Search

Complex dialog systems often use retrieved evidence to facilitate factual responses. Such RAG (Retrieval Augmented Generation) systems retrieve from massive heterogeneous data stores that are usually architected as multiple indexes or APIs instead of a single monolithic source. For a given query, relevant evidence needs to be retrieved from one or a small subset of possible retrieval sources. Complex queries can even require multi-step retrieval. For example, a conversational agent on a retail site answering customer questions about past orders will need to retrieve the appropriate customer order first and then the evidence relevant to the customer's question in the context of the ordered product. Most RAG Agents handle such Chain-of-Thought (CoT) tasks by interleaving reasoning and retrieval steps. However, each reasoning step directly adds to the latency of the system. For large models this latency cost is significant -- in the order of multiple seconds. Multi-agent systems may classify the query to a single Agent associated with a retrieval source, though this means that a (small) classification model dictates the performance of a large language model. In this work we present REAPER (REAsoning-based PlannER) - an LLM based planner to generate retrieval plans in conversational systems. We show significant gains in latency over Agent-based systems and are able to scale easily to new and unseen use cases as compared to classification-based planning. Though our method can be applied to any RAG system, we show our results in the context of a conversational shopping assistant.

Leveraging large-scale pre-training, vision foundational models showcase notable performance benefits. Recent segmentation algorithms for natural scenes have advanced significantly. However, existing models still struggle to automatically segment personalized instances in dense and crowded scenarios, where severe occlusions, scale variations, and background clutter pose a challenge to accurately delineate densely packed instances of the target object. To address this, we propose PerSense, an end-to-end, training-free, and model-agnostic one-shot framework for Personalized instance Segmentation in dense images. We develop a new baseline capable of automatically generating instance-level point prompts via proposing a novel Instance Detection Module (IDM) that leverages density maps, encapsulating spatial distribution of objects in an image. To mitigate false positives within generated point prompts, we design Point Prompt Selection Module (PPSM). Both IDM and PPSM transform density maps into personalized precise point prompts for instance-level segmentation and offer a seamless integration in our model-agnostic framework. We also introduce a feedback mechanism which enables PerSense to improve the accuracy of density maps by automating the exemplar selection process for density map generation. To promote algorithmic advances and effective tools for this relatively underexplored task, we introduce PerSense-D, a diverse dataset exclusive to personalized instance segmentation in dense images. Our extensive experiments establish PerSense superiority in dense scenarios by achieving an mIoU of 71.61% on PerSense-D, outperforming recent SOTA models by significant margins of +47.16%, +42.27%, +8.83%, and +5.69%. Additionally, our qualitative findings demonstrate the adaptability of our framework to images captured in-the-wild., Comment: Technical report of PerSense

Identifying the targets of hate speech is a crucial step in grasping the nature of such speech and, ultimately, in improving the detection of offensive posts on online forums. Much harmful content on online platforms uses implicit language especially when targeting vulnerable and protected groups such as using stereotypical characteristics instead of explicit target names, making it harder to detect and mitigate the language. In this study, we focus on identifying implied targets of hate speech, essential for recognizing subtler hate speech and enhancing the detection of harmful content on digital platforms. We define a new task aimed at identifying the targets even when they are not explicitly stated. To address that task, we collect and annotate target spans in three prominent implicit hate speech datasets: SBIC, DynaHate, and IHC. We call the resulting merged collection Implicit-Target-Span. The collection is achieved using an innovative pooling method with matching scores based on human annotations and Large Language Models (LLMs). Our experiments indicate that Implicit-Target-Span provides a challenging test bed for target span detection methods.

Recent research has shown that transformer networks can be used as differentiable search indexes by representing each document as a sequences of document ID tokens. These generative retrieval models cast the retrieval problem to a document ID generation problem for each given query. Despite their elegant design, existing generative retrieval models only perform well on artificially-constructed and small-scale collections. This has led to serious skepticism in the research community on their real-world impact. This paper represents an important milestone in generative retrieval research by showing, for the first time, that generative retrieval models can be trained to perform effectively on large-scale standard retrieval benchmarks. For doing so, we propose RIPOR- an optimization framework for generative retrieval that can be adopted by any encoder-decoder architecture. RIPOR is designed based on two often-overlooked fundamental design considerations in generative retrieval. First, given the sequential decoding nature of document ID generation, assigning accurate relevance scores to documents based on the whole document ID sequence is not sufficient. To address this issue, RIPOR introduces a novel prefix-oriented ranking optimization algorithm. Second, initial document IDs should be constructed based on relevance associations between queries and documents, instead of the syntactic and semantic information in the documents. RIPOR addresses this issue using a relevance-based document ID construction approach that quantizes relevance-based representations learned for documents. Evaluation on MSMARCO and TREC Deep Learning Track reveals that RIPOR surpasses state-of-the-art generative retrieval models by a large margin (e.g., 30.5% MRR improvements on MS MARCO Dev Set), and perform better on par with popular dense retrieval models.

Genome-wide association studies (GWAS) are used to identify relationships between genetic variations and specific traits. When applied to high-dimensional medical imaging data, a key step is to extract lower-dimensional, yet informative representations of the data as traits. Representation learning for imaging genetics is largely under-explored due to the unique challenges posed by GWAS in comparison to typical visual representation learning. In this study, we tackle this problem from the mutual information (MI) perspective by identifying key limitations of existing methods. We introduce a trans-modal learning framework Genetic InfoMax (GIM), including a regularized MI estimator and a novel genetics-informed transformer to address the specific challenges of GWAS. We evaluate GIM on human brain 3D MRI data and establish standardized evaluation protocols to compare it to existing approaches. Our results demonstrate the effectiveness of GIM and a significantly improved performance on GWAS., Comment: 17 pages, 7 figures

Augmented Reality (AR) based on Head-Mounted Displays (HMDs) has gained significant traction over the recent years. Nevertheless, it remains unclear what AR HMD-based applications have been developed over the years and what their system performance is when they are run on HMDs. In this paper, we aim to shed light into this direction. Our study focuses on the applications available on the Microsoft Hololens application store given the wide use of the Hololens headset. Our study has two major parts: (i) we collect metadata about the applications available on the Microsoft Hololens application store to understand their characteristics (e.g., categories, pricing, permissions requested, hardware and software compatibility); and (ii) we interact with these applications while running on a Hololens 2 headset and collect data about systems-related metrics (e.g., memory and storage usage, time spent on CPU and GPU related operations) to investigate the systems performance of applications. Our study has resulted in several interesting findings, which we share with the research community., Comment: This paper has been accepted for publication by IEEE ICC workshops 2023

Deep neural models for named entity recognition (NER) have shown impressive results in overcoming label scarcity and generalizing to unseen entities by leveraging distant supervision and auxiliary information such as explanations. However, the costs of acquiring such additional information are generally prohibitive. In this paper, we present a novel two-stage framework (AutoTriggER) to improve NER performance by automatically generating and leveraging ``entity triggers'' which are human-readable cues in the text that help guide the model to make better decisions. Our framework leverages post-hoc explanation to generate rationales and strengthens a model's prior knowledge using an embedding interpolation technique. This approach allows models to exploit triggers to infer entity boundaries and types instead of solely memorizing the entity words themselves. Through experiments on three well-studied NER datasets, AutoTriggER shows strong label-efficiency, is capable of generalizing to unseen entities, and outperforms the RoBERTa-CRF baseline by nearly 0.5 F1 points on average., Comment: 15 pages, 13 figures, EACL 2023

Pretrained contextualized representations offer great success for many downstream tasks, including document ranking. The multilingual versions of such pretrained representations provide a possibility of jointly learning many languages with the same model. Although it is expected to gain big with such joint training, in the case of cross lingual information retrieval (CLIR), the models under a multilingual setting are not achieving the same level of performance as those under a monolingual setting. We hypothesize that the performance drop is due to the translation gap between query and documents. In the monolingual retrieval task, because of the same lexical inputs, it is easier for model to identify the query terms that occurred in documents. However, in the multilingual pretrained models that the words in different languages are projected into the same hyperspace, the model tends to translate query terms into related terms, i.e., terms that appear in a similar context, in addition to or sometimes rather than synonyms in the target language. This property is creating difficulties for the model to connect terms that cooccur in both query and document. To address this issue, we propose a novel Mixed Attention Transformer (MAT) that incorporates external word level knowledge, such as a dictionary or translation table. We design a sandwich like architecture to embed MAT into the recent transformer based deep neural models. By encoding the translation knowledge into an attention matrix, the model with MAT is able to focus on the mutually translated words in the input sequence. Experimental results demonstrate the effectiveness of the external knowledge and the significant improvement of MAT embedded neural reranking model on CLIR task.

Online harassment in the form of hate speech has been on the rise in recent years. Addressing the issue requires a combination of content moderation by people, aided by automatic detection methods. As content moderation is itself harmful to the people doing it, we desire to reduce the burden by improving the automatic detection of hate speech. Hate speech presents a challenge as it is directed at different target groups using a completely different vocabulary. Further the authors of the hate speech are incentivized to disguise their behavior to avoid being removed from a platform. This makes it difficult to develop a comprehensive data set for training and evaluating hate speech detection models because the examples that represent one hate speech domain do not typically represent others, even within the same language or culture. We propose an unsupervised domain adaptation approach to augment labeled data for hate speech detection. We evaluate the approach with three different models (character CNNs, BiLSTMs and BERT) on three different collections. We show our approach improves Area under the Precision/Recall curve by as much as 42% and recall by as much as 278%, with no loss (and in some cases a significant gain) in precision.

Automated event extraction in social science applications often requires corpus-level evaluations: for example, aggregating text predictions across metadata and unbiased estimates of recall. We combine corpus-level evaluation requirements with a real-world, social science setting and introduce the IndiaPoliceEvents corpus--all 21,391 sentences from 1,257 English-language Times of India articles about events in the state of Gujarat during March 2002. Our trained annotators read and label every document for mentions of police activity events, allowing for unbiased recall evaluations. In contrast to other datasets with structured event representations, we gather annotations by posing natural questions, and evaluate off-the-shelf models for three different tasks: sentence classification, document ranking, and temporal aggregation of target events. We present baseline results from zero-shot BERT-based models fine-tuned on natural language inference and passage retrieval tasks. Our novel corpus-level evaluations and annotation approach can guide creation of similar social-science-oriented resources in the future., Comment: To appear in Findings of ACL 2021

We propose a novel framework for cross-lingual content flagging with limited target-language data, which significantly outperforms prior work in terms of predictive performance. The framework is based on a nearest-neighbour architecture. It is a modern instantiation of the vanilla k-nearest neighbour model, as we use Transformer representations in all its components. Our framework can adapt to new source-language instances, without the need to be retrained from scratch. Unlike prior work on neighbourhood-based approaches, we encode the neighbourhood information based on query--neighbour interactions. We propose two encoding schemes and we show their effectiveness using both qualitative and quantitative analysis. Our evaluation results on eight languages from two different datasets for abusive language detection show sizable improvements of up to 9.5 F1 points absolute (for Italian) over strong baselines. On average, we achieve 3.6 absolute F1 points of improvement for the three languages in the Jigsaw Multilingual dataset and 2.14 points for the WUL dataset., Comment: Accepted to appear in Transactions of the Association for Computational Linguistics (TACL) -- this is a pre-MIT Press publication version

The proliferation of harmful content on online platforms is a major societal problem, which comes in many different forms including hate speech, offensive language, bullying and harassment, misinformation, spam, violence, graphic content, sexual abuse, self harm, and many other. Online platforms seek to moderate such content to limit societal harm, to comply with legislation, and to create a more inclusive environment for their users. Researchers have developed different methods for automatically detecting harmful content, often focusing on specific sub-problems or on narrow communities, as what is considered harmful often depends on the platform and on the context. We argue that there is currently a dichotomy between what types of harmful content online platforms seek to curb, and what research efforts there are to automatically detect such content. We thus survey existing methods as well as content moderation policies by online platforms in this light and we suggest directions for future work., Comment: The paper has been accepted for publication to ACM Computing Surveys (CSUR)

The widely used Doyler-Fuller-Newman (DFN) model for lithium-ion batteries is too computationally expensive for certain applications, which has motivated the appearance of a plethora of simpler models. These models are usually posed in an ad hoc manner, leading to inconsistencies with the DFN model and to multiple formulations of the same model, with the Single Particle Model (SPM) being a very good example of the latter. In this work, we discuss the concept of SPM-type models showing that, despite the multiple formulations found in the literature, these models always follow the same structure, and we extend this discussion to models accounting for thermal effects. Then, we present a Thermal Single Particle Model with electrolyte (TSPMe) derived in a systematic manner using asymptotic techniques. The validation of the TSPMe against a thermal DFN model shows very high accuracy with a computational cost over forty times smaller. The comparison against experimental data shows that the model does a reasonable job predicting the behaviour of a real battery, but a very good parameter set is required to obtain accurate predictions., Comment: Accepted version to Electrochimca Acta, 27 pages, 13 pages of appendix, 12 figures

State-of-the-art neural dialogue systems excel at syntactic and semantic modelling of language, but often have a hard time establishing emotional alignment with the human interactant during a conversation. In this work, we bring Affect Control Theory (ACT), a socio-mathematical model of emotions for human-human interactions, to the neural dialogue generation setting. ACT makes predictions about how humans respond to emotional stimuli in social situations. Due to this property, ACT and its derivative probabilistic models have been successfully deployed in several applications of Human-Computer Interaction, including empathetic tutoring systems, assistive healthcare devices and two-person social dilemma games. We investigate how ACT can be used to develop affect-aware neural conversational agents, which produce emotionally aligned responses to prompts and take into consideration the affective identities of the interactants.

Service quality possesses the vital prominence in usability of innovative products and services. As technological innovation has made the life synchronized and effective, Internet of Things (IoT) is matter of discussion everywhere. From users' perspective, IoT services are always embraced by various system characteristics of security and performance. A service quality model can better present the preference of such technology customers. the study intends to project theoretical model of service quality for internet of things (IoT). Based on the existing models of service quality and the literature in internet of things, a framework is proposed to conceptualize and measure service quality for internet of things.This study established the IoT-Servqual model with four dimensions (i.e., Privacy, Functionality, Efficiency, and Tangibility) of multiple service quality models. These dimensions are essential and inclined towards the users' leaning of IoT Services. This paper contributes to research on internet of things services by development of a comprehensive framework for customers' quality apprehension. This model will previse the expression of information secrecy of users related with internet of things (IoT). This research will advance understanding of service quality in modern day technology and assist firms to devise the fruitful service structure., Comment: 11 pages, 01 figure, 01 table

A common approach for knowledge-base entity search is to consider an entity as a document with multiple fields. Models that focus on matching query terms in different fields are popular choices for searching such entity representations. An instance of such a model is FSDM (Fielded Sequential Dependence Model). We propose to integrate field-level semantic features into FSDM. We use FSDM to retrieve a pool of documents, and then to use semantic field-level features to re-rank those documents. We propose to represent queries as bags of terms as well as bags of entities, and eventually, use their dense vector representation to compute semantic features based on query document similarity. Our proposed re-ranking approach achieves significant improvement in entity retrieval on the DBpedia-Entity (v2) dataset over existing FSDM model. Specifically, for all queries we achieve 2.5% and 1.2% significant improvement in NDCG@10 and NDCG@100, respectively.

We describe a multi-task learning approach to train a Neural Machine Translation (NMT) model with a Relevance-based Auxiliary Task (RAT) for search query translation. The translation process for Cross-lingual Information Retrieval (CLIR) task is usually treated as a black box and it is performed as an independent step. However, an NMT model trained on sentence-level parallel data is not aware of the vocabulary distribution of the retrieval corpus. We address this problem with our multi-task learning architecture that achieves 16% improvement over a strong NMT baseline on Italian-English query-document dataset. We show using both quantitative and qualitative analysis that our model generates balanced and precise translations with the regularization effect it achieves from multi-task learning paradigm., Comment: Accepted for publication at ACL 2019

How non-spherical particles orient as they settle in a flow has important practical implications in a number of scientific and engineering problems. In a quiescent fluid, a slowly settling particle orients so that it settles with its broad side first. This is an effect of the torque due to convective inertia of the fluid set in motion by the settling particle, which maximises the drag experienced by the particle. Turbulent flows tend to randomise the particle orientation. Recently the settling of non-spherical particles in turbulence was analysed neglecting the effect of convective fluid inertia, but taking into account the effect of the turbulent fluid-velocity gradients on the particle orientation. These studies reached the opposite conclusion, namely that a rod settles preferentially with its tip first, wheras a disk settles with its edge first, therefore minimizing the drag on the particle. Here, we consider both effects, the convective inertial torque as well as the torque due to fluctuating velocity gradients, and ask under which circumstances either one or the other dominate. To this end we estimate the ratio of the magnitudes of the two torques. Our estimates suggest that the fluid-inertia torque prevails in high-Reynolds number flows. In this case non-spherical particles are expected to settle with a maximal drag. But when the Reynolds number is small then the torque due to fluid-velocity gradients may dominate, causing the particle to settle with its broad side first.

Microfluidic cytometers based on coulter principle have recently shown a great potential for point of care biosensors for medical diagnostics. In this study, the design of coulter based microfluidic cytometer is investigated by using electrical circuit simulations. We explore the effects of physical dimensions of the microelectrodes, the measurement volume, size/morphology of the targeted cells, electrical properties of the reagents in the measurement volume, and, the impedance of external readout circuit, on the sensitivity of the sensor. We show that the effect of microelectrode's surface area and the dielectric properties of the suspension medium should be carefully considered when characterizing the output response of the sensor. In particular, the area of microelectrodes can have significant effect on cells electrical opacity( the ratio of the cell impedance at high to low frequency) which is commonly used to distinguish between sub-population of the target cells( e.g. lymphocytes vs monocytes when counting white blood cells).Moreover, we highlight that the opacity response vs frequency can significantly vary depending upon whether the absolute cell impedance or the differential output impedance is used in the calculation. These insights can provide valuable guidelines for the design and characterization of coulter based microfluidic sensors., Comment: 18 pages,10 figures

Diabetic Retinopathy (DR) is a constantly deteriorating disease, being one of the leading causes of vision impairment and blindness. Subtle distinction among different grades and existence of many significant small features make the task of recognition very challenging. In addition, the present approach of retinopathy detection is a very laborious and time-intensive task, which heavily relies on the skill of a physician. Automated detection of diabetic retinopathy is essential to tackle these problems. Early-stage detection of diabetic retinopathy is also very important for diagnosis, which can prevent blindness with proper treatment. In this paper, we developed a novel deep convolutional neural network, which performs the early-stage detection by identifying all microaneurysms (MAs), the first signs of DR, along with correctly assigning labels to retinal fundus images which are graded into five categories. We have tested our network on the largest publicly available Kaggle diabetic retinopathy dataset, and achieved 0.851 quadratic weighted kappa score and 0.844 AUC score, which achieves the state-of-the-art performance on severity grading. In the early-stage detection, we have achieved a sensitivity of 98% and specificity of above 94%, which demonstrates the effectiveness of our proposed method. Our proposed architecture is at the same time very simple and efficient with respect to computational time and space are concerned., Comment: Accepted in MIND 2019

The present gap between the amount of available protein sequence due to the development of next generation sequencing technology (NGS) and slow and expensive experimental extraction of useful information like annotation of protein sequence in different functional aspects, is ever widening, which can be reduced by employing automatic function prediction (AFP) approaches. Gene Ontology (GO), comprising of more than 40, 000 classes, defines three aspects of protein function names Biological Process (BP), Cellular Component (CC), Molecular Function (MF). Multiple functions of a single protein, has made automatic function prediction a large-scale, multi-class, multi-label task. In this paper, we present DEEPGONET, a novel cascaded convolutional and recurrent neural network, to predict the top-level hierarchy of GO ontology. The network takes the primary sequence of protein as input which makes it more useful than other prevailing state-of-the-art deep learning based methods with multi-modal input, making them less applicable for proteins where only primary sequence is available. All predictions of different protein functions of our network are performed by the same architecture, a proof of better generalization as demonstrated by promising performance on a variety of organisms while trained on Homo sapiens only, which is made possible by efficient exploration of vast output space by leveraging hierarchical relationship among GO classes. The promising performance of our model makes it a potential avenue for directing experimental protein functions exploration efficiently by vastly eliminating possible routes which is done by the exploring only the suggested routes from our model. Our proposed model is also very simple and efficient in terms of computational time and space compared to other architectures in literature., Comment: Accepted in ICCIT 2018

Traditional information retrieval treats named entity recognition as a pre-indexing corpus annotation task, allowing entity tags to be indexed and used during search. Named entity taggers themselves are typically trained on thousands or tens of thousands of examples labeled by humans. However, there is a long tail of named entities classes, and for these cases, labeled data may be impossible to find or justify financially. We propose exploring named entity recognition as a search task, where the named entity class of interest is a query, and entities of that class are the relevant "documents". What should that query look like? Can we even perform NER-style labeling with tens of labels? This study presents an exploration of CRF-based NER models with handcrafted features and of how we might transform them into search queries., Comment: 6 pages

We address the role of a user in Contextual Named Entity Retrieval (CNER), showing (1) that user identification of important context-bearing terms is superior to automated approaches, and (2) that further gains are possible if the user indicates the relative importance of those terms. CNER is similar in spirit to List Question answering and Entity disambiguation. However, the main focus of CNER is to obtain user feedback for constructing a profile for a class of entities on the fly and use that to retrieve entities from free text. Given a sentence, and an entity selected from that sentence, CNER aims to retrieve sentences that have entities similar to query entity. This paper explores obtaining term relevance feedback and importance weighting from humans in order to improve a CNER system. We report our findings based on the efforts of IR researchers as well as crowdsourced workers.

In this paper we describe our machine learning solution for the RecSys Challenge, 2015. We have proposed a time efficient two-stage cascaded classifier for the prediction of buy sessions and purchased items within such sessions. Based on the model, several interesting features found, and formation of our own test bed, we have achieved a reasonable score. Usage of Random Forests helps us to cope with the effect of the multiplicity of good models depending on varying subsets of features in the purchased items prediction and, in its turn, boosting is used as a suitable technique to overcome severe class imbalance of the buy-session prediction.

Mobile operators commonly use macro cells with traditional wide beam antennas for wider coverage in the cell, but future capacity demands cannot be achieved by using them only. It is required to achieve maximum practical capacity from macro cells by employing higher order sectorization and by utilizing all possible antenna solutions including smart antennas. This paper presents enhanced tessellation for 6-sector sites and proposes novel layout for 12-sector sites. The main target of this paper is to compare the performance of conventional wide beam antenna, switched beam smart antenna, adaptive beam antenna and different network layouts in terms of offering better received signal quality and user throughput. Splitting macro cell into smaller micro or pico cells can improve the capacity of network, but this paper highlights the importance of higher order sectorization and advance antenna techniques to attain high Signal to Interference plus Noise Ratio (SINR), along with improved network capacity. Monte Carlo simulations at system level were done for Dual Cell High Speed Downlink Packet Access (DC-HSDPA) technology with multiple (five) users per Transmission Time Interval (TTI) at different Intersite Distance (ISD). The obtained results validate and estimate the gain of using smart antennas and higher order sectorization with proposed network layout., Comment: 20 pages, 18 Figures, and 5 Tables

Management of project planning, monitoring, scheduling, estimation and risk management are critical issues faced by a project manager during development life cycle of software. In RUP, project management is considered as core discipline whose activities are carried in all phases during development of software products. On other side service monitoring is considered as best practice of SOA which leads to availability, auditing, debugging and tracing process. In this paper, authors define a strategy to incorporate the service monitoring of SOA into RUP to improve the artifacts of project management activities. Moreover, the authors define the rules to implement the features of service monitoring, which help the project manager to carry on activities in well define manner. Proposed frame work is implemented on RB (Resuming Bank) application and obtained improved results on PM (Project Management) work.

SOA (Service Oriented Architecture) is a new trend towards increasing the profit margins in an organization due to incorporating business services to business practices. Rational Unified Process (RUP) is a unified method planning form for large business applications that provides a language for describing method content and processes. The well defined mapping of SOA and RUP leads to successful completion of RUP software projects to provide services to their users. DOA (Digital Office Assistant) is a multi user SOA type application that provides appropriate viewer for each user to assist him through services. In this paper authors proposed the mapping strategy of SOA with RUP by considering DOA as case study., Comment: Journal of Computing, Vol. 2, Issue 1, January 2010, https://sites.google.com/site/journalofcomputing/