Search

Edge Computing (EC) became popular again with the rise of IoT, Cloud Computing, and Industry 4.0. In this paper, difficulties of application development in the EC environment are discussed and a container-based solution using remote debugging at the edge is proposed. This container allows application developers to write code in the production environment. Our implementation increases the development speed and facilitates in-place debugging for EC environments.

It is necessary to develop accurate, yet simple and efficient models that can be used with high-speed industrial data streams. In this paper, we develop a mode identification technique using stream analytics and show that it may be more effective than batch models, especially for time-varying systems. These industrial systems continuously monitor hundreds of sensors, but the relationships among variables change over time, which are identified as different operational modes. To detect drifts among modes, predictive modeling techniques such as regression analysis, K-means and DBSCAN clustering are used over sensor data streams from an oil refinery and models are updated in real-time using window-based analysis. Finally, an adaptive window size tuning approach based on the TCP congestion control algorithm is discussed, which reduces model update costs as well as prediction errors.

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Internet of Things (IoT) is a new computing paradigm that is proliferated by wide adoption of application level protocols such as MQTT and CoAP, each of which defines different styles of sequential interaction of events. Even though there is a considerable effort in the literature for verification of such complex and distributed systems, a practical solution for IoT systems that supports runtime system verification is still missing. In this paper, we present a runtime monitoring approach for IoT systems that exploits event relations expressed in terms of sequential interaction messaging model of Constrained Application Protocol (CoAP). We propose the use of Complex-Event Processing (CEP) to detect failures at runtime by exploiting complex event patterns defined via predetermined event algebra. We further present a simple case scenario to demonstrate the applicability of the approach on Wireless Token Ring Protocol execution. TÜBiTAK BiDEB

Due to copyright restrictions, the access to the full text of this article is only available via subscription. To extract useful and actionable information in real-time, the information technology (IT) world is coping with big data problems today. In this paper, we present implementation details and performance results of ReCEPtor, our system for "online" Association Rule Mining (ARM) over big and fast data streams. Specifically, we added Apriori and two different FP-Growth algorithms inside Esper Complex Event Processing (CEP) engine and compared their performances using LastFM social music site data. Our most important findings show that online ARM can generate (1) more unique rules, (2) with higher throughput, and (3) much sooner (lower latency) than offline rule mining. In addition, we have found many interesting and realistic musical preference rules such as "George HarrisonàBeatles". We demonstrate a sustained rate of ~15K rows/sec per core. We hope that our findings can shed light on the design and implementation of other fast data analytics systems in the future. European Union ; TÜBİTAK ; IBM Shared University Research program ; Turkish Telecomm ; Avea Labs

Today, the IT world is trying to cope with “big data” problems (data volume, velocity, variety, veracity) on the path to obtaining useful information. In this paper, we present implementation details and performance results of realizing “online” Association Rule Mining (ARM) over big data streams for the first time in the literature. Specifically, we added Apriori and FP-Growth algorithms for stream mining inside an event processing engine, called Esper. Using the system, these two algorithms were compared over LastFM social music site data and by using tumbling windows. The better-performing FP-Growth was selected and used in creation of a real-time rule-based recommendation engine. Our most important findings show that online association rule mining can generate (1) more rules, (2) much faster and more efficiently, and (3) much sooner than offline rule mining. In addition, we have found many interesting and realistic musical preference rules such as “George Harrison⇒Beatles”. We hope that our findings can shed light on the design and implementation of other big data analytics systems in the future.

Interpolative decomposition (ID) is a matrix factorization which aims to represent the data matrix via a subset of its own columns. These selected columns are supposed to hold the salient features expressing the data. A very common ID approach in the literature is based on importance sampling where a statistical leverage score is computed for each column and K columns are randomly selected using these scores. These randomized methods aim a better low-rank approximation of the matrix by seeking for the columns that express the range of the matrix the best. This makes ID a good alternative to Singular Value Decomposition (SVD) since it favors sparsity and the bases correspond to real data points. However, the columns leading to the best low-rank approximation are usually not the ones in terms of representativeness if the underlying data is composed of several clusters which is very common in real life. In this paper, we introduce an alternative ID approach based on clustering. We employ K-medoids to be employed as an ID method for better interpretability and respresentativeness. We apply ID on handwritten digit recognition and supply comparative results of the proposed approach to the state-of-the-art method in the literature. We show its superiority in terms of representativeness of the data. We demonstrate that most of the data can be discarded without compromising the accuracy.

Interpolative decomposition (ID) is a low-rank matrix decomposition where the data matrix is expressed via a sub-set of its own columns. In this work, we propose a novel probabilistic method for ID where it is expressed as a statistical model within a Bayesian framework. The proposed method considerably differs from other ID methods in the literature: It handles the model selection automatically and enables the construction of problem-specific interpolative decompositions. We derive the analytical solution for the normal distribution and we provide a numerical solution for the generic case. Simulation results on synthetic data are provided to illustrate that the method converges to the true decomposition, independent of the initialization; and it can successfully handle noise. In addition, we apply probabilistic ID to the problem of automatic polyphonic music transcription to extract important information from a huge dictionary of spectrum instances. We supply comparative results with the other proposed techniques in the literature and show that it performs better. Probabilistic interpolative decomposition serves as a promising feature selection and de-noising tool to be exploited in big data problems.

Omnidirectional cameras have a lot of potential for surveillance and ambient intelligence applications, since they provide increased coverage with fewer cameras. We introduce the new BOMNI dataset, collected with two omnidirectional cameras simultaneously. The dataset contains single subject and multi-subject interaction scenarios, as well as actions relevant for ambient assisted living, such as falling down. We describe evaluation protocols on this dataset, and provide benchmarking baseline results for two tracking systems based on bounding box and interest point matching after foreground-background segmentation, respectively.

The aim of this work is to perform polyphonic music transcription in an efficient way. The problem is formulated as a linear model and the speed is improved by a randomized SVD-based method. The method is shown to compete with the best resulting approaches in literature. The conventional methods seem to fail in this era of big data whereas the proposed method efficiently handles this by use of randomized algorithms for matrix decompositions. The method is able to improve time and space complexity without compromising the high success rate.

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Due to prevalent use of sensors and network monitoring tools, big volumes of data or “big data” today traverse the enterprise data processing pipelines in a streaming fashion. While some companies prefer to deploy their data processing infrastructures and services as private clouds, others completely outsource these services to public clouds. In either case, attempting to store the data first for subsequent analysis creates additional resource costs and unwanted delays in obtaining actionable information. As a result, enterprises increasingly employ data or event stream processing systems and further want to extend them with complex online analytic and mining capabilities. In this paper, we present implementation details for doing both correlation analysis and association rule mining (ARM) over streams. Specifically, we implement Pearson-Product Moment Correlation for analytics and Apriori & FPGrowth algorithms for stream mining inside a popular event stream processing engine called Esper. As a unique contribution, we conduct experiments and present performance results of these new tools with different tumbling and sliding time-windows over two different stream types: one for moving bus trajectories and another for web logs from a music site. We find that while tumbling windows may be more preferable for performance in certain applications, sliding windows can provide additional benefits with rule mining. We hope that our findings can shed light on the design of other cloud analytics systems. Avea Labs ; TÜBİTAK ; European Commission ; IBM Shared University Research Program

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Clustering is a common technique, in all areas where information is obtained from the collected data. In this work, three well-known clustering algorithms namely, K-means, Spectral and DBSCAN are investigated in terms of their validity using four clustering validity indexes, Rand, Adjusted Rand, Jaccard, Silhouette. These clustering algorithms are applied on three data sets which have different characteristics. Thus steps have been taken for an automated clustering optimization system. TÜBİTAK ; European Commission

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Event correlation engines help us find events of interest inside raw sensor data streams and help reduce the data volume, simultaneously. This paper discusses some of the challenges faced in finding event correlations over federated wireless sensor networks (WSNs) including high data volumes, uncertain or missing data, application-specific dependencies and widely varying data ranges and sampling frequencies. Analysisover real geo-tracking data of moving objects confirms some of these challenges. Federation at the data layer above the WSNs is presented as a feasible alternative. TÜBİTAK ; IBM Shared University Research program ; European Commission

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Complex Event Processing (CEP) has become increasingly important for tracking and monitoring complex event anomalies and trends in event streams emitted from business processes such as supply chain management to online stores in e-commerce. These monitoring applications submit complex event queries to track sequences of events that match a given pattern. The state-of-the-art CEP systems mostly focus on the execution of flat sequence queries, we instead support the execution of nested CEP queries specified by our NEsted Event Language NEEL. However, the iterative execution of nested CEP expressions often results in the repeated recomputation of the same or similar results for nested subexpressions as the window slides over the event stream. In this work we thus propose to optimize NEEL execution performance by caching intermediate results. In particular we design two methods of applying selective caching of intermediate results namely Object Caching and the Interval-Driven Semantic Caching. Techniques for incrementally loading, purging and exploiting the cache content are described. Our experimental study using real-world stock trades evaluates the performance of our proposed caching strategies for different query types. HP Labs Innovation Research Program ; NSF ; TÜBİTAK

Automatic recognition of emotional states is a challenging problem which takes the attention of researchers studying in many areas such as human-computer interaction, verbal-nonverbal communication, data-driven facial animation and analysis of the faces of autistic children. In this work, emotion recognition problem is investigated; a novel approach is proposed that works in real-time and recognizes surprise, anger, happiness, sadness, fear, disgust and neutral expression. In the proposed system, facial landmarks are tracked using the tracker based on the multi-resolution active shape models. Afterwards, the classification is done using the differences among the feature vectors belonging to emotional expressions to the feature vector involving high level features extracted from the locations of the landmarks and the gradients of specific facial regions. In experiments conducted on 5 subjects, success rate of 75.23% is achieved for the seven classes mentioned whereas a success rate of 100% is found when four classes (surprise, anger, happiness and neutral expression) are used. The system gives promising results also on partially occluded faces. The implementation works in real-time and includes modules for person and environment adaptation.

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Complex event processing (CEP) over event streams has become increasingly important for real-time applications ranging from health care, supply chain management to business intelligence. These monitoring applications submit complex queries to track sequences of events that match a given pattern. As these systems mature the need for increasingly complex nested sequence query support arises, while the state-of-art CEP systems mostly support the execution of flat sequence queries only. To assure real-time responsiveness and scalability for pattern detection even on huge volume high-speed streams, efficient processing techniques must be designed. In this paper, we first analyze the prevailing nested pattern query processing strategy and identify several serious shortcomings. Not only are substantial subsequences first constructed just to be subsequently discarded, but also opportunities for shared execution of nested subexpressions are overlooked. As foundation, we introduce NEEL, a CEP query language for expressing nested CEP pattern queries composed of sequence, negation, AND and OR operators. To overcome deficiencies, we design rewriting rules for pushing negation into inner subexpressions. Next, we devise a normalization procedure that employs these rules for flattening a nested complex event expression. To conserve CPU and memory consumption, we propose several strategies for efficient shared processing of groups of normalized NEEL subexpressions. These strategies include prefix caching, suffix clustering and customized “bit-marking” execution strategies. We design an optimizer to partition the set of all CEP subexpressions in a NEEL normal form into groups, each of which can then be mapped to one of our shared execution operators. Lastly, we evaluate our technologies by conducting a performance study to assess the CPU processing time using real-world stock trades data. Our results confirm that our NEEL execution in many cases performs 100 fold fast er than the traditional iterative nested execution strategy for real stock market query workloads. HP Labs Innovation Research Program ; NSF ; TÜBİTAK

Automatic recognition of facial expressions and head gestures plays an important role in a wide range of research area including sign language recognition and human-computer interaction. In this work, we adopt the well-performing self-similarity based action recognition method to classify facial expressions and head gestures. Additionally, we propose a novel approach for facial gesture recognition based on the histogram of tracked facial landmarks. We fuse the presented techniques with our previous Hidden Markov Model based approach [1] and get 15% increase in classification results.

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Many modern applications including tag based mass transit systems, RFID-based supply chain management systems and online financial feeds require special purpose event stream processing technology to analyze vast amounts of sequential multi-dimensional data available in real-time data feeds. Traditional online analytical processing (OLAP) systems are not designed for real-time pattern-based operations, while Complex Event Processing (CEP) systems are designed for sequence detection and do not support OLAP operations. We will demonstrate a novel E-Cube model that combines CEP and OLAP techniques for multi-dimensional event pattern analysis at different abstraction levels. A London transit scenario will be given to demonstrate the utility and performance of this proposed technology. NSF

The Mobile Shopping Assistant (MSA) is a mobile application platform to deliver real-time, in-store, and personalized services, such as personalized product offerings and in-store customer advisory support, to improve the shopping experiences of in-store customers. The service delivery network that powers the MSA involves retail stores and their business partners such as manufacturers. This paper presents the core technologies that we developed in this cross-organizational service network to support the MSA and its personalized services, with focus on service delivery, customer behavior understanding and information sharing. Our event-based techniques allow customers, stores and manufacturers to deliver and consume the services in a loosely coupled manner, thus solving a critical store-specific real-time engagement problem in a seamless way. Service response tracking enables the stores to construct a comprehensive view of a customer’s in-store shopping behavior. Finally, the cross-organizational authorization-based access control mechanism effectively enforces information sharing between the stores and their partners.

In this paper, we describe the design of our architecture for Continuous, Heterogeneous Analysis Over Streams, aka CHAOS that combines stream processing, approximation techniques, mining, complex event processing and visualization. CHAOS, with the novel concept of Computational Stream Analysis Cube, provides an effective, scalable platform for near real time processing of business and enterprise streams. We describe our approach with a real data center temperature analysis application.

Facial expressions play a vital role in sign language by changing the meaning of the performed sign. In this work, we propose a system composed of a facial feature tracker based on Active Shape Models and a classifier based on Hidden Markov Models to recognize common facial expressions used in sign language. Face tracker works in multi-resolution and multi-view to track faces in different poses fast and effectively. Detailed tests are prepared to report the accuracy of the tracker and classifier, and it is seen that the system works in high accuracy and robustly.

Expressions carry vital information in sign language. In this study, we have implemented a multi-resolution active shape model (MR-ASM) tracker, which tracks 116 facial landmarks on videos. Since the expressions involve significant amount of head rotation, we employ multiple ASM models to deal with different poses. The tracked landmark points are used to extract motion features which are used by a support vector machine (SVM) based classifier. We obtained above 90% classification accuracy in a data set containing 7 expressions.

Sign languages are visual languages. The message is not only transferred via hand gestures (manual signs) but also head/body motion and facial expressions (non-manual signs). In this article, we present a database of non-manual signs in Turkish sign language (TSL). There are eight non-manual signs in the database, which are frequently used in TSL. The database contains the videos of these signs as well as a ground truth data of 60 manually landmarked points of the face.

The storage traffic for different logical units (LUs) of a disk array converge at the array's cache. The cache is allocated among the LUs approximately according to their relative I/O rates. In the case of nonuniform I/O rates and sensitivity to storage response times between differing applications in a storage area network (SAN), undesirable cache interference between LUs can result in unacceptable storage performance for some LUs. This paper describes SANBoost, a SAN-level caching approach that can be enabled selectively on a per-LU basis to provide a performance isolation mechanism for response time metrics related to storage quality of service (QoS). SANBoost automates hot data detection and migration processes in block-level storage. The design consists of a migration module implemented in a fabric-based SAN virtualization appliance and a solid-state disk (SSD) that acts as a cache resource within the same SAN. Simulation results quantify the impact of a specific static SANBoost caching policy on the SPC-1 benchmark workload and address the relative impact of adapting a threshold in the placement algorithm.