collection. Enforcement agencies---the target end-users of this framework---face a myriad of challenges in discovering and acting upon shipments with illegal timber that violate national and transnational laws due to volume and complexity of shipment data, coupled with logistical hurdles. This necessitates an automated framework based upon shipment data that can address this task---through solving problems of discovery, analysis and actionability. The overall problem is decomposed into self contained sub-problems that address the associated specific research questions. These comprise of anomaly detection in multiple types of high dimensional tabular data, improving precision of anomaly detection through expert feedback and algorithmic recourse for anomaly detection. We present data mining and machine learning solutions to each of the sub-problems that overcome limitations and inapplicability of prior approaches. Further, we address two broader research questions. First is incorporation domain knowledge into the framework, which we accomplish through collaboration with domain experts from environmental conservation organizations. Secondly, we address the issue of explainability in anomaly detection for tabular data in multiple contexts. Such real world data presents with challenges of complexity and scalability, especially given the tabular format of the data that presents it's own set of challenges in terms of machine learning. The solutions presented to these machine learning problems associated with each of components of the framework provide an end-to-end solution to it's requirements. More importantly, the models and approaches presented in this dissertation have applicability beyond the application scenario with similar data and application specific challenges.

In recent years the usage of public transit services has been rapidly increased, thanks to huge progress on network technologies. However, the disruptions in modern public transit services also increased, due to aging infrastructure, non-comprehensive system design and the needs for maintenance. Any disruptions happened in current transit networks can cause to major disasters on passengers who use these networks for their daily commutes. Although we have lots of usage on transit network, still most current disruptions detection systems either lack of network coverage or did not have real-time system. The goal of this thesis was to create a system that can leverage Twitter data to help in detecting service disruptions in their early stage. This work involves a web applications which contains front-end, back-end and database, along with data mining techniques that obtain Tweets from a live Twitter stream related to the Washington Metropolitan Area Transit Authority (WMATA) metro system. The fundamental features of the system includes real-time incidents panel, historical events review, activities search near specific metro station and recent news review, which allowing people to have more relatively information based on their needs. After the initial functionalities is being settled, we further developed storytelling and sentiment analysis applications, which allowed people have more comprehensive information about the incidents that are happened around metro stations. Also, with the emergency report we developed, the developer can have immediate notification when an urgent event occurred. After fully testified the system's case study on storytelling, sentiment analysis and emergency report, the outcomes are extreme convincing and trustworthy. Master of Science As public transit network become more and more accessible, people around the world rely on these network for their daily commutes. It is clearly that service disruptions among these system will affect passengers severely, especially when there are more and more people using it. This thesis is dedicated to build a web application that will not only allowing people to search latest information, but also assisting on the early detection of the disruptions. In this work we have developed an web application which has easy to use user interface, along with data mining techniques that connected with live data from Twitter to identify these disruptions. Our website is a real-time platform that contains real-time incidents panel, historical events review, activities search near specific metro station and recent news review based on latest tweets and news. By collecting live data from Twitter and various news website, we further developed storytelling and sentiment analysis features. For storytelling, we applied a machine learning model to help us clustering the related tweets/news, after summarize and track the evolution of tweets/news, we converted into stories and displayed it with interactive timelines. For sentiment analysis, we integrated a machine learning model which will scaled the emotional strength of tweets/news, then show the feelings of particular tweets/news. Additionally, we create an emergency report functionality, since it is important for the authority to where and when the incidents happened as soon as possible. The outcome of the system has been well-testify based on the daily case studies, and the results not only meet the ground truth, but also provide with various information.

The use of Machine Learning for malware detection is essential to counter the massive growth in malware types compared with the traditional signature-based detection system. However, machine learning models could also be extremely vulnerable and sensible to transferable adversarial example (AE) attacks. The transfer AE attack does not require extra information from the victim model such as gradient information. Researchers explore mainly 2 lines of transfer-based adversarial example attacks: ensemble models and ensemble samples. \\ Although comprehensive innovations and progress have been achieved in transfer AE attacks, few works have investigated how these techniques perform in malware data. Besides, generating adversarial examples on an android APK file is not as easy and convenient as it is on image data since the generated AE of malware should also remain its functionality and executability after perturbation. Therefore, it is urgent to validate whether previous methodologies could still have their effect on malware considering the differences compared to image data. \\ In this thesis, we first have a thorough literature review for the AE attacks on malware data and general transfer AE attacks. Then we design our algorithm for the transfer AE attack. We formulate the optimization problem based on the intuition that the contribution evenness of features towards the final prediction result is highly correlated to the AE transferability. We then solve the optimization problem by gradient descent and evaluate it through extensive experiments. Implementing and experimenting with the state-of-the-art AE algorithms and transferability enhancement techniques, we analyze and summarize the weaknesses and strengths of each method. Master of Science Machine learning models have been widely applied to malware detection systems in recent years due to the massive growth in malware types. However, these models are vulnerable to adversarial attacks. Malicious attackers can add some small imperceptible perturbations to the original testing samples and mislead the classification results at a very low cost. Research on adversarial attacks would help us gain a better understanding of the attacker's side and inspire defenses against them. Among all adversarial attacks, the transfer-based adversarial example attack is one of the most devastating attacks since it does not require extra information from the targeted victim model such as gradient information or query from the model. Although plenty of researchers has explored the transfer AE attack lately, few works focus on malware (e.g., Android) data. Compared with image data, perturbing malware is more complicated and challenging since the generated adversarial examples of malware need to remain functionality and executability. To validate how transfer AE attack methods perform on malware, we implement the state-of-the-art (SOTA) works in this thesis and experiment with them on real Android data. Besides, we develop a new transfer-based AE attack method based on the contribution of each feature for generating AE. We then do comprehensive evaluations and draw comparisons between SOTA works and our proposed method.

Information mining and knowledge learning from sequential data is a field of growing importance in both industrial and academic fields. Sequential data, which is the natural representation format of the information flow in many applications, usually carries enormous information and is able to help researchers gain insights for various tasks such as airport threat detection, cyber-attack detection, recommender system, point-of-interest (POI) prediction, and citation forecasting. This dissertation focuses on developing the methods for sequential data-driven applications and evolutionary dynamics characterization for various topics such as transit service disruption detection, early event detection on social media, technology opportunity discovery, and traffic incident impact analysis. In particular, four specific applications are studied with four proposed novel methods, including a spatiotemporal feature learning framework for transit service disruption detection, a multi-task learning framework for cybersecurity event detection, citation dynamics modeling via multi-context attentional recurrent neural networks, and traffic incident impact forecasting via hierarchical spatiotemporal graph neural networks. For the first of these methods, the existing transit service disruption detection methods usually suffer from two significant shortcomings: 1) failing to modulate the sparsity of the social media feature domain, i.e., only a few important ``particles'' are indeed related to service disruption among the massive volume of data generated every day and 2) ignoring the real-world geographical connections of transit networks as well as the semantic consistency existing in the problem space. This work makes three contributions: 1) developing a spatiotemporal learning framework for metro disruption detection using open-source data, 2) modeling semantic similarity and spatial connectivity among metro lines in feature space, and 3) developing an optimization algorithm for solving the multi-convex and non-smooth objective function efficiently. For the second of these methods, the conventional studies in cybersecurity detection suffer from the following shortcomings: 1) unable to capture weak signals generated by the cyber-attacks on small organizations or individual accounts, 2) lack of generalization of distinct types of security incidents, and 3) failing to consider the relatedness across different types of cyber-attacks in the feature domain. Three contributions are made in this work: 1) formulating the problem of social media-based cyber-attack detection into the multi-task learning framework, 2) modeling multi-type task relatedness in feature space, and 3) developing an efficient algorithm to solve the non-smooth model with inequality constraints. For the third of these methods, conventional citation forecasting methods are using the traditional temporal point process, which suffers from several drawbacks: 1) unable to predict the technological categories of citing documents and thus are incapable of technological diversity assessment, and 2) require prior domain knowledge and thus are hard to extend to different research areas. Two contributions are made in this work: 1) formulating a novel framework to provide long-term citation predictions in an end-to-end fashion by integrating the process of learning intensity function representations and the process of predicting future citations and 2) designing two novel temporal attention mechanisms to improve the model's ability to modulate complicated temporal dependencies and to allow the model to dynamically combine the observation and prediction sides during the learning process. For the fourth of these methods, the previous work treats the traffic sensor readings as the features and views the incident duration prediction as a feature-driven regression, which typically suffers from three drawbacks: 1) ignoring the existence of the road-sensor hierarchical structure in the real-world traffic network, 2) unable to learn and modulate the hidden temporal patterns in the sensor readings, and 3) lack of consideration of the spatial connectivity between arterial roads and traffic sensors. This work makes three significant contributions: 1) designing a hierarchical graph convolutional network architecture for modeling the road-sensor hierarchy, 2) proposing novel spatiotemporal attention mechanism on the sensor- and road-level features for representation learning, and 3) presenting a graph convolutional network-based method for incident representation learning via spatial connectivity modeling and traffic characteristics modulation. Doctor of Philosophy Information mining and knowledge learning from sequential data is a field of growing importance in both industrial and academic fields. Sequential data, which is the natural representation format of the information flow in many applications, usually carries enormous information and is able to help researchers gain insights for various tasks such as airport threat detection, cyber-attack detection, recommender system, point-of-interest (POI) prediction, and citation forecasting. This dissertation focuses on developing the methods for sequential data-driven applications and evolutionary dynamics characterization for various topics such as transit service disruption detection, early event detection on social media, technology opportunity discovery, and traffic incident impact analysis. In particular, four specific applications are studied with four proposed novel methods, including a spatiotemporal feature learning framework for transit service disruption detection, a multi-task learning framework for cybersecurity event detection, citation dynamics modeling via multi-context attentional recurrent neural networks, and traffic incident impact forecasting via hierarchical spatiotemporal graph neural networks. For the first of these methods, the existing transit service disruption detection methods usually suffer from two significant shortcomings: 1) failing to modulate the sparsity of the social media feature domain, i.e., only a few important ``particles'' are indeed related to service disruption among the massive volume of data generated every day and 2) ignoring the real-world geographical connections of transit networks as well as the semantic consistency existing in the problem space. This work makes three contributions: 1) developing a spatiotemporal learning framework for metro disruption detection using open-source data, 2) modeling semantic similarity and spatial connectivity among metro lines in feature space, and 3) developing an optimization algorithm for solving the multi-convex and non-smooth objective function efficiently. For the second of these methods, the conventional studies in cybersecurity detection suffer from the following shortcomings: 1) unable to capture weak signals generated by the cyber-attacks on small organizations or individual accounts, 2) lack of generalization of distinct types of security incidents, and 3) failing to consider the relatedness across different types of cyber-attacks in the feature domain. Three contributions are made in this work: 1) formulating the problem of social media-based cyber-attack detection into the multi-task learning framework, 2) modeling multi-type task relatedness in feature space, and 3) developing an efficient algorithm to solve the non-smooth model with inequality constraints. For the third of these methods, conventional citation forecasting methods are using the traditional temporal point process, which suffers from several drawbacks: 1) unable to predict the technological categories of citing documents and thus are incapable of technological diversity assessment, and 2) require prior domain knowledge and thus are hard to extend to different research areas. Two contributions are made in this work: 1) formulating a novel framework to provide long-term citation predictions in an end-to-end fashion by integrating the process of learning intensity function representations and the process of predicting future citations and 2) designing two novel temporal attention mechanisms to improve the model's ability to modulate complicated temporal dependencies and to allow the model to dynamically combine the observation and prediction sides during the learning process. For the fourth of these methods, the previous work treats the traffic sensor readings as the features and views the incident duration prediction as a feature-driven regression, which typically suffers from three drawbacks: 1) ignoring the existence of the road-sensor hierarchical structure in the real-world traffic network, 2) unable to learn and modulate the hidden temporal patterns in the sensor readings, and 3) lack of consideration of the spatial connectivity between arterial roads and traffic sensors. This work makes three significant contributions: 1) designing a hierarchical graph convolutional network architecture for modeling the road-sensor hierarchy, 2) proposing novel spatiotemporal attention mechanism on the sensor- and road-level features for representation learning, and 3) presenting a graph convolutional network-based method for incident representation learning via spatial connectivity modeling and traffic characteristics modulation.

Cybergrooming refers to the crime of establishing personal close relationships with potential victims, commonly teens, for the purpose of sexual exploitation or abuse via online social media platforms. Cybergrooming has been recognized as a serious social problem. However, there have been insufficient programs to provide proactive prevention to protect the youth users from cybergrooming. In this thesis, we present a generative chatbot framework, called SERI (Stop cybERgroomIng), that can generate simulated conversations between a perpetrator chatbot and a potential victim chatbot. To realize the simulation of authentic conversations in the context of cybergrooming, we take deep reinforcement learning (DRL)-based dialogue generation to simulate the authentic conversations between a perpetrator and a potential victim. The design and development of the SERI are motivated to provide a safe and authentic chatting environment to enhance the youth's precautionary awareness and sensitivity of cybergrooming while any unnecessary ethical issues (e.g., the potential misuse of the SERI) are removed or minimized. We developed the SERI as a preliminary platform that the perpetrator chatbot can be deployed in social media environments to interact with human users (i.e., youth) and observe the conversations that the youth users respond to strangers or acquaintances when they are asked for private or sensitive information by the perpetrator. We evaluated the quality of conversations generated by the SERI based on open-source, referenced, and unreferenced metrics as well as human evaluation. The evaluation results show that the SERI can generate authentic conversations between two chatbots compared to the original conversations from the used datasets in perplexity and MaUde scores. Master of Science Cybergrooming refers to the crime of building personal close relationships with potential victims, especially youth users such as children and teenagers, for the purpose of sexual exploitation or abuse via online social media platforms. Cybergrooming has been recognized as a serious social problem. However, there have been insufficient methods to provide proactive protection for the youth users from cybergrooming. In this thesis, we present a generative chatbot framework, called SERI (Stop cybERgroomIng), that can generate simulated authentic conversations between a perpetrator chatbot and a potential victim chatbot by applying advanced natural language generation models. The design and development of the SERI are motivated to ensure a safe and authentic environment to strengthen the youth's precautionary awareness and sensitivity of cybergrooming while any unnecessary ethical issues (e.g., the potential misuse of the SERI) are removed or minimized. We used different metrics and methods to evaluate the quality of conversations generated by the SERI. The evaluation results show that the SERI can generate authentic conversations between two chatbots compared to the original conversations from the used datasets.

During the unprecedented global pandemic of Covid-19, the general public is suffering from inaccurate Covid-19 related information including outdated information and fake news. The most used media: TV, social media, newspaper, and radio are incompetent in providing certitude and flash updates that people are seeking. In order to cope with this challenge, several public data resources that are dedicated to providing Covid-19 information were born. They rallied with experts from different fields to provide authoritative and up-to-date pandemic updates. However, the general public cannot still make complete use of such resources since the learning curve is too steep, especially for the aged and under-aged users. To address this problem, in this Thesis, we propose a question answering system that can be interacted with using simple natural language-based sentences. While building this system, we investigate qualified public data resources and from the data content they are providing, and we collect a set of frequently asked questions for Covid-19 tracking. We further build a dedicated dataset named CovidQA for evaluating the performance of the question answering system with different models. Based on the new dataset, we assess multiple machine learning-based models that are built for retrieving relevant information from databases, and then propose two empirical models which utilize the pre-defined templates to generate SQL queries. In our experiments, we demonstrate both quantitative and qualitative results and provide a comprehensive comparison between different types of methods. The results show that the proposed template-based methods are simple but effective in building question answering systems for specific domain problems. Master of Science During the unprecedented global pandemic of Covid-19, the general public is suffering from inaccurate Covid-19 related information including outdated information and fake news. The most used media: TV, social media, newspaper, and radio are incompetent in providing certitude and flash updates that people are seeking. In order to cope with this challenge, several public data resources that are dedicated to providing Covid-19 information were born. They rallied with experts from different fields to provide authoritative and up-to-date pandemic updates. However, there is room for improvement in terms of user experience. To address this problem, in this Thesis, we propose a system that can be interacted with using natural questions. While building this system, we evaluate and choose six qualified public data providers as the data sources. We further build a testing dataset for evaluating the performance of the system. We assess two Artificial Intelligence-powered models for the system, and then propose two rule-based models for the researched problem. In our experiments, we provide a comprehensive comparison between different types of methods. The results show that the proposed rule-based methods are simple but effective in building such systems.

Electronic Health Records (EHR) data contain comprehensive longitudinal patient information, which is usually stored in databases in the form of either multi-relational structured tables or unstructured texts, e.g., clinical notes. EHR provides a useful resource to assist doctors' decision making, however, they also present many unique challenges that limit the efficient use of the valuable information, such as large data volume, heterogeneous and dynamic information, medical term abbreviations, and noisy nature caused by misspelled words. This dissertation focuses on the development and evaluation of advanced machine learning algorithms to solve the following research questions: (1) How to seek answers from EHR for clinical activity related questions posed in human language without the assistance of database and natural language processing (NLP) domain experts, (2) How to discover underlying relationships of different events and entities in structured tabular EHRs, and (3) How to predict when a medical event will occur and estimate its probability based on previous medical information of patients. First, to automatically retrieve answers for natural language questions from the structured tables in EHR, we study the question-to-SQL generation task by generating the corresponding SQL query of the input question. We propose a translation-edit model driven by a language generation module and an editing module for the SQL query generation task. This model helps automatically translate clinical activity related questions to SQL queries, so that the doctors only need to provide their questions in natural language to get the answers they need. We also create a large-scale dataset for question answering on tabular EHR to simulate a more realistic setting. Our performance evaluation shows that the proposed model is effective in handling the unique challenges about clinical terminologies, such as abbreviations and misspelled words. Second, to automatically identify answers for natural language questions from unstructured clinical notes in EHR, we propose to achieve this goal by querying a knowledge base constructed based on fine-grained document-level expert annotations of clinical records for various NLP tasks. We first create a dataset for clinical knowledge base question answering with two sets: clinical knowledge base and question-answer pairs. An attention-based aspect-level reasoning model is developed and evaluated on the new dataset. Our experimental analysis shows that it is effective in identifying answers and also allows us to analyze the impact of different answer aspects in predicting correct answers. Third, we focus on discovering underlying relationships of different entities (e.g., patient, disease, medication, and treatment) in tabular EHR, which can be formulated as a link prediction problem in graph domain. We develop a self-supervised learning framework for better representation learning of entities across a large corpus and also consider local contextual information for the down-stream link prediction task. We demonstrate the effectiveness, interpretability, and scalability of the proposed model on the healthcare network built from tabular EHR. It is also successfully applied to solve link prediction problems in a variety of domains, such as e-commerce, social networks, and academic networks. Finally, to dynamically predict the occurrence of multiple correlated medical events, we formulate the problem as a temporal (multiple time-points) and multi-task learning problem using tensor representation. We propose an algorithm to jointly and dynamically predict several survival problems at each time point and optimize it with the Alternating Direction Methods of Multipliers (ADMM) algorithm. The model allows us to consider both the dependencies between different tasks and the correlations of each task at different time points. We evaluate the proposed model on two real-world applications and demonstrate its effectiveness and interpretability. Doctor of Philosophy Healthcare is an important part of our lives. Due to the recent advances of data collection and storing techniques, a large amount of medical information is generated and stored in Electronic Health Records (EHR). By comprehensively documenting the longitudinal medical history information about a large patient cohort, this EHR data forms a fundamental resource in assisting doctors' decision making including optimization of treatments for patients and selection of patients for clinical trials. However, EHR data also presents a number of unique challenges, such as (i) large-scale and dynamic data, (ii) heterogeneity of medical information, and (iii) medical term abbreviation. It is difficult for doctors to effectively utilize such complex data collected in a typical clinical practice. Therefore, it is imperative to develop advanced methods that are helpful for efficient use of EHR and further benefit doctors in their clinical decision making. This dissertation focuses on automatically retrieving useful medical information, analyzing complex relationships of medical entities, and detecting future medical outcomes from EHR data. In order to retrieve information from EHR efficiently, we develop deep learning based algorithms that can automatically answer various clinical questions on structured and unstructured EHR data. These algorithms can help us understand more about the challenges in retrieving information from different data types in EHR. We also build a clinical knowledge graph based on EHR and link the distributed medical information and further perform the link prediction task, which allows us to analyze the complex underlying relationships of various medical entities. In addition, we propose a temporal multi-task survival analysis method to dynamically predict multiple medical events at the same time and identify the most important factors leading to the future medical events. By handling these unique challenges in EHR and developing suitable approaches, we hope to improve the efficiency of information retrieval and predictive modeling in healthcare.

Event extraction refers to extracting specific knowledge of incidents from natural language text and consolidating it into a structured form. Some important applications of event extraction include search, retrieval, question answering and event forecasting. However, before events can be extracted it is imperative to detect events i.e. identify which documents from a large collection contain events of interest and from those extracting the sentences that might contain the event related information. This task is challenging because it is easier to obtain labels at the document level than finegrained annotations at the sentence level. Current approaches for this task are suboptimal because they directly aggregate sentence probabilities estimated by a classifier to obtain document probabilities resulting in error propagation. To alleviate this problem we propose to leverage recent advances in representation learning by using attention mechanisms. Specifically, for event detection we propose a method to compute document embeddings from sentence embeddings by leveraging attention and training a document classifier on those embeddings to mitigate the error propagation problem. However, we find that existing attention mechanisms are inept for this task, because either they are suboptimal or they use a large number of parameters. To address this problem we propose a lean attention mechanism which is effective for event detection. Current approaches for event extraction rely on finegrained labels in specific domains. Extending extraction to new domains is challenging because of difficulty of collecting finegrained data. Machine reading comprehension(MRC) based approaches, that enable zero-shot extraction struggle with syntactically complex sentences and long-range dependencies. To mitigate this problem, we propose a syntactic sentence simplification approach that is guided by MRC model to improve its performance on event extraction. Doctor of Philosophy Event extraction is the task of extracting events of societal importance from natural language texts. The task has a wide range of applications from search, retrieval, question answering to forecasting population level events like civil unrest, disease occurrences with reasonable accuracy. Before events can be extracted it is imperative to identify the documents that are likely to contain the events of interest and extract the sentences that mention those events. This is termed as event detection. Current approaches for event detection are suboptimal. They assume that events are neatly partitioned into sentences and obtain document level event probabilities directly from predicted sentence level probabilities. In this dissertation, under the same assumption by leveraging representation learning we mitigate some of the shortcomings of the previous event detection methods. Current approaches to event extraction are only limited to restricted domains and require finegrained labeled corpora for their training. One way to extend event extraction to new domains in by enabling zero-shot extraction. Machine reading comprehension (MRC) based approach provides a promising way forward for zero-shot extraction. However, this approach suffers from the long-range dependency problem and faces difficulty in handling syntactically complex sentences with multiple clauses. To mitigate this problem we propose a syntactic sentence simplification algorithm that is guided by the MRC system to improves its performance.

The study of information extraction and knowledge exploration in the urban environment is gaining popularity. Ubiquitous sensors and a plethora of statistical reports provide an immense amount of heterogeneous urban data, such as traffic data, crime activity statistics, social media messages, and street imagery. The development of methods for heterogeneous urban data-based event identification and impacts analysis for a variety of event topics and assumptions is the subject of this dissertation. A graph convolutional neural network for crime prediction, a multitask learning system for traffic incident prediction with spatiotemporal feature learning, social media-based transportation event detection, and a graph convolutional network-based cyberbullying detection algorithm are the four methods proposed. Additionally, based on the sensitivity of these urban sensor data, a comprehensive discussion on ethical issues of urban computing is presented. This work makes the following contributions in urban perception predictions: 1) Create a preference learning system for inferring crime rankings from street view images using a bidirectional convolutional neural network (bCNN). 2) Propose a graph convolutional networkbased solution to the current urban crime perception problem; 3) Develop street view image retrieval algorithms to demonstrate real city perception. This work also makes the following contributions in traffic incident effect analysis: 1) developing a novel machine learning system for predicting traffic incident duration using temporal features; 2) modeling traffic speed similarity among road segments using spatial connectivity in feature space; and 3) proposing a sparse feature learning method for identifying groups of temporal features at a higher level. In transportation-related incidents detection, this work makes the following contributions: 1) creating a real-time social media-based traffic incident detection platform; 2) proposing a query expansion algorithm for traffic-related tweets; and 3) developing a text summarization tool for redundant traffic-related tweets. Cyberbullying detection from social media platforms is one of the major focus of this work: 1) Developing an online Dynamic Query Expansion process using concatenated keyword search. 2) Formulating a graph structure of tweet embeddings and implementing a Graph Convolutional Network for fine-grained cyberbullying classification. 3) Curating a balanced multiclass cyberbullying dataset from DQE, and making it publicly available. Additionally, this work seeks to identify ethical vulnerabilities from three primary research directions of urban computing: urban safety analysis, urban transportation analysis, and social media analysis for urban events. Visions for future improvements in the perspective of ethics are addressed. Doctor of Philosophy The ubiquitously deployed urban sensors such as traffic speed meters, street-view cameras, and even smartphones in everybody's pockets are generating terabytes of data every hour. How do we refine the valuable intelligence out of such explosions of urban data and information became one of the profitable questions in the field of data mining and urban computing. In this dissertation, four innovative applications are proposed to solve real-world problems with big data of the urban sensors. In addition, the foreseeable ethical vulnerabilities in the research fields of urban computing and event predictions are addressed. The first work explores the connection between urban perception and crime inferences. StreetNet is proposed to learn crime rankings from street view images. This work presents the design of a street view images retrieval algorithm to improve the representation of urban perception. A data-driven, spatiotemporal algorithm is proposed to find unbiased label mappings between the street view images and the crime ranking records. The second work proposes a traffic incident duration prediction model that simultaneously predicts the impact of the traffic incidents and identifies the critical groups of temporal features via a multi-task learning framework. Such functionality provided by this model is helpful for the transportation operators and first responders to judge the influences of traffic incidents. In the third work, a social media-based traffic status monitoring system is established. The system is initiated by a transportation-related keyword generation process. A state-of-the-art tweets summarization algorithm is designed to eliminate the redundant tweets information. In addition, we show that the proposed tweets query expansion algorithm outperforms the previous methods. The fourth work aims to investigate the viability of an automatic multiclass cyberbullying detection model that is able to classify whether a cyberbully is targeting a victim's age, ethnicity, gender, religion, or other quality. This work represents a step forward for establishing an active anti-cyberbullying presence in social media and a step forward towards a future without cyberbullying. Finally, a discussion of the ethical issues in the urban computing community is addressed. This work seeks to identify ethical vulnerabilities from three primary research directions of urban computing: urban safety analysis, urban transportation analysis, and social media analysis for urban events. Visions for future improvements in the perspective of ethics are pointed out.

Researchers have made significant progress in information extraction from short documents in the last few years, including social media interaction, news articles, and email excerpts. This research aims to extract technical entities like hardware resources, computing platforms, compute time, programming language, and libraries from scholarly research articles. Research articles are generally long documents having both salient as well as non-salient entities. Analyzing the cross-sectional relation, filtering the relevant information, measuring the saliency of mentioned entities, and extracting novel entities are some of the technical challenges involved in this research. This work presents a detailed study about the performance, effectiveness, and scalability of rule-based weakly supervised algorithms. We also develop an automated end-to-end Research Entity and Relationship Extractor (E2R Extractor). Additionally, we perform a comprehensive study about the effectiveness of existing deep learning-based information extraction tools like Dygie, Dygie++, SciREX. The research also contributes a dataset containing novel entities annotated in BILUO format and represents the baseline results using the E2R extractor on the proposed dataset. The results indicate that the E2R extractor successfully extracts salient entities from research articles. Master of Science Information extraction is a process of automatically extracting meaningful information from unstructured text such as articles, news feeds and presenting it in a structured format. Researchers have made significant progress in this domain over the past few years. However, their work primarily focuses on short documents such as social media interactions, news articles, email excerpts, and not on long documents such as scholarly articles and research papers. Long documents contain a lot of redundant data, so filtering and extracting meaningful information is quite challenging. This work focuses on extracting entities such as hardware resources, compute platforms, and programming languages used in scholarly articles. We present a deep learning-based model to extract such entities from research articles and research papers. We evaluate the performance of our deep learning model against simple rule-based algorithms and other state-of-the-art models for extracting the desired entities. Our work also contributes a labeled dataset containing the entities mentioned above and results obtained on this dataset using our deep learning model.

In this work, we propose a PRivAcy-Diversity-Aware Team Formation framework, namely PRADA-TF, that can be deployed based on the trust relationships between users in online social networks (OSNs). Our proposed PRADA-TF is mainly designed to reflect team members' domain expertise and privacy preserving preferences when a task requires a wide range of diverse domain expertise for its successful completion. The proposed PRADA-TF aims to form a team for maximizing its productivity based on members' characteristics in their diversity, privacy preserving, and information sharing. We leveraged a game theory called Mechanism Design in order for a mechanism designer as a team leader to select team members that can maximize the team's social welfare, which is the sum of all team members' utilities considering team productivity, members' privacy preserving, and potential privacy loss caused by information sharing. To screen a set of candidate teams in the OSN, we built an expert social network based on real co-authorship datasets (i.e., Netscience) with 1,590 scientists, used the semi-synthetic datasets to construct a trust network based on a belief model called Subjective Logic, and identified trustworthy users as candidate team members. Via our extensive simulation experiments, we compared the seven different TF schemes, including our proposed and existing TF algorithms, and analyzed the key factors that can significantly impact the expected and actual social welfare, expected and actual potential privacy leakout, and team diversity of a selected team. Master of Science In this work, we propose a PRivAcy-Diversity-Aware Team Formation framework, namely PRADA-TF, that can be deployed based on the trust relationships between users in online social networks (OSNs). Our proposed PRADA-TF is mainly designed to reflect team members' domain expertise and privacy preserving preferences when a task requires a wide range of diverse domain expertise for its successful completion. The proposed PRADA-TF aims to form a team based on members' characteristics in their diversity, privacy preserving, and information sharing so as to maximize the performance of the team. We leveraged a game theory called Mechanism Design in order for a mechanism designer as a team leader to select team members that can maximize the team's social welfare, which is the sum of all team members' utilities considering team productivity, members' privacy preserving, and potential privacy loss caused by information sharing. To screen a set of candidate teams in the OSN, we built an expert social network based on real co-authorship datasets with 1,590 scientists, used the semi-synthetic datasets to construct a trust network representing the trust relationship between the users in OSNs, and identified trustworthy users as candidate team members. Via our extensive simulation experiments, we compared the seven different team formation (TF) schemes, including our proposed and existing TF algorithms, and analyzed the key factors that can significantly impact the expected and actual social welfare, expected and actual potential privacy leakout, and team diversity of a selected team.

In Virginia Tech's Digital Library Research Laboratory, we support subject-matter-experts (SMEs) in their pursuit of research goals. Their goals include everything from data collection to analysis to reporting. Their research commonly involves an analysis of an extensive collection of data such as tweets or web pages. Without support -- such as by our lab, developers, or data analysts/scientists -- they would undertake the data analysis themselves, using available analytical tools, frameworks, and languages. Then, to extract and produce the information needed to achieve their goals, the researchers/users would need to know what sequences of functions or algorithms to run using such tools, after considering all of their extensive functionality. Our research addresses these problems directly by designing a system that lowers the information barriers. Our approach is broken down into three parts. In the first two parts, we introduce a system that supports discovery of both information and supporting services. In the first part, we describe the methodology that incorporates User eXperience (UX) research into the process of workflow design. Through the methodology, we capture (a) what are the different user roles and goals, (b) how we break down the user goals into tasks and sub-tasks, and (c) what functions and services are required to solve each (sub-)task. In the second part, we identify and describe key components of the infrastructure implementation. This implementation captures the various goals/tasks/services associations in a manner that supports information inquiry of two types: (1) Given an information goal as query, what is the workflow to derive this information? and (2) Given a data resource, what information can we derive using this data resource as input? We demonstrate both parts of the approach, describing how we teach and apply the methodology, with three case studies. In the third part of this research, we rely on formalisms used in describing digital libraries to explain the components that make up the information system. The formal description serves as a guide to support the development of information systems that generate workflows to support SME information needs. We also specifically describe an information system meant to support information goals that relate to Twitter data. Doctor of Philosophy In Virginia Tech's Digital Library Research Laboratory, we support subject-matter-experts (SMEs) in their pursuit of research goals. This includes everything from data collection to analysis to reporting. Their research commonly involves an analysis of an extensive collection of data such as tweets or web pages. Without support -- such as by our lab, developers, or data analysts/scientists -- they would undertake the data analysis themselves, using available analytical tools, frameworks, and languages. Then, to extract and produce the information needed to achieve their goals, the researchers/users would need to know what sequences of functions or algorithms to run using such tools, after considering all of their extensive functionality. Further, as more algorithms are being discovered and datasets are getting larger, the information processing effort is getting more and more complicated. Our research aims to address these problems directly by attempting to lower the barriers, through a methodology that integrates the full life cycle, including the activities carried out by User eXperience (UX), analysis, development, and implementation experts. We devise a three part approach to this research. The first two parts concern building a system that supports discovery of both information and supporting services. First, we describe the methodology that introduces UX research into the process of workflow design. Second, we identify and describe key components of the infrastructure implementation. We demonstrate both parts of the approach, describing how we teach and apply the methodology, with three case studies. In the third part of this research, we extend formalisms used in describing digital libraries to encompass the components that make up our new type of extensible information system.

In recent years we have seen an increase in the number of public transit service disruptions due to aging infrastructure, system failures and the regular need for maintenance. With the fleeting growth in the usage of these transit networks there has been an increase in the need for the timely detection of such disruptions. Any types of disruptions in these transit networks can lead to delays which can have major implications on the daily passengers. Most current disruption detection systems either do not operate in real-time or lack transit network coverage. The theme of this thesis was to leverage Twitter data to help in earlier detection of service disruptions. This work involves developing a pure Data Mining approach and a couple different approaches that use Graph Neural Networks to identify transit disruption related information in Tweets from a live Twitter stream related to the Washington Metropolitan Area Transit Authority (WMATA) metro system. After developing three different models, a Dynamic Query Expansion model, a Tweet-GCN and a Tweet-Level GCN to represent the data corpus we performed various experiments and benchmark evaluations against other existing baseline models, to justify the efficacy of our approaches. After seeing astounding results across both the Tweet-GCN and Tweet-Level GCN, with an average accuracy of approximately 87.3% and 89.9% we can conclude that not only are these two graph neural models superior for basic NLP text classification, but they also outperform other models in identifying transit disruptions. Master of Science Millions of people worldwide rely on public transit networks for their daily commutes and day to day movements. With the growth in the number of people using the service, there has been an increase in the number of daily passengers affected by service disruptions. This thesis and research involves proposing and developing three different approaches to help aid in the timely detection of these disruptions. In this work we have developed a pure data mining approach along with two deep learning models using neural networks and live data from Twitter to identify these disruptions. The data mining approach uses a set of dirsuption related input keywords to identify similar keywords within the live Twitter data. By collecting historical data we were able to create deep learning models that represent the vocabulary from the disruptions related Tweets in the form of a graph. A graph is a collection of data values where the data points are connected to one another based on their relationships. A longer chain of connection between two words defines a weak relationship, a shorter chain defines a stronger relationship. In our graph, words with similar contextual meanings are connected to each other over shorter distances, compared to words with different meanings. At the end we use a neural network as a classifier to scan this graph to learn the semantic relationships within our data. Afterwards, this learned information can be used to accurately classify the disruption related Tweets within a pool of random Tweets. Once all the proposed approaches have been developed, a benchmark evaluation is performed against other existing text classification techniques, to justify the effectiveness of the approaches. The final results indicate that the proposed graph based models achieved a higher accuracy, compared to the data mining model, and also outperformed all the other baseline models. Our Tweet-Level GCN had the highest accuracy of 89.9%.

With the huge volume of biological data generated by new technologies and the booming of new machine learning based analytical tools, we expect to advance life science and human health at an unprecedented pace. Unfortunately, there is a significant gap between the complex raw biological data from real life and the data required by mathematical and statistical tools. This gap is contributed by two fundamental and universal problems in biological data that are both related to confounding effects. The first is the intrinsic complexities of the data. An observed sample could be the mixture of multiple underlying sources and we may be only interested in one or part of the sources. The second type of complexities come from the acquisition process of the data. Different samples may be gathered at different time and/or from different locations. Therefore, each sample is associated with specific distortion that must be carefully addressed. These confounding effects obscure the signals of interest in the acquired data. Specifically, this dissertation will address the two major challenges in confounding effects removal: alignment and deconvolution. Liquid chromatography–mass spectrometry (LC-MS) is a standard method for proteomics and metabolomics analysis of biological samples. Unfortunately, it suffers from various changes in the retention time (RT) of the same compound in different samples, and these must be subsequently corrected (aligned) during data processing. Classic alignment methods such as in the popular XCMS package often assume a single time-warping function for each sample. Thus, the potentially varying RT drift for compounds with different masses in a sample is neglected in these methods. Moreover, the systematic change in RT drift across run order is often not considered by alignment algorithms. Therefore, these methods cannot effectively correct all misalignments. To utilize this information, we develop an integrated reference-free profile alignment method, neighbor-wise compound-specific Graphical Time Warping (ncGTW), that can detect misaligned features and align profiles by leveraging expected RT drift structures and compound-specific warping functions. Specifically, ncGTW uses individualized warping functions for different compounds and assigns constraint edges on warping functions of neighboring samples. We applied ncGTW to two large-scale metabolomics LC-MS datasets, which identifies many misaligned features and successfully realigns them. These features would otherwise be discarded or uncorrected using existing methods. When the desired signal is buried in a mixture, deconvolution is needed to recover the pure sources. Many biological questions can be better addressed when the data is in the form of individual sources, instead of mixtures. Though there are some promising supervised deconvolution methods, when there is no a priori information, unsupervised deconvolution is still needed. Among current unsupervised methods, Convex Analysis of Mixtures (CAM) is the most theoretically solid and strongest performing one. However, there are some major limitations of this method. Most importantly, the overall time complexity can be very high, especially when analyzing a large dataset or a dataset with many sources. Also, since there are some stochastic and heuristic steps, the deconvolution result is not accurate enough. To address these problems, we redesigned the modules of CAM. In the feature clustering step, we propose a clustering method, radius-fixed clustering, which could not only control the space size of the cluster, but also find out the outliers simultaneously. Therefore, the disadvantages of K-means clustering, such as instability and the need of cluster number are avoided. Moreover, when identifying the convex hull, we replace Quickhull with linear programming, which decreases the computation time significantly. To avoid the not only heuristic but also approximated step in optimal simplex identification, we propose a greedy search strategy instead. The experimental results demonstrate the vast improvement of computation time. The accuracy of the deconvolution is also shown to be higher than the original CAM. Doctor of Philosophy Due to the complexity of biological data, there are two major pre-processing steps: alignment and deconvolution. The alignment step corrects the time and location related data acquisition distortion by aligning the detected signals to a reference signal. Though many alignment methods are proposed for biological data, most of them fail to consider the relationships among samples carefully. This piece of structure information can help alignment when the data is noisy and/or irregular. To utilize this information, we develop a new method, Neighbor-wise Compound-specific Graphical Time Warping (ncGTW), inspired by graph theory. This new alignment method not only utilizes the structural information but also provides a reference-free solution. We show that the performance of our new method is better than other methods in both simulations and real datasets. When the signal is from a mixture, deconvolution is needed to recover the pure sources. Many biological questions can be better addressed when the data is in the form of single sources, instead of mixtures. There is a classic unsupervised deconvolution method: Convex Analysis of Mixtures (CAM). However, there are some limitations of this method. For example, the time complexity of some steps is very high. Thus, when facing a large dataset or a dataset with many sources, the computation time would be extremely long. Also, since there are some stochastic and heuristic steps, the deconvolution result may be not accurate enough. We improved CAM and the experimental results show that the speed and accuracy of the deconvolution is significantly improved.

With significant growth in modern forms of communication such as social media and micro- blogs we are able to gain a real-time understanding into events happening in many parts of the world. In addition, these modern forms of communication have helped shed light into the increasing instabilities across the world via the design of anticipatory intelligence systems [45, 43, 20] that can forecast population level events like civil unrest, disease occurrences with reasonable accuracy. Event forecasting systems are generally prone to become outdated (model drift) as they fail to keep-up with constantly changing patterns and thus require regular re-training in order to sustain their accuracy and reliability. In this dissertation we try to address some of the issues associated with design and maintenance of event forecasting systems in general. We propose and showcase performance results for a drift adaptation technique in event forecasting systems and also build a hybrid system for event coding which is cognizant of and seeks human intervention in uncertain prediction contexts to maintain a good balance between prediction-fidelity and cost of human effort. Specifically we identify several micro-tasks for event coding and build separate pipelines for each with uncertainty estimation capabilities and thereby be able to seek human feedback whenever required for each micro-task independent of the rest. Doctor of Philosophy Event forecasting systems help reduce violence, loss/damage to humans and property. They find applicability in supply chain management, prioritizing citizen grievances, designing mea- sures to control violence and minimize disruptions and also in applications like health/tourism by providing timely travel alerts. Several issues exist with the design and maintenance of such event forecasting systems in general. Predictions from such systems may drift away from ground reality over time if not adapted to various shifts (or changes) in event occurrence patterns in real-time. A continuous source of ground-truth events is of paramount necessity for the continuous maintenance of forecasting systems. However ground-truth events used for training may not be reliable but often information about their uncertainty is not reflected in the systems that are used to build the ground truth. This dissertation focuses on addressing such issues pertaining to design and maintenance of event forecasting systems. We propose a framework for online drift-adaptation and also build machine learning methods capable of modeling and capturing uncertainty in event detection systems. Finally we propose and built a hybrid event coding system that can capture the best of both automated and manual event coders. We breakdown the overall event coding pipeline into several micro-tasks and propose individual methods for each micro-task. Each method is built with the capability to know what it doesn't know and thus is capable of balancing quality vs throughput based on available human resources.

This dissertation focuses on identifying various types of anomalous information pattern in social media and news outlets. We focus on three types of anomalous information, including (1) media censorship in news outlets, which is information that should be published but is actually missing, (2) fake news in social media, which is unreliable information shown to the public, and (3) media propaganda in news outlets, which is trustworthy information but being over-populated. For the first problem, existing approaches on censorship detection mostly rely on monitoring posts in social media. However, media censorship in news outlets has not received nearly as much attention, mostly because it is difficult to systematically detect. The contributions of our work include: (1) a hypothesis testing framework to identify and evaluate censored clusters of keywords, (2) a near-linear-time algorithm to identify the highest scoring clusters as indicators of censorship, and (3) extensive experiments on six Latin American countries for performance evaluation. For the second problem, existing approaches studying fake news in social media primarily focus on topic-level modeling or prediction based on a set of aggregated features from a col- lection of posts. However, the credibility of various information components within the same topic can be quite different. The contributions of our work in this space include: (1) a new benchmark dataset for fake news research, (2) a cluster-based approach to improve instance- level prediction of information credibility, and (3) extensive experiments for performance evaluations. For the last problem, existing approaches to media propaganda detection primarily focus on investigating the pattern of information shared over social media or evaluation from domain experts. However, these approaches cannot be generalized to a large-scale analysis of media propaganda in news outlets. The contributions of our work include: (1) non- parametric scan statistics to identify clusters of over-populated keywords, (2) a near-linear-time algorithm to identify the highest scoring clusters as indicators of propaganda, and (3) extensive experiments on two Latin American countries for performance evaluation. Doctor of Philosophy Nowadays, massive information is available through a variety of social media platforms. However, the information accessed by the audience might be not exactly correct in different ways. In order for the audience being able to get access to the correct information, we develop various machine learning algorithms to uncover the anomalous information pattern in social media and explain the reason behind this behavior. Our algorithms can be used to learn what different information patterns can exist in the open data source.

Machine Learning and Graph techniques are complex and challenging to distribute. Generally, they are distributed by modeling the problem in a similar way as single node sequential techniques except applied on smaller chunks of data and compute and the results combined. These techniques focus on stitching the results from smaller chunks as the best possible way to have the outcome as close to the sequential results on entire data as possible. This approach is not feasible in numerous kernel, matrix, optimization, graph, and other techniques where the algorithm needs access to all the data during execution. In this work, we propose key-value pair based distribution techniques that are widely applicable to statistical machine learning techniques along with matrix, graph, and time series based algorithms. The crucial difference with previously proposed techniques is that all operations are modeled on key-value pair based fine or coarse-grained steps. This allows flexibility in distribution with no compounding error in each step. The distribution is applicable not only in robust disk-based frameworks but also in in-memory based systems without significant changes. Key-value pair based techniques also provide the ability to generate the same result as sequential techniques with no edge or overlap effects in structures such as graphs or matrices to resolve. This thesis focuses on key-value pair based distribution of applied machine learning techniques on a variety of problems. For the first method key-value pair distribution is used for storytelling at scale. Storytelling connects entities (people, organizations) using their observed relationships to establish meaningful storylines. When performed sequentially these computations become a bottleneck because the massive number of entities make space and time complexity untenable. We present DISCRN, or DIstributed Spatio-temporal ConceptseaRch based StorytelliNg, a distributed framework for performing spatio-temporal storytelling. The framework extracts entities from microblogs and event data, and links these entities using a novel ConceptSearch to derive storylines in a distributed fashion utilizing key-value pair paradigm. Performing these operations at scale allows deeper and broader analysis of storylines. The novel parallelization techniques speed up the generation and filtering of storylines on massive datasets. Experiments with microblog posts such as Twitter data and GDELT(Global Database of Events, Language and Tone) events show the efficiency of the techniques in DISCRN. The second work determines brand perception directly from people's comments in social media. Current techniques for determining brand perception, such as surveys of handpicked users by mail, in person, phone or online, are time consuming and increasingly inadequate. The proposed DERIV system distills storylines from open data representing direct consumer voice into a brand perception. The framework summarizes the perception of a brand in comparison to peer brands with in-memory key-value pair based distributed algorithms utilizing supervised machine learning techniques. Experiments performed with open data and models built with storylines of known peer brands show the technique as highly scalable and accurate in capturing brand perception from vast amounts of social data compared to sentiment analysis. The third work performs event categorization and prospect identification in social media. The problem is challenging due to endless amount of information generated daily. In our work, we present DISTL, an event processing and prospect identifying platform. It accepts as input a set of storylines (a sequence of entities and their relationships) and processes them as follows: (1) uses different algorithms (LDA, SVM, information gain, rule sets) to identify themes from storylines; (2) identifies top locations and times in storylines and combines with themes to generate events that are meaningful in a specific scenario for categorizing storylines; and (3) extracts top prospects as people and organizations from data elements contained in storylines. The output comprises sets of events in different categories and storylines under them along with top prospects identified. DISTL utilizes in-memory key-value pair based distributed processing that scales to high data volumes and categorizes generated storylines in near real-time. The fourth work builds flight paths of drones in a distributed manner to survey a large area taking images to determine growth of vegetation over power lines allowing for adjustment to terrain and number of drones and their capabilities. Drones are increasingly being used to perform risky and labor intensive aerial tasks cheaply and safely. To ensure operating costs are low and flights autonomous, their flight plans must be pre-built. In existing techniques drone flight paths are not automatically pre-calculated based on drone capabilities and terrain information. We present details of an automated flight plan builder DIMPL that pre-builds flight plans for drones tasked with surveying a large area to take photographs of electric poles to identify ones with hazardous vegetation overgrowth. DIMPL employs a distributed in-memory key-value pair based paradigm to process subregions in parallel and build flight paths in a highly efficient manner. The fifth work highlights scaling graph operations, particularly pruning and joins. Linking topics to specific experts in technical documents and finding connections between experts are crucial for detecting the evolution of emerging topics and the relationships between their influencers in state-of-the-art research. Current techniques that make such connections are limited to similarity measures. Methods based on weights such as TF-IDF and frequency to identify important topics and self joins between topics and experts are generally utilized to identify connections between experts. However, such approaches are inadequate for identifying emerging keywords and experts since the most useful terms in technical documents tend to be infrequent and concentrated in just a few documents. This makes connecting experts through joins on large dense graphs challenging. We present DIGDUG, a framework that identifies emerging topics by applying graph operations to technical terms. The framework identifies connections between authors of patents and journal papers by performing joins on connected topics and topics associated with the authors at scale. The problem of scaling the graph operations for topics and experts is solved through dense graph pruning and graph joins categorized under their own scalable separable dense graph class based on key-value pair distribution. Comparing our graph join and pruning technique against multiple graph and join methods in MapReduce revealed a significant improvement in performance using our approach. Doctor of Philosophy Distribution of Machine Learning and Graph algorithms is commonly performed by modeling the core algorithm in the same way as the sequential technique except implemented on distributed framework. This approach is satisfactory in very few cases, such as depth-first search and subgraph enumerations in graphs, k nearest neighbors, and few additional common methods. These techniques focus on stitching the results from smaller data or compute chunks as the best possible way to have the outcome as close to the sequential results on entire data as possible. This approach is not feasible in numerous kernel, matrix, optimization, graph, and other techniques where the algorithm needs to perform exhaustive computations on all the data during execution. In this work, we propose key-value pair based distribution techniques that are exhaustive and widely applicable to statistical machine learning algorithms along with matrix, graph, and time series based operations. The crucial difference with previously proposed techniques is that all operations are modeled as key-value pair based fine or coarse-grained steps. This allows flexibility in distribution with no compounding error in each step. The distribution is applicable not only in robust disk-based frameworks but also in in-memory based systems without significant changes. Key-value pair based techniques also provide the ability to generate the same result as sequential techniques with no edge or overlap effects in structures such as graphs or matrices to resolve.

Effective information analysis generally boils down to the geometry of the data represented by a graph. Typical applications include social networks, transportation networks, the spread of epidemic disease, brain's neuronal networks, gene data on biological regulatory networks, telecommunication networks, knowledge graph, which are lying on the non-Euclidean graph domain. To describe the geometric structures, graph matrices such as adjacency matrix or graph Laplacian can be employed to reveal latent patterns. This thesis focuses on the theoretical analysis of graph neural networks and the development of methods for specific applications using graph representation. Four methods are proposed, including rational neural networks for jump graph signal estimation, RemezNet for robust attribute prediction in the graph, ICNet for integrated circuit security, and CNF-Net for dynamic circuit deobfuscation. For the first method, a recent important state-of-art method is the graph convolutional networks (GCN) nicely integrate local vertex features and graph topology in the spectral domain. However, current studies suffer from drawbacks: graph CNNs rely on Chebyshev polynomial approximation which results in oscillatory approximation at jump discontinuities since Chebyshev polynomials require degree $Omega$(poly(1/$epsilon$)) to approximate a jump signal such as $|x|$. To reduce complexity, RatioanlNet is proposed to integrate rational function and neural networks for graph node level embeddings. For the second method, we propose a method for function approximation which suffers from several drawbacks: non-robustness and infeasibility issue; neural networks are incapable of extracting analytical representation; there is no study reported to integrate the superiorities of neural network and Remez. This work proposes a novel neural network model to address the above issues. Specifically, our method utilizes the characterizations of Remez to design objective functions. To avoid the infeasibility issue and deal with the non-robustness, a set of constraints are imposed inspired by the equioscillation theorem of best rational approximation. The third method proposes an approach for circuit security. Circuit obfuscation is a recently proposed defense mechanism to protect digital integrated circuits (ICs) from reverse engineering. Estimating the deobfuscation runtime is a challenging task due to the complexity and heterogeneity of graph-structured circuit, and the unknown and sophisticated mechanisms of the attackers for deobfuscation. To address the above-mentioned challenges, this work proposes the first graph-based approach that predicts the deobfuscation runtime based on graph neural networks. The fourth method proposes a representation for dynamic size of circuit graph. By analyzing SAT attack method, a conjunctive normal form (CNF) bipartite graph is utilized to characterize the complexity of this SAT problem. To overcome the difficulty in capturing the dynamic size of the CNF graph, an energy-based kernel is proposed to aggregate dynamic features. Doctor of Philosophy Graph data is pervasive throughout most fields, including pandemic spread network, social network, transportation roads, internet, and chemical structure. Therefore, the applications modeled by graph benefit people's everyday life, and graph mining derives insightful opinions from this complex topology. This paper investigates an emerging technique called graph neural newton (GNNs), which is designed for graph data mining. There are two primary goals of this thesis paper: (1) understanding the GNNs in theory, and (2) apply GNNs for unexplored and values real-world scenarios. For the first goal, we investigate spectral theory and approximation theory, and a unified framework is proposed to summarize most GNNs. This direction provides a possibility that existing or newly proposed works can be compared, and the actual process can be measured. Specifically, this result demonstrates that most GNNs are either an approximation for a function of graph adjacency matrix or a function of eigenvalues. Different types of approximations are analyzed in terms of physical meaning, and the advantages and disadvantages are offered. Beyond that, we proposed a new optimization for a highly accurate but low efficient approximation. Evaluation of synthetic data proves its theoretical power, and the tests on two transportation networks show its potentials in real-world graphs. For the second goal, the circuit is selected as a novel application since it is crucial, but there are few works. Specifically, we focus on a security problem, a high-value real-world problem in industry companies such as Nvidia, Apple, AMD, etc. This problem is defined as a circuit graph as apply GNN to learn the representation regarding the prediction target such as attach runtime. Experiment on several benchmark circuits shows its superiority on effectiveness and efficacy compared with competitive baselines. This paper provides exploration in theory and application with GNNs, which shows a promising direction for graph mining tasks. Its potentials also provide a wide range of innovations in graph-based problems.

Causal graphs are a useful notation to represent the interplay between the actors as well as the polarity and strength of the relationship that they share. They are used extensively in educational, professional, and industrial contexts to simulate different scenarios, validate behavioral aspects, visualize the connections between different processes, and explore the adversarial effects of changing certain nodes. However, as the size of the causal graphs increase, interpreting them also becomes increasingly tougher. In such cases, new analytical tools are required to enhance the user's comprehension of the graph, both in terms of correctness and speed. To this purpose, this thesis introduces 1) a system that allows for causal exploration of directed graphs, while enabling the user to see the effect of interventions on the target nodes, 2) the use of natural language generation techniques to create a coherent passage explaining the propagation effects, and 3) results of an expert user study validating the efficacy of the narratives in enhancing the user's understanding of the causal graphs. In overall, the system aims to enhance user experience and promote further causal exploration. Master of Science Narrative generation is the art of creating coherent snippets of text that cumulatively describe a succession of events, played across a period of time. These goals of narrative generation are also shared by causal graphs – models that encapsulate inferences between the nodes through the strength and polarity of the connecting edges. Causal graphs are an useful mechanism to visualize changes propagating amongst nodes in the system. However, as the graph starts addressing real-world actors and their interactions, it becomes increasingly difficult to understand causal inferences between distant nodes, especially if the graph is cyclic. Moreover, if the value of more than a single node is altered and the cumulative effect of the change is to be perceived on a set of target nodes, it becomes extremely difficult to the human eye. This thesis attempts to alleviate this problem by generating dynamic narratives detailing the effect of one or more interventions on one or more target nodes, incorporating time-series analysis, Wikification, and spike detection. Moreover, the narrative enhances the user's understanding of the change propagation occurring in the system. The efficacy of the narrative was further corroborated by the results of user studies, which concluded that the presence of the narrative aids the user's confidence level, correctness, and speed while exploring the causal network.

The sun is constantly emitting electrons and ions as magnetized plasma, forming the solar wind and carrying with it the sun’s rotating magnetic field. The solar wind subsequently interacts with Earth’s magnetosphere and the magnetospheres of other planets. Occasionally large amounts of magnetized plasma are released at one time in a Corona mass ejection (CME). When the CME interacts with Earth’s magnetic field, it induces perturbations that may have a significant impact on critical infrastructure, for instance, by disturbing communication systems, and inducing currents on pipelines and electric power lines, which in turn may cause increased corrosion or blackouts, among other effects. In this dissertation, we analyze measured electrical signals provided to us by the Finnish Meteorological Institute, which were induced by geomagnetic storms on pipelines located in Finland and recorded in 2003. Specifically, we perform a wavelet leader multifractal analysis of these current signals to generate singularity spectra, and then using the bootstrapping technique, we apply statistical tests to show that these signals exhibit multifractal characteristics. We also performed surrogate testing to show that these characteristics were unique to the signal. Ph.D. Earth’s weather affects all of us every day. The solar space environment has weather of its own that affects us as well. Storms of a size that far exceed anything on Earth can impact Earth and affect our infrastructure. One of the most powerful phenomena that occur, called solar corona mass ejections, results when the sun ejects a large amount of plasma. This can interact with the Earth’s magnetic field, which in turn induces perturbations that may have a significant impact on critical infrastructure, for instance, by disturbing communication systems, and inducing currents on pipelines and electric power lines. The currents can cause increased corrosion or blackouts, among other effects. In this dissertation, we analyze measured electrical signals provided to us by the Finnish Meteorological Institute, which were induced by geomagnetic storms on pipelines located in Finland and recorded in 2003. Specifically, we perform a statistical analysis of these current signals to decide whether they exhibit multifractal characteristics.

The presence of noise and corruption in real-world data can be inevitably caused by accidental outliers, transmission loss, or even adversarial data attacks. Unlike traditional random noise usually assume a specific distribution with low corruption ratio, the data collected from crowdsourcing or labeled by weak annotators can contain adversarial data corruption. More challenge, the adversarial data corruption can be arbitrary, unbounded and do not follow any specific distribution. In addition, in the era of data explosion, the fast-growing amount of data makes the robust models more difficult to handle large-scale data sets. This thesis focuses on the development of methods for scalable robust models under the adversarial data corruption assumptions. Four methods are proposed, including robust regression via heuristic hard-thresholding, online and distributed robust regression with adversarial noises, self-paced robust learning for leveraging clean labels in noisy data, and robust regression via online feature selection with adversarial noises. Moreover, I extended the self-paced robust learning method to its distributed version for the scalability of the proposed algorithm, named distributed self-paced learning in alternating direction method of multiplier. Last, a robust multi-factor personality prediction model is proposed to hand the correlated data noises. For the first method, existing solutions for robust regression lack rigorous recovery guarantee of regression coefficients under the adversarial data corruption with no prior knowledge of corruption ratio. The proposed contributions of our work include: (1) Propose efficient algorithms to address the robust least-square regression problem; (2) Design effective approaches to estimate the corruption ratio; (3) Provide a rigorous robustness guarantee for regression coefficient recovery; and (4) Conduct extensive experiments for performance evaluation. For the second method, existing robust learning methods typically focus on modeling the entire dataset at once; however, they may meet the bottleneck of memory and computation as more and more datasets are becoming too large to be handled integrally. The proposed contributions of our work for this task include: (1) Formulate a framework for the scalable robust least-squares regression problem; (2) Propose online and distributed algorithms to handle the adversarial corruption; (3) Provide a rigorous robustness guarantee for regression coefficient recovery; and (4) Conduct extensive experiments for performance evaluations. For the third method, leveraging the prior knowledge of clean labels in noisy data is actually a crucial issue in practice, but existing robust learning methods typically focus more on eliminating noisy data. However, the data collected by ``weak annotator" or crowd-sourcing can be too noisy for existing robust methods to train an accurate model. Moreover, existing work that utilize additional clean labels are usually designed for some specific problems such as image classification. These methods typically utilize clean labels in large-scale noisy data based on their additional domain knowledge; however, these approaches are difficult to handle extremely noisy data and relied on their domain knowledge heavily, which makes them difficult be used in more general problems. The proposed contributions of our work for this task include: (1) Formulating a framework to leverage the clean labels in noisy data; (2) Proposing a self-paced robust learning algorithm to train models under the supervision of clean labels; (3) Providing a theoretical analysis for the convergence of the proposed algorithm; and (4) Conducting extensive experiments for performance evaluations. For the fourth method, the presence of data corruption in user-generated streaming data, such as social media, motivates a new fundamental problem that learns reliable regression coefficient when features are not accessible entirely at one time. Until now, several important challenges still cannot be handled concurrently: 1) corrupted data estimation when only partial features are accessible; 2) online feature selection when data contains adversarial corruption; and 3) scaling to a massive dataset. This paper proposes a novel RObust regression algorithm via Online Feature Selection (textit{RoOFS}) that concurrently addresses all the above challenges. Specifically, the algorithm iteratively updates the regression coefficients and the uncorrupted set via a robust online feature substitution method. We also prove that our algorithm has a restricted error bound compared to the optimal solution. Extensive empirical experiments in both synthetic and real-world data sets demonstrated that the effectiveness of our new method is superior to that of existing methods in the recovery of both feature selection and regression coefficients, with very competitive efficiency. For the fifth method, existing self-paced learning approaches typically focus on modeling the entire dataset at once; however, this may introduce a bottleneck in terms of memory and computation, as today's fast-growing datasets are becoming too large to be handled integrally. The proposed contributions of our work for this task include: (1) Reformulate the self-paced problem into a distributed setting.; (2) A distributed self-paced learning algorithm based on consensus ADMM is proposed to solve the textit{SPL} problem in a distributed setting; (3) A theoretical analysis is provided for the convergence of our proposed textit{DSPL} algorithm; and (4) Extensive experiments have been conducted utilizing both synthetic and real-world data based on a robust regression task. For the last method, personality prediction in multiple factors, such as openness and agreeableness, is growing in interest especially in the context of social media, which contains massive online posts or likes that can potentially reveal an individual's personality. However, the data collected from social media inevitably contains massive amounts of noise and corruption. To address it, traditional robust methods still suffer from several important challenges, including 1) existence of correlated corruption among multiple factors, 2) difficulty in estimating the corruption ratio in multi-factor data, and 3) scalability to massive datasets. This paper proposes a novel robust multi-factor personality prediction model that concurrently addresses all the above challenges by developing a distributed robust regression algorithm. Specifically, the algorithm optimizes regression coefficients of each factor in parallel with a heuristically estimated corruption ratio and then consolidates the uncorrupted set from multiple factors in two strategies: global consensus and majority voting. We also prove that our algorithm benefits from strong guarantees in terms of convergence rates and coefficient recovery, which can be utilized as a generic framework for the multi-factor robust regression problem with correlated corruption property. Extensive experiment on synthetic and real dataset demonstrates that our algorithm is superior to those of existing methods in both effectiveness and efficiency. Doctor of Philosophy Social media has experienced a rapid growth during the past decade. Millions of users of sites such as Twitter have been generating and sharing a wide variety of content including texts, images, and other metadata. In addition, social media can be treated as a social sensor that reflects different aspects of our society. Event analytics in social media have enormous significance for applications like disease surveillance, business intelligence, and disaster management. Social media data possesses a number of important characteristics including dynamics, heterogeneity, noisiness, timeliness, big volume, and network properties. These characteristics cause various new challenges and hence invoke many interesting research topics, which will be addressed here. This dissertation focuses on the development of five novel methods for social media-based spatiotemporal event detection and forecasting. The first of these is a novel unsupervised approach for detecting the dynamic keywords of spatial events in targeted domains. This method has been deployed in a practical project for monitoring civil unrest events in several Latin American regions. The second builds on this by discovering the underlying development progress of events, jointly considering the structural contexts and spatiotemporal burstiness. The third seeks to forecast future events using social media data. The basic idea here is to search for subtle patterns in specific cities as indicators of ongoing or future events, where each pattern is defined as a burst of context features (keywords) that are relevant to a specific event. For instance, an initial expression of discontent gas price increases could actually be a potential precursor to a more general protest about government policies. Beyond social media data, in the fourth method proposed here, multiple data sources are leveraged to reflect different aspects of the society for event forecasting. This addresses several important problems, including the common phenomena that different sources may come from different geographical levels and have different available time periods. The fifth study is a novel flu forecasting method based on epidemics modeling and social media mining. A new framework is proposed to integrate prior knowledge of disease propagation mechanisms and real-time information from social media.

Among many smart grid technologies, demand response (DR) is gaining increasing popularity. Many utility companies provide a variety of programs to encourage DR participation. Under these circumstances, various building energy management (BEM) systems have emerged to facilitate the building control during a DR event. Nonetheless, due to the cost and return on investment, these solutions mainly target homes and large commercial buildings, leaving aside small- and medium-sized commercial buildings (SMCB). SMCB, however, accounts for 90% of commercial buildings in the US, and offer great potential of load reduction during peak hours. With the advent of Internet-of-Things (IoT) devices and technologies, low cost smart building solutions have become possible for the SMCB; nonetheless, related intelligent algorithms are not widely available. This dissertation work investigates automated building control algorithms, tailored for the SMCB, to realize automatic device control during DR events. To be specific, a control framework for Air-Conditioning (AC) units' coordination is proposed. The goal of such framework is to reduce the aggregated AC power consumption while maintaining the thermal comfort inside a building during DR events. To achieve this goal, three major components of the framework were studied: building thermal property modeling, AC power consumption modeling and control algorithms design. Firstly, to consider occupants' thermal comfort, a reverse thermal model was designed to predict the indoor temperature of thermal zones under different AC control signals. The model was trained with supervised learning using coarse-grained temperature data recorded by smart thermostats; thus, it requires no lengthy configuration as a forward model does. The cost efficiency and plug-and-play feature of the model make it appropriate for SMCB. Secondly, a power disaggregation algorithm is proposed to model the power-outdoor temperature relationship of multiple AC units, using data from a single power meter and thermostats. Finally, algorithms based on mixed integer linear programming (MILP) and reinforcement learning (RL) were devised to coordinate multiple AC units in a building during a DR event. Integrated with the thermal model and AC power consumption model, these algorithms minimize occupants' thermal discomfort while restricting the aggregated AC power consumption below the DR limit. The efficiency of these control algorithms was tested, which demonstrate that they can generate AC control schedule in short notice (5 minutes) ahead of a DR event. Verification and validation of the proposed framework was conducted in both simulation and actual building environments. In addition, though the framework is designed for SMCBs, it can also be applied to large homes with multiple AC units to coordinate. This work is expected to give an insight into the BEM sector, helping the popularization of implementing DR in buildings. The research findings from this dissertation work shows the validity of the proposed algorithms, which can be used in BEM systems and cloud-based smart thermostats to exploit the untapped DR resource in SMCB. PHD For power system operation, the demand and supply should be equal at all time. During peak hours, the demand becomes very high. One way to keep the balance is to provide more generation capacity, and thus more expensive and less efficient generators are brought online, which causes higher production cost and more pollution. Instead, an alternative is to encourage the load reduction via demand response (DR): customers reduce load upon receiving a signal sent by the utility company, usually in exchange for some monetary payback. For buildings to participate in DR, an affordable automation system and related control algorithms are needed. This dissertation proposed a cost-effective, self-learning and data-driven framework to facilitate small- and medium-sized commercial buildings or large homes in air-conditioner (AC) units control during DR events. The devised framework requires little human configuration; it learns the building behavior by analyzing the operation data. Two algorithms are proposed to coordinate multiple AC units in a building with two goals: firstly, reducing the total AC power consumption below certain limit, as agreed between the building owners and their utility company. Secondly, minimizing occupants’ thermal discomfort caused by limiting AC operation. The effectiveness of the framework is investigated in this dissertation based on data collected from a real building.

Identification of travelers' transportation modes is a fundamental step for various problems that arise in the domain of transportation such as travel demand analysis, transport planning, and traffic management. This thesis aims to identify travelers' transportation modes purely based on their GPS trajectories. First, a segmentation process is developed to partition a user's trip into GPS segments with only one transportation mode. A majority of studies have proposed mode inference models based on hand-crafted features, which might be vulnerable to traffic and environmental conditions. Furthermore, the classification task in almost all models have been performed in a supervised fashion while a large amount of unlabeled GPS trajectories has remained unused. Accordingly, a deep SEmi-Supervised Convolutional Autoencoder (SECA) architecture is proposed to not only automatically extract relevant features from GPS segments but also exploit useful information in unlabeled data. The SECA integrates a convolutional-deconvolutional autoencoder and a convolutional neural network into a unified framework to concurrently perform supervised and unsupervised learning. The two components are simultaneously trained using both labeled and unlabeled GPS segments, which have already been converted into an efficient representation for the convolutional operation. An optimum schedule for varying the balancing parameters between reconstruction and classification errors are also implemented. The performance of the proposed SECA model, trip segmentation, the method for converting a raw trajectory into a new representation, the hyperparameter schedule, and the model configuration are evaluated by comparing to several baselines and alternatives for various amounts of labeled and unlabeled data. The experimental results demonstrate the superiority of the proposed model over the state-of-the-art semi-supervised and supervised methods with respect to metrics such as accuracy and F-measure. Master of Science Identifying users' transportation modes (e.g., bike, bus, train, and car) is a key step towards many transportation related problems including (but not limited to) transport planning, transit demand analysis, auto ownership, and transportation emissions analysis. Traditionally, the information for analyzing travelers' behavior for choosing transport mode(s) was obtained through travel surveys. High cost, low-response rate, time-consuming manual data collection, and misreporting are the main demerits of the survey-based approaches. With the rapid growth of ubiquitous GPS-enabled devices (e.g., smartphones), a constant stream of users' trajectory data can be recorded. A user's GPS trajectory is a sequence of GPS points, recorded by means of a GPS-enabled device, in which a GPS point contains the information of the device geographic location at a particular moment. In this research, users' GPS trajectories, rather than traditional resources, are harnessed to predict their transportation mode by means of statistical models. With respect to the statistical models, a wide range of studies have developed travel mode detection models using on hand-designed attributes and classical learning techniques. Nonetheless, hand-crafted features cause some main shortcomings including vulnerability to traffic uncertainties and biased engineering justification in generating effective features. A potential solution to address these issues is by leveraging deep learning frameworks that are capable of capturing abstract features from the raw input in an automated fashion. Thus, in this thesis, deep learning architectures are exploited in order to identify transport modes based on only raw GPS tracks. It is worth noting that a significant portion of trajectories in GPS data might not be annotated by a transport mode and the acquisition of labeled data is a more expensive and labor-intensive task in comparison with collecting unlabeled data. Thus, utilizing the unlabeled GPS trajectory (i.e., the GPS trajectories that have not been annotated by a transport mode) is a cost-effective approach for improving the prediction quality of the travel mode detection model. Therefore, the unlabeled GPS data are also leveraged by developing a novel deep-learning architecture that is capable of extracting information from both labeled and unlabeled data. The experimental results demonstrate the superiority of the proposed models over the state-of-the-art methods in literature with respect to several performance metrics.

Dynamic query expansion is a method of automatically identifying terms relevant to a target domain based on an incomplete query input. With the explosive growth of online media, such tools are essential for efficient search result refining to track emerging themes in noisy, unstructured text streams. It's crucial for large-scale predictive analytics and decision-making, systems which use open source indicators to find meaningful information rapidly and accurately. The problems of information overload and semantic mismatch are systemic during the Information Retrieval (IR) tasks undertaken by such systems. In this dissertation, we develop approaches to dynamic query expansion algorithms that can help improve the efficacy of such systems using only a small set of seed queries and requires no training or labeled samples. We primarily investigate four significant problems related to the retrieval and assessment of event-related information, viz. (1) How can we adapt the query expansion process to support rank-based analysis when tracking a fixed set of entities? A scalable framework is essential to allow relative assessment of emerging themes such as airport threats. (2) What visual knowledge discovery framework to adopt that can incorporate users' feedback back into the search result refinement process? A crucial step to efficiently integrate real-time `situational awareness' when monitoring specific themes using open source indicators. (3) How can we contextualize query expansions? We focus on capturing semantic relatedness between a query and reference text so that it can quickly adapt to different target domains. (4) How can we synchronously perform knowledge discovery and characterization (unstructured to structured) during the retrieval process? We mainly aim to model high-order, relational aspects of event-related information from microblog texts. Ph. D. Analysis of real-time, social media can provide critical insights into ongoing societal events. Where consequences and implications of specific events include monetary losses, threats to critical infrastructure and national security, disruptions to daily life, and a potential to cause loss of life and physical property. It is imperative for developing good ‘ground truth’ to develop adequate data-driven information systems, i.e., an authoritative record of events reported in the media cataloged alongside important dimensions. Availability of high-quality ground truth events can support various analytic efforts, e.g., identifying precursors of attacks, developing predictive indicators using surrogate data sources, and tracking the progression of events over space and time. A dynamic search result refinement is useful for expanding a general set of user queries into a more relevant collection. The challenges of information overload and misalignment of context between the user query and retrieved results can overwhelm both human and machine. In this dissertation, we focus our efforts on these specific challenges. With the ever-increasing volume of user-generated data large-scale analysis is a tedious task. Our first focus is to develop a scalable model that dynamically tracks and ranks evolving topics as they appear in social media. Then to simplify the cognitive tasks involving sense-making of evolving themes, we take a visual approach to retrieve situationally critical and emergent information effectively. This visual analytics approach learns from user’s interactions during the exploratory process and then generates a better representation of the data. Thus, improving the situational understanding and usability of underlying data models. Such features are crucial for big-data based decision & support systems. To make the event-focused retrieval process more robust, we developed a context-rich procedure that adds new relevant key terms to the user’s original query by utilizing the linguistic structures in text. This context-awareness allows the algorithm to retrieve those relevant characteristics that can help users to gain adequate information from social media about real-world events. Online social commentary about events is very informal and can be incomplete. However, to get the complete picture and adequately describe these events we develop an approach that models the underlying relatedness of information and iteratively extract meaning and denotations from event-related texts. We learn how to express the high-order relationships between events and entities and group them to identify those attributes that best explain the events the user is trying to uncover. In all the augmentations we develop, our strategy is to allow only very minimal human supervision using just a small set of seed event triggers and requires no training or labeled samples. We show a comprehensive evaluation of these augmentations on real-world domains - threats on airports, cyber attacks, and protests. We also demonstrate their applicability as for real-time analysis that provides vital event characteristics, and contextually consistent information can be a beneficial aid for emergency responders.

This dissertation is focused on the design and evaluation of machine learning algorithms for modeling information precursors for use in event modeling and forecasting. Given an online stream of information (e.g., news articles, social media postings), how can we model and understand how events unfold, how they influence each other, and how they can act as determinants of future events? First, we study information reciprocity in joint news and social media streams to capture how events evolve. We present an online story chaining algorithm which links related news articles together in a low complexity manner and a mechanism to classify the interaction between a news article and social media (Twitter) activity into four categories. This is followed by identification of major information sources for a given story chain based on the interaction states of news and Twitter. We demonstrate through this study that Twitter as a social network platform serves as a fast way to draw attention from the public to many social events such as sports, whereas news media is quicker to report events regarding political, economical, and business issues. In the second problem we focus on forecasting and understanding large-scale societal events from open source datasets. Our goal here is to develop algorithms that can automatically reconstruct precursors to societal events. We develop a nested framework involving multi-instance learning for mining precursors by harnessing temporal constraints. We evaluate the proposed model for various event categories in multiple geo-locations with comprehensive experiments. Next, to reinforce the fact that events are typically inter-connected and influenced by events in other locations, we develop an approach that creates personalized models for exploring spatio-temporal event correlations; this approach also helps tackle data/label sparsity problems across geolocations. Finally, this dissertation demonstrates how our algorithms can be used to study key characteristics of mass events such as protests. Some mass gatherings run the risk of turning violent, causing damage to both property and people. We propose a tailored solution for uncovering triggers from both news media and social media for violent event analysis. This work was partially supported by the Intelligence Advanced Research Projects Activity (IARPA) via Department of Interior National Business Center (DoI/NBC) contract number D12PC000337, the Office of Naval Research under contract N00014-16-C-1054, and the U.S. Department of Homeland Security under Grant Award Number 2017-ST-061-CINA01. The US Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright annotation thereon. The views and conclusions contained herein are those of the author and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of NSF, IARPA, DoI/NBC, or the US Government. Ph. D. Today, massive open source information is widely available through news and social media, but analyzing this information is a complex task. It is imperative to develop algorithms that can automatically reconstruct the clues to societal events that are reported in news or social media. The focus of this dissertation is on simultaneously uncovering precursors to societal events and using such precursors to forecast upcoming events. We develop various machine learning algorithms that can model event-related data and determine the key happenings prior to an event that have the greatest predictability to such events in the future. We use our algorithms to understand the nature of precursors to civil unrest events (protests, strikes, and ‘occupy’ events) and why some of these events turn violent.

Temporal data is ubiquitous nowadays and can be easily found in many applications. Consider the extensively studied social media website Twitter. All the information can be associated with time stamps, and thus form different types of data sequences: a sequence of feature values of users who retweet a message, a sequence of tweets from a certain user, or a sequence of the evolving friendship networks. Mining these data sequences is an important task, which reveals patterns in the sequences, and it is a very challenging task as it usually requires different techniques for different sequences. The problem becomes even more complicated when the sequences are correlated. In this dissertation, we study the following two types of data sequences, and we show how to carefully exploit within-sequence and across-sequence correlations to develop more effective and scalable algorithms. 1. Multi-dimensional value sequences: We study sequences of multi-dimensional values, where each value is associated with a time stamp. Such value sequences arise in many domains such as epidemiology (medical records), social media (keyword trends), etc. Our goals are: for individual sequences, to find a segmentation of the sequence to capture where the pattern changes; for multiple correlated sequences, to use the correlations between sequences to further improve our segmentation; and to automatically find explanations of the segmentation results. 2. Social media post sequences: Driven by applications from popular social media websites such as Twitter and Weibo, we study the modeling of social media post sequences. Our goal is to understand how the posts (like tweets) are generated and how we can gain understanding of the users behind these posts. For individual social media post sequences, we study a prediction problem to find the users' latent state changes over the sequence. For dependent post sequences, we analyze the social influence among users, and how it affects users in generating posts and links. Our models and algorithms lead to useful discoveries, and they solve real problems in Epidemiology, Social Media and Critical Infrastructure Systems. Further, most of the algorithms and frameworks we propose can be extended to solve sequence mining problems in other domains as well. Ph. D.

Epigenetic modifications, such as DNA methylation and histone modifications, play important roles in gene expression and regulation, and are highly involved in cellular processes such as stem cell pluripotency/differentiation and tumorigenesis. Chromatin immunoprecipitation (ChIP) is the technique of choice for examining in vivo DNA-protein interactions and has been a great tool for studying epigenetic mechanisms. However, conventional ChIP assays require millions of cells for tests and are not practical for examination of samples from lab animals and patients. Automated microfluidic chips offer the advantage to handle small sample sizes and facilitate rapid reaction. They also eliminate cumbersome manual handling. In this report, I will talk about three different projects that utilized microfluidic immunoprecipitation followed by next genereation sequencing technologies to enable low input and high through epigenomics profiling. First, I examined RNA polymerase II transcriptional regulation with microfluidic chromatin immunoprecipitation followed by next generation sequencing (ChIP-seq) assays. Second, I probed the temporal dynamics in the DNA methylome during cancer development using a transgenic mouse model with microfluidic methylated DNA immunoprecipitation followed by next generation sequencing (MeDIP-seq) assays. Third, I explored negative enrichment of circulating tumor cells (CTCs) followed by microfluidic ChIP-seq technology for studying temporal dynamic histone modification (H3K4me3) of patient-derived tumor xenograft on an immunodeficient mouse model during the course of cancer metastasis. In the first study, I adapted microfluidic ChIP-seq devices to achieve ultrahigh sensitivity to study Pol2 transcriptional regulation from scarce cell samples. I dramatically increased the assay sensitivity to an unprecedented level (~50 K cells for pol2 ChIP-seq). Importantly, this is three orders of magnitude more sensitive than the prevailing pol2 ChIP-seq assays. I showed that MNase digestion provided better ChIP-seq signal than sonication, and two-steps fixation with MNase digestion provided the best ChIP-seq quality followed by one-step fixation with MNase digestion, and lastly, no fixation with MNase digestion. In the second study, I probed dynamic epigenomic changes during tumorigenesis using mice often require profiling epigenomes using a tiny quantity of tissue samples. Conventional epigenomic tests do not support such analysis due to the large amount of materials required by these assays. In this study, I developed an ultrasensitive microfluidics-based methylated DNA immunoprecipitation followed by next-generation sequencing (MeDIP-seq) technology for profiling methylomes using as little as 0.5 ng DNA (or ~100 cells) with 1.5 h on-chip process for immunoprecipitation. This technology enabled me to examine genome-wide DNA methylation in a C3(1)/SV40 T-antigen transgenic mouse model during different stages of mammary cancer development. Using this data, I identified differentially methylated regions and their associated genes in different periods of cancer development. Interestingly, the results showed that methylomic features are dynamic and change with tumor developmental stage. In the last study, I developed a negative enrichment of CTCs followed by ultrasensitive microfluidic ChIP-seq technology for profiling histone modification (H3K4Me3) of CTCs to resolve the technical challenges associated with CTC isolation and difficulties related with tools for profiling whole genome histone modification on tiny cell samples. Ph. D.

Automated event extraction from free text remains an open problem, particularly when the goal is to identify all relevant events. Manual extraction is currently the only alternative for comprehensive and reliable extraction. Therefore, it is required to have a system that can comprehensively extract events reported in news articles (high recall) and is also scalable enough to handle a large number of articles. In this dissertation, we explore various methods to develop an event extraction system that can mitigate these challenges. We primarily investigate three major problems related to event extraction as follows. (i) What are the strengths and weaknesses of the automated event extractors? A thorough understanding of what can be automated with high success and what leads to common pitfalls is crucial before we could develop a superior event extraction system. (ii) How can we build a hybrid event extraction system that can bridge the gap between manual and automated event extraction? Hybrid extraction is a semi-automated approach that uses an ecosystem of machine learning models along with a carefully designed user interface for extracting events. Since this method is semi-automated it also requires a meticulous understanding of user behavior in order to identify tasks that humans can perform with ease while diverting the more tedious task to the machine learning methods (iii) Finally, we explore methods for displaying extracted events that could simplify the analytical and inference generation processes for an analyst. We particularly aim to develop visualizations that would allow analysts can perform macro and micro level analysis of significant societal events. Ph. D.

Event extraction is a type of information extraction(IE) that works on extracting the specific knowledge of certain incidents from texts. Nowadays the amount of available information (such as news, blogs, and social media) grows in exponential order. Therefore, it becomes imperative to develop algorithms that automatically extract the machine-readable information from large volumes of text data. In this dissertation, we focus on three problems in obtaining event-related information from news articles. (1) The first effort is to comprehensively analyze the performance and challenges in current large-scale event encoding systems. (2) The second problem involves event detection and critical information extractions from news articles. (3) Third, the efforts concentrate on event-encoding which aims to extract event extent and arguments from texts. We start by investigating the two large-scale event extraction systems (ICEWS and GDELT) in the political science domain. We design a set of experiments to evaluate the quality of the extracted events from the two target systems, in terms of reliability and correctness. The results show that there exist significant discrepancies between the outputs of automated systems and hand-coded system and the accuracy of both systems are far away from satisfying. These findings provide preliminary background and set the foundation for using advanced machine learning algorithms for event related information extraction. Inspired by the successful application of deep learning in Natural Language Processing (NLP), we propose a Multi-Instance Convolutional Neural Network (MI-CNN) model for event detection and critical sentences extraction without sentence level labels. To evaluate the model, we run a set of experiments on a real-world protest event dataset. The result shows that our model could be able to outperform the strong baseline models and extract the meaningful key sentences without domain knowledge and manually designed features. We also extend the MI-CNN model and propose an MIMTRNN model for event extraction with distant supervision to overcome the problem of lacking fine level labels and small size training data. The proposed MIMTRNN model systematically integrates the RNN, Multi-Instance Learning, and Multi-Task Learning into a unified framework. The RNN module aims to encode into the representation of entity mentions the sequential information as well as the dependencies between event arguments, which are very useful in the event extraction task. The Multi-Instance Learning paradigm makes the system does not require the precise labels in entity mention level and make it perfect to work together with distant supervision for event extraction. And the Multi-Task Learning module in our approach is designed to alleviate the potential overfitting problem caused by the relatively small size of training data. The results of the experiments on two real-world datasets(Cyber-Attack and Civil Unrest) show that our model could be able to benefit from the advantage of each component and outperform other baseline methods significantly. Ph. D.

The rise of big data, especially social media data (e.g., Twitter, Facebook, Youtube), gives new opportunities to the understanding of human behavior. Consequently, novel computing methods for mining patterns in social media data are therefore desired. Through applying these approaches, it has become possible to aggregate public available data to capture triggers underlying events, detect on-going trends, and forecast future happenings. This thesis focuses on developing methods for social media analysis. Specifically, five directions are proposed here: 1) semi-supervised detection for targeted-domain events, 2) topical interaction study among multiple datasets, 3) discriminative learning about the identifications for common and distinctive topics, 4) epidemics modeling for flu forecasting with simulation via signals from social media data, 5) storyline generation for massive unorganized documents. Ph. D.

Networks are a mathematical abstraction of the interactions between a set of entities, with extensive applications in social science, epidemiology, bioinformatics, and cybersecurity, among others. There are many fundamental problems when analyzing network data, such as anomaly detection, dense subgraph mining, motif finding, information diffusion, and epidemic spread. A common underlying task in all these problems is finding an "interesting subgraph"; that is, finding a part of the graph---usually small relative to the whole---that optimizes a score function and has some property of interest, such as connectivity or a minimum density. Finding subgraphs that satisfy common constraints of interest, such as the ones above, is computationally hard in general, and state-of-the-art algorithms for many problems in network analysis are heuristic in nature. These methods are fast and usually easy to implement. However, they come with no theoretical guarantees on the quality of the solution, which makes it difficult to assess how the discovered subgraphs compare to an optimal solution, which in turn affects the data mining task at hand. For instance, in anomaly detection, solutions with low anomaly score lead to sub-optimal detection power. On the other end of the spectrum, there have been significant advances on approximation algorithms for these challenging graph problems in the theoretical computer science community. However, these algorithms tend to be slow, difficult to implement, and they do not scale to the large datasets that are common nowadays. The goal of this dissertation is developing scalable algorithms with theoretical guarantees for various network analysis problems, where the underlying task is to find subgraphs with constraints. We find interesting subgraphs with guarantees by adapting techniques from parameterized complexity, convex optimization, and submodularity optimization. These techniques are well-known in the algorithm design literature, but they lead to slow and impractical algorithms. One unifying theme in the problems that we study is that our methods are scalable without sacrificing the theoretical guarantees of these algorithm design techniques. We accomplish this combination of scalability and rigorous bounds by exploiting properties of the problems we are trying to optimize, decomposing or compressing the input graph to a manageable size, and parallelization. We consider problems on network analysis for both static and dynamic network models. And we illustrate the power of our methods in applications, such as public health, sensor data analysis, and event detection using social media data. Ph. D.

The rapid development of biotechnology has enabled researchers to collect high-throughput data for studying various biological processes at the genomic level, transcriptomic level, and proteomic level. Due to the large noise in the data and the high complexity of diseases (such as cancer), it is a challenging task for researchers to extract biologically meaningful information that can help reveal the underlying molecular mechanisms. The challenges call for more efforts in developing efficient and effective computational methods to analyze the data at different levels so as to understand the biological systems in different aspects. In this dissertation research, we have developed novel Bayesian approaches to infer alternative splicing mechanisms in biological systems using RNA sequencing data. Specifically, we focus on two research topics in this dissertation: isoform identification and phenotype-specific transcript assembly. For isoform identification, we develop a computational approach, SparseIso, to jointly model the existence and abundance of isoforms in a Bayesian framework. A spike-and-slab prior is incorporated into the model to enforce the sparsity of expressed isoforms. A Gibbs sampler is developed to sample the existence and abundance of isoforms iteratively. For transcript assembly, we develop a Bayesian approach, IntAPT, to assemble phenotype-specific transcripts from multiple RNA sequencing profiles. A two-layer Bayesian framework is used to model the existence of phenotype-specific transcripts and the transcript abundance in individual samples. Based on the hierarchical Bayesian model, a Gibbs sampling algorithm is developed to estimate the joint posterior distribution for phenotype-specific transcript assembly. The performances of our proposed methods are evaluated with simulation data, compared with existing methods and benchmarked with real cell line data. We then apply our methods on breast cancer data to identify biologically meaningful splicing mechanisms associated with breast cancer. For the further work, we will extend our methods for de novo transcript assembly to identify novel isoforms in biological systems; we will incorporate isoform-specific networks into our methods to better understand splicing mechanisms in biological systems. Ph. D.

Trust is the cornerstone of success in any relationship between two or more parties. Generally, we do not socialize, seek advice, consult, cooperate, buy or sell goods and services from/to others unless we establish some level of mutual trust between interacting parties. When e-commerce was merging infancy, the concept of trusting an entity in a virtual world was a huge obstacle. Gradually, increasingly-sophisticated, largely generic reputation scoring and management systems were embedded into the evolving marketplaces. Current technologies to include cloud computing, social networking, and mobile applications, coupled with the explosion in storage capacity and processing power, are evolving large-scale global marketplaces for a wide variety of resources and services, such as Amazon.com, BitTorrent, WebEx and Skype. In such marketplaces, user entities, or users for short; namely, consumers, providers and brokers, are largely autonomous with vastly diverse requirements, capabilities and trust profiles. Users' requirements may include service quality levels, cost, ease of use, etc. Users' capabilities may include assets owned or outsourced. Trustors' profiles may include what is promised and commitments to keep these promises. In such a large-scale heterogeneous marketplace, the trustworthy interactions and transactions in services and resources constitute a challenging endeavor. Currently, solving such issues generally adopts a "one-size fits all" trust models and systems. Such approach is limiting due to variations in technology, conflicts between users' requirements and/or conflicts between user requirements and service outcomes. Additionally, this approach may result in service providers being overwhelmed by adding new resources to satisfy all possible requirements, while having no information or guarantees about the level of trust they gain in the network. Accordingly, we hypothesize the need for personalizable customizable Trust Management Systems (TMSs) for the robustness and wide-scale adoption of large-scale marketplaces for resources and services. Most contemporary TMSs suffer from the following drawbacks: • Oblivious to diversities in trustors' requirements, • Primarily utilize feedback and direct or indirect experience as the only form of credentials and trust computations, • Trust computation methodologies are generally hardcoded and not reconfigurable, • Trust management operations, which we identify as monitoring, data management, analysis, expectation management, and decision making, are tightly coupled. Such coupling impedes customizability and personalization, and • Do not consider context in trust computations, where trust perspectives may vary from a context to another. Given these drawbacks and the large scale of the global marketplace of resources and services, a reference architecture for trust management systems is needed, which can incorporate current systems and may be used in guidance and development of a wide spectrum of trust management systems ranging from un-personalized to fully personalized systems. Up to our knowledge, no TMS reference architecture exists in the literature. In this dissertation, we propose a new Reference Architecture for Trust Management (RATM). The proposed reference architecture applies separation of concern among trust management operations; namely, decision expectation, analytics, data management and monitoring. RATM defines trust management operations through five reconfigurable components which collectively can be used to implement a wide spectrum of trust management systems ranging from generic to highly personalized systems. We used RATM for trust personalization, where we propose a Personalized Trust Management System (PTMS) based on RATM. We evaluated PTMS's scalability and demonstrated its effectiveness, efficiency and resilience by contrasting against a Generic Trust Management System (GTMS). We used two case studies for our evaluations; namely, BitTorrent and a video conferencing application. Intellectual Merit In this work, we propose RATM, a reference architecture for trust management systems that can be used to implement a wide variety of trust management systems ranging from generic systems (un-personalized) to highly personalized systems. We assume service-based environment where consumers, brokers and service providers are interacting and transacting in services and resources to satisfying their own trust requirements. We used RATM to implement a personalized trust management system (TMS). The main contributions of this work are as follows: • Proposing RATM for the guidance and development of a wide spectrum of TMSs ranging from un-personalized to fully personalized systems, and • Utilizing our RATM to propose and implement a personalized, scalable TMS with varying trust computation models. Broader Impact RATM provides reference architecture for trust management which can be used to develop and evaluate a wide spectrum of TMSs. Personalization, in general, paves the road for reaching high levels of satisfaction, where interacting parties' requirements are individually considered and thus consumers are served the best suited service(s). Hence, we claim that PTMS would largely enhance large-scale heterogeneous systems offering services and resources. This could lead to more cooperation, more transactions, more satisfaction, less malicious behavior and lower costs. PHD

Microfluidics has revolutionized how molecular biology studies are conducted. It permits profiling of genomic and epigenomic features for a wide range of applications. Microfluidics has been proven to be highly complementary to NGS technology with its unique capabilities for handling small volumes of samples and providing platforms for automation, integration, and multiplexing. In this thesis, we focus on three projects (diffusion-based PCR, MID-RRBS, and SurfaceChIP-seq), which improved the sensitivities of conventional assays by coupling with microfluidic technology. MID-RRBS and SurfaceChIP-seq projects were designed to profiling genome-wide DNA methylation and histone modifications, respectively. These assays dramatically improved the sensitivities of conventional approaches over 1000 times without compromising genomic coverages. We applied these assays to examine the neuronal/glial nuclei isolated from mouse brain tissues. We successfully identified the distinctive epigenomic signatures from neurons and glia. Another focus of this thesis is applying electrical field to investigate the intracellular contents. We report two projects, drug delivery to encapsulated bacteria and mRNA extraction under ultra-high electrical field intensity. We envision rapid growth in these directions, driven by the needs for testing scarce primary cells samples from patients in the context of precision medicine. Ph. D.

Transportation engineering is a wide area that covers different topics including traffic planning, highway design, pavement design, traffic safety, and traffic control. Certain concepts in those topics are challenging and are hard to understand based on textbooks and lectures. In this work, we developed five web games targeting the five topics in transportation engineering education to improve students’ understanding of those hard concepts. The games are hosted in a website server. Students can play these games online after register and login. The server stores the users’ information and their gameplay data. We conducted a Before-and-After study to test the effectiveness of the games in terms of improving the learning outcomes of the students. The results showed that the games could increase the students’ understanding of hard concepts significantly. The developed games can be used in transportation education. This game framework can serve as a reference for other education game developers. We envision that more educational games will be developed by transportation and education communities in the recent future. There will be more than one game for the same topic. We need an approach to select games for different students group. We proposed a gravity model for evaluating the engagement of the students for the educational games. We found that different games have different properties in terms of attracting students’ engagement. The proposed model can be used in the future for selecting educational games for specific students group. Master of Science This thesis presents the effort of making computer games for education purpose. I developed five games corresponding to the five areas in transportation engineering courses. The objective of this work is to let the students understand the hard concepts in transportation engineering by playing the developed games. The students can play the games online, and their gameplay data will be recorded as they play. The effectiveness of this work was tested using before-and-after quizzes. We designed a set of quizzes that are within transportation engineering and can be solved using the knowledge learned in the games. We asked the students to do the quizzes and, without any feedback, do the same quizzes again after playing the games. The result showed that their scores improved in general, which means their understanding of transportation engineering was improved. Using the data collected from the gameplay, quizzes, and the students’ course scores, I proposed a gravity model that describes how students were engaged in the games. I found that different games could attract different students.

With the advent of modern sensor technologies, significant opportunities have opened up to help conserve energy in residential and commercial buildings. Moreover, the rapid urbanization we are witnessing requires optimized energy distribution. This dissertation focuses on two sub-problems in improving energy conservation; energy disaggregation and occupancy prediction. Energy disaggregation attempts to separate the energy usage of each circuit or each electric device in a building using only aggregate electricity usage information from the meter for the whole house. The second problem of occupancy prediction can be accomplished using non-invasive indoor activity tracking to predict the locations of people inside a building. We cast both problems as temporal mining problems. We exploit motif mining with constraints to distinguish devices with multiple states, which helps tackle the energy disaggregation problem. Our results reveal that motif mining is adept at distinguishing devices with multiple power levels and at disentangling the combinatorial operation of devices. For the second problem we propose time-gap constrained episode mining to detect activity patterns followed by the use of a mixture of episode generating HMM (EGH) models to predict home occupancy. Finally, we demonstrate that the mixture EGH model can also help predict the location of a person to address non-invasive indoor activities tracking. Ph. D.

One of the most common measures of clinical outcomes is the survival time. Accurately linking cancer molecular profiling with survival outcome advances clinical management of cancer. However, existing survival analysis relies intensively on statistical evidence from a single level of data, without paying much attention to the integration of interacting multi-level data and the underlying biology. Advances in genomic techniques provide unprecedented power of characterizing the cancer tissue in a more complete manner than before, opening the opportunity of designing biologically informed and integrative approaches for survival analysis. Many cancer tissues have been profiled for gene expression levels and genomic variants (such as copy number alterations, sequence mutations, DNA methylation, and histone modification). However, it is not clear how to integrate the gene expression and genetic variants to achieve a better prediction and understanding of the cancer survival. To address this challenge, we propose two approaches for data integration in order to both biologically and statistically boost the features selection process for proper detection of the true predictive players of survival. The first approach is data-driven yet biologically informed. Consistent with the biological hierarchy from DNA to RNA, we prioritize each survival-relevant feature with two separate scores, predictive and mechanistic. With mRNA expression levels in concern, predictive features are those mRNAs whose variation in expression levels are associated with the survival outcome, and mechanistic features are those mRNAs whose variation in expression levels are associated with genomic variants (copy number alterations (CNAs) in this study). Further, we propose simultaneously integrating information from both the predictive model and the mechanistic model through our new approach GEMPS (Gene Expression as a Mediator for Predicting Survival). Applied on two cancer types (ovarian and glioblastoma multiforme), our method achieved better prediction power than peer methods. Gene set enrichment analysis confirms that the genes utilized for the final survival analysis are biologically important and relevant. The second approach is a generic mathematical framework to biologically regularize the Cox's proportional hazards model that is widely used in survival analysis. We propose a penalty function that both links the mechanistic model to the clinical model and reflects the biological downstream regulatory effect of the genomic variants on the mRNA expression levels of the target genes. Fast and efficient optimization principles like the coordinate descent and majorization-minimization are adopted in the inference process of the coefficients of the Cox model predictors. Through this model, we develop the regulator-target gene relationship to a new one: regulator-target-outcome relationship of a disease. Assessed via a simulation study and analysis of two real cancer data sets, the proposed method showed better performance in terms of selecting the true predictors and achieving better survival prediction. The proposed method gives insightful and meaningful interpretability to the selected model due to the biological linking of the mechanistic model and the clinical model. Other important forms of clinical outcomes are monitoring angiogenesis (formation of new blood vessels necessary for tumor to nourish itself and sustain its existence) and assessing therapeutic response. This can be done through dynamic imaging, in which a series of images at different time instances are acquired for a specific tumor site after injection of a contrast agent. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a noninvasive tool to examine tumor vasculature patterns based on accumulation and washout of the contrast agent. DCE-MRI gives indication about tumor vasculature permeability, which in turn indicates the tumor angiogenic activity. Observing this activity over time can reflect the tumor drug responsiveness and efficacy of the treatment plan. However, due to the limited resolution of the imaging scanners, a partial-volume effect (PVE) problem occurs, which is the result of signals from two or more tissues combining together to produce a single image concentration value within a pixel, with the effect of inaccurate estimation to the values of the pharmacokinetic parameters. A multi-tissue compartmental modeling (CM) technique supported by convex analysis of mixtures is used to mitigate the PVE by clustering pixels and constructing a simplex whose vertices are of a single compartment type. CAM uses the identified pure-volume pixels to estimate the kinetics of the tissues under investigation. We propose an enhanced version of CAM-CM to identify pure-volume pixels more accurately. This includes the consideration of the neighborhood effect on each pixel and the use of a barycentric coordinate system to identify more pure-volume pixels and to test those identified by CAM-CM. Tested on simulated DCE-MRI data, the enhanced CAM-CM achieved better performance in terms of accuracy and reproducibility. Ph. D.

Online social networks have become a staging ground for many modern movements, with the Arab Spring being the most prominent example. In an effort to understand and predict those movements, social media can be regarded as a valuable social sensor for disclosing underlying behaviors and patterns. To fully understand mass movement information propagation patterns in social networks, several problems need to be considered and addressed. Specifically, modeling mass movements that incorporate multiple spaces, a dynamic network structure, and misinformation propagation, can be exceptionally useful in understanding information propagation in social media. This dissertation explores four research problems underlying efforts to identify and track the adoption of mass movements in social media. First, how do mass movements become mobilized on Twitter, especially in a specific geographic area? Second, can we detect protest activity in social networks by observing group anomalies in graph? Third, how can we distinguish real movements from rumors or misinformation campaigns? and fourth, how can we infer the indicators of a specific type of protest, say climate related protest? A fundamental objective of this research has been to conduct a comprehensive study of how mass movement adoption functions in social networks. For example, it may cross multiple spaces, evolve with dynamic network structures, or consist of swift outbreaks or long term slowly evolving transmissions. In many cases, it may also be mixed with misinformation campaigns, either deliberate or in the form of rumors. Each of those issues requires the development of new mathematical models and algorithmic approaches such as those explored here. This work aims to facilitate advances in information propagation, group anomaly detection and misinformation distinction and, ultimately, help improve our understanding of mass movements and their adoption in social networks. Ph. D.

Nowadays, knowledge discovery on social media is attracting growing interest. Social media has become more than a communication tool, effectively functioning as a social sensor for our society. This dissertation focuses on the development of methods for social media-based spatiotemporal event detection and forecasting for a variety of event topics and assumptions. Five methods are proposed, namely dynamic query expansion for event detection, a generative framework for event forecasting, multi-task learning for spatiotemporal event forecasting, multi-source spatiotemporal event forecasting, and deep learning based epidemic modeling for forecasting influenza outbreaks. For the first of these methods, existing solutions for spatiotemporal event detection are mostly supervised and lack the flexibility to handle the dynamic keywords used in social media. The contributions of this work are: (1) Develop an unsupervised framework; (2) Design a novel dynamic query expansion method; and (3) Propose an innovative local modularity spatial scan algorithm. For the second of these methods, traditional solutions are unable to capture the spatiotemporal context, model mixed-type observations, or utilize prior geographical knowledge. The contributions of this work include: (1) Propose a novel generative model for spatial event forecasting; (2) Design an effective algorithm for model parameter inference; and (3) Develop a new sequence likelihood calculation method. For the third method, traditional solutions cannot deal with spatial heterogeneity or handle the dynamics of social media data effectively. This work's contributions include: (1) Formulate a multi-task learning framework for event forecasting; (2) simultaneously model static and dynamic terms; and (3) Develop efficient parameter optimization algorithms. For the fourth method, traditional multi-source solutions typically fail to consider the geographical hierarchy or cope with incomplete data blocks among different sources. The contributions here are: (1) Design a framework for event forecasting based on hierarchical multi-source indicators; (2) Propose a robust model for geo-hierarchical feature selection; and (3) Develop an efficient algorithm for model parameter optimization. For the last method, existing work on epidemic modeling either cannot ensure timeliness, or cannot characterize the underlying epidemic propagation mechanisms. The contributions of this work include: (1) Propose a novel integrated framework for computational epidemiology and social media mining; (2) Develop a semi-supervised multilayer perceptron for mining epidemic features; and (3) Design an online training algorithm. Ph. D.

Computational biology currently faces challenges in a big data world with thousands of data samples across multiple disease types including cancer. The challenging problem is how to extract biologically meaningful information from large-scale genomic data. Next-generation Sequencing (NGS) can now produce high quality data at DNA and RNA levels. However, in cells there exist a lot of non-specific (background) signals that affect the detection accuracy of true (foreground) signals. In this dissertation work, under Bayesian framework, we aim to develop and apply approaches to learn the distribution of genomic signals in each type of NGS data for reliable identification of specific foreground signals. We propose a novel Bayesian approach (ChIP-BIT) to reliably detect transcription factor (TF) binding sites (TFBSs) within promoter or enhancer regions by jointly analyzing the sample and input ChIP-seq data for one specific TF. Specifically, a Gaussian mixture model is used to capture both binding and background signals in the sample data; and background signals are modeled by a local Gaussian distribution that is accurately estimated from the input data. An Expectation-Maximization algorithm is used to learn the model parameters according to the distributions on binding signal intensity and binding locations. Extensive simulation studies and experimental validation both demonstrate that ChIP-BIT has a significantly improved performance on TFBS detection over conventional methods, particularly on weak binding signal detection. To infer cis-regulatory modules (CRMs) of multiple TFs, we propose to develop a Bayesian integration approach, namely BICORN, to integrate ChIP-seq and RNA-seq data of the same tissue. Each TFBS identified from ChIP-seq data can be either a functional binding event mediating target gene transcription or a non-functional binding. The functional bindings of a set of TFs usually work together as a CRM to regulate the transcription processes of a group of genes. We develop a Gibbs sampling approach to learn the distribution of CRMs (a joint distribution of multiple TFs) based on their functional bindings and target gene expression. The robustness of BICORN has been validated on simulated regulatory network and gene expression data with respect to different noise settings. BICORN is further applied to breast cancer MCF-7 ChIP-seq and RNA-seq data to identify CRMs functional in promoter or enhancer regions. In tumor cells, the normal regulatory mechanism may be interrupted by genome mutations, especially those somatic mutations that uniquely occur in tumor cells. Focused on a specific type of genome mutation, structural variation (SV), we develop a novel pattern-based probabilistic approach, namely PSSV, to identify somatic SVs from whole genome sequencing (WGS) data. PSSV features a mixture model with hidden states representing different mutation patterns; PSSV can thus differentiate heterozygous and homozygous SVs in each sample, enabling the identification of those somatic SVs with a heterozygous status in the normal sample and a homozygous status in the tumor sample. Simulation studies demonstrate that PSSV outperforms existing tools. PSSV has been successfully applied to breast cancer patient WGS data for identifying somatic SVs of key factors associated with breast cancer development. In this dissertation research, we demonstrate the advantage of the proposed distributional learning-based approaches over conventional methods for NGS data analysis. Distributional learning is a very powerful approach to gain biological insights from high quality NGS data. Successful applications of the proposed Bayesian methods to breast cancer NGS data shed light on underlying molecular mechanisms of breast cancer, enabling biologists or clinicians to identify major cancer drivers and develop new therapeutics for cancer treatment. Ph. D.

With the proliferation of fairly powerful mobile devices and ubiquitous wireless technology, traditional mobile ad hoc networks (MANETs) now migrate into a new era of service-oriented MANETs wherein a node can provide and receive service from other nodes it encounters and interacts with. This dissertation research concerns trust management and its applications for service-oriented MANETs to answer the challenges of MANET environments, including no centralized authority, dynamically changing topology, limited bandwidth and battery power, limited observations, unreliable communication, and the presence of malicious nodes who act to break the system functionality as well as selfish nodes who act to maximize their own gain. We propose a context-aware trust management model called CATrust for service-oriented ad hoc networks. The novelty of our design lies in the use of logit regression to dynamically estimate trustworthiness of a service provider based on its service behavior patterns in a context environment, treating channel conditions, node status, service payoff, and social disposition as 'context' information. We develop a recommendation filtering mechanism to effectively screen out false recommendations even in extremely hostile environments in which the majority recommenders are malicious. We demonstrate desirable convergence, accuracy, and resiliency properties of CATrust. We also demonstrate that CATrust outperforms contemporary peer-to-peer and Internet of Things trust models in terms of service trust prediction accuracy against collusion recommendation attacks. We validate the design of trust-based service management based on CATrust with a node-to-service composition and binding MANET application and a node-to-task assignment MANET application with multi-objective optimization (MOO) requirements. For either application, we propose a trust-based algorithm to effectively filter out malicious nodes exhibiting various attack behaviors by penalizing them with trust loss, which ultimately leads to high user satisfaction. Our trust-based algorithm is efficient with polynomial runtime complexity while achieving a close-to-optimal solution. We demonstrate that our trust-based algorithm built on CATrust outperforms a non-trust-based counterpart using blacklisting techniques and trust-based counterparts built on contemporary peer-to-peer trust protocols. We also develop a dynamic table-lookup method to apply the best trust model parameter settings upon detection of rapid MANET environment changes to maximize MOO performance. Ph. D.