Search

This study investigates the dynamics of pedestrian crossing flows with varying crossing angles $\alpha$ to classify different scenarios and derive implications for crowd management. Probability density functions of four key features$-$velocity $v$, density $\rho$, avoidance number $Av$, and intrusion number $In$$-$were analyzed to characterize pedestrian behavior. Velocity-density fundamental diagrams were constructed for each $\alpha$ and fitted with functional forms from existing literature. Classification attempts using $Av$-$In$ and $v$-$\rho$ phase spaces revealed significant overlaps, highlighting the limitations of these metrics alone for scenario differentiation. To address this, machine learning models, including logistic regression and random forest, were employed using all four features. Results showed robust classification performance, with $v$ and $Av$ contributing most significantly. Insights from feature importance metrics and classification accuracy offer practical guidance for managing high-density crowds, optimizing pedestrian flow, and designing safer public spaces. These findings provide a data-driven framework for advancing pedestrian dynamics research., Comment: 25 pages, 9 figures, 5 tables

In a setting where segmentation models have to be built for multiple datasets, each with its own corresponding label set, a straightforward way is to learn one model for every dataset and its labels. Alternatively, multi-task architectures with shared encoders and multiple segmentation heads or shared weights with compound labels can also be made use of. This work proposes a novel label sharing framework where a shared common label space is constructed and each of the individual label sets are systematically mapped to the common labels. This transforms multiple datasets with disparate label sets into a single large dataset with shared labels, and therefore all the segmentation tasks can be addressed by learning a single model. This eliminates the need for task specific adaptations in network architectures and also results in parameter and data efficient models. Furthermore, label sharing framework is naturally amenable for incremental learning where segmentations for new datasets can be easily learnt. We experimentally validate our method on various medical image segmentation datasets, each involving multi-label segmentation. Furthermore, we demonstrate the efficacy of the proposed method in terms of performance and incremental learning ability vis-a-vis alternative methods.

In task-oriented dialogue systems, intent detection is crucial for interpreting user queries and providing appropriate responses. Existing research primarily addresses simple queries with a single intent, lacking effective systems for handling complex queries with multiple intents and extracting different intent spans. Additionally, there is a notable absence of multilingual, multi-intent datasets. This study addresses three critical tasks: extracting multiple intent spans from queries, detecting multiple intents, and developing a multi-lingual multi-label intent dataset. We introduce a novel multi-label multi-class intent detection dataset (MLMCID-dataset) curated from existing benchmark datasets. We also propose a pointer network-based architecture (MLMCID) to extract intent spans and detect multiple intents with coarse and fine-grained labels in the form of sextuplets. Comprehensive analysis demonstrates the superiority of our pointer network-based system over baseline approaches in terms of accuracy and F1-score across various datasets., Comment: Accepted at EMNLP 2024 Findings (Long Paper)

We investigate a dynamical model of opinion formation in which an individual's opinion is influenced by interactions with a group of other agents. We introduce a bias towards one of the opinions in a manner not considered earlier to the best of our knowledge. When the bias is neutral, the model is reduced to a mean-field voter model. We analyze the behavior and steady states of the system, identifying three distinct regimes based on the bias level: one favoring negative opinions, one favoring positive opinions, and a neutral case. In large systems, the equilibrium properties become independent of the size of the group, indicating that only the bias influences the final outcome. However, for small groups, the time to reach equilibrium depends on the size of the group. Our results show that even a small initial bias leads to a consensus where all agents eventually share the same opinion when the bias is not neutral. The system exhibits universal behavior, with critical slowing down occurring near the neutral bias point, marking it as a critical dynamical threshold. The time required to reach consensus scales logarithmically when the bias is non-neutral and linearly when it is neutral. Although short-term dynamics depends on group size for small groups, long-term behavior is governed solely by the bias., Comment: 34 pages, 10 figures

The ever-increasing volume of digital information necessitates efficient methods for users to extract key insights from lengthy documents. Aspect-based summarization offers a targeted approach, generating summaries focused on specific aspects within a document. Despite advancements in aspect-based summarization research, there is a continuous quest for improved model performance. Given that large language models (LLMs) have demonstrated the potential to revolutionize diverse tasks within natural language processing, particularly in the problem of summarization, this paper explores the potential of fine-tuning LLMs for the aspect-based summarization task. We evaluate the impact of fine-tuning open-source foundation LLMs, including Llama2, Mistral, Gemma and Aya, on a publicly available domain-specific aspect based summary dataset. We hypothesize that this approach will enable these models to effectively identify and extract aspect-related information, leading to superior quality aspect-based summaries compared to the state-of-the-art. We establish a comprehensive evaluation framework to compare the performance of fine-tuned LLMs against competing aspect-based summarization methods and vanilla counterparts of the fine-tuned LLMs. Our work contributes to the field of aspect-based summarization by demonstrating the efficacy of fine-tuning LLMs for generating high-quality aspect-based summaries. Furthermore, it opens doors for further exploration of using LLMs for targeted information extraction tasks across various NLP domains.

For pedestrians moving without spatial constraints, extensive research has been devoted to develop methods of density estimation. In this paper we present a new approach based on Voronoi cells, offering a means to estimate density for individuals in small, unbounded pedestrian groups. A thorough evaluation of existing methods, encompassing both Lagrangian and Eulerian approaches employed in similar contexts, reveals notable limitations. Specifically, these methods turn out to be ill-defined for realistic density estimation along a pedestrian's trajectory, exhibiting systematic biases and fluctuations that depend on the choice of parameters. There is thus a need for a parameter-independent method to eliminate this bias. We propose a modification of the widely used Voronoi-cell based density estimate to accommodate pedestrian groups, irrespective of their size. The advantages of this modified Voronoi method are that it is an instantaneous method that requires only knowledge of the pedestrians' positions at a give time, does not depend on the choice of parameter values, gives us a realistic estimate of density in an individual's neighborhood, and has appropriate physical meaning for both small and large human crowds in a wide variety of situations. We conclude with general remarks about the meaning of density measurements for small groups of pedestrians., Comment: 27 pages, 10 figures. arXiv admin note: text overlap with arXiv:2212.01338

In the dynamic realm of One Day International (ODI) cricket, the sport has undergone significant transformations over the past four decades. This study digs into the intricate evolution of ODI cricket from 1987 to 2023, analyzing about 4000 matches to uncover pivotal performance indicators such as batting prowess, bowling efficiency, and partnership dynamics. Employing statistical methodologies, including Cohen's effect size, we scrutinize the observed changes that have shaped ODI cricket's landscape. Our findings reveal nuanced trends: while first innings scores have shown stability with sporadic high outliers in recent years, the impact of achieving scores exceeding 300 has notably increased. Furthermore, batting depth and early wickets lost in the first innings continue to significantly influence match outcomes, highlighting strategic shifts in team approaches. We also observe improvements in second innings bowling effectiveness, particularly in wicket-taking ability, underscoring evolving defensive strategies. This research contributes a statistical foundation to comprehensively understand the evolving dynamics of ODI cricket, offering insights crucial for strategic decision-making and further analysis in sports analytics., Comment: 32 pages, 10 figures

In an ever-expanding world of domain-specific knowledge, the increasing complexity of consuming, and storing information necessitates the generation of summaries from large information repositories. However, every persona of a domain has different requirements of information and hence their summarization. For example, in the healthcare domain, a persona-based (such as Doctor, Nurse, Patient etc.) approach is imperative to deliver targeted medical information efficiently. Persona-based summarization of domain-specific information by humans is a high cognitive load task and is generally not preferred. The summaries generated by two different humans have high variability and do not scale in cost and subject matter expertise as domains and personas grow. Further, AI-generated summaries using generic Large Language Models (LLMs) may not necessarily offer satisfactory accuracy for different domains unless they have been specifically trained on domain-specific data and can also be very expensive to use in day-to-day operations. Our contribution in this paper is two-fold: 1) We present an approach to efficiently fine-tune a domain-specific small foundation LLM using a healthcare corpus and also show that we can effectively evaluate the summarization quality using AI-based critiquing. 2) We further show that AI-based critiquing has good concordance with Human-based critiquing of the summaries. Hence, such AI-based pipelines to generate domain-specific persona-based summaries can be easily scaled to other domains such as legal, enterprise documents, education etc. in a very efficient and cost-effective manner.

Quantifying cyber risks is essential for organizations to grasp their vulnerability to threats and make informed decisions. However, current approaches still need to work on blending economic viewpoints to provide insightful analysis. To bridge this gap, we introduce QBER approach to offer decision-makers measurable risk metrics. The QBER evaluates losses from cyberattacks, performs detailed risk analyses based on existing cybersecurity measures, and provides thorough cost assessments. Our contributions involve outlining cyberattack probabilities and risks, identifying Technical, Economic, and Legal (TEL) impacts, creating a model to gauge impacts, suggesting risk mitigation strategies, and examining trends and challenges in implementing widespread Cyber Risk Quantification (CRQ). The QBER approach serves as a guided approach for organizations to assess risks and strategically invest in cybersecurity., Comment: 10 pages, 9 equations, 3 tables, 2 figures

The escalating process of urbanization has raised concerns about incidents arising from overcrowding, necessitating a deep understanding of large human crowd behavior and the development of effective crowd management strategies. This study employs computational methods to analyze real-world crowd behaviors, emphasizing self-organizing patterns. Notably, the intersection of two streams of individuals triggers the spontaneous emergence of striped patterns, validated through both simulations and live human experiments. Addressing a gap in computational methods for studying these patterns, previous research utilized the pattern-matching technique, employing the Nelder-Mead Simplex algorithm for fitting a two-dimensional sinusoidal function to pedestrian coordinates. This paper advances the pattern-matching procedure by introducing Simulated Annealing as the optimization algorithm and employing a two-dimensional square wave for data fitting. The amalgamation of Simulated Annealing and the square wave significantly enhances pattern fitting quality, validated through statistical hypothesis tests. The study concludes by outlining potential applications of this method across diverse scenarios., Comment: 12 pages, 7 figures. Accepted for publication in Journal of Mathematics in Industry

Super-Resolution (SR) is a time-hallowed image processing problem that aims to improve the quality of a Low-Resolution (LR) sample up to the standard of its High-Resolution (HR) counterpart. We aim to address this by introducing Super-Resolution Generator (SuRGe), a fully-convolutional Generative Adversarial Network (GAN)-based architecture for SR. We show that distinct convolutional features obtained at increasing depths of a GAN generator can be optimally combined by a set of learnable convex weights to improve the quality of generated SR samples. In the process, we employ the Jensen-Shannon and the Gromov-Wasserstein losses respectively between the SR-HR and LR-SR pairs of distributions to further aid the generator of SuRGe to better exploit the available information in an attempt to improve SR. Moreover, we train the discriminator of SuRGe with the Wasserstein loss with gradient penalty, to primarily prevent mode collapse. The proposed SuRGe, as an end-to-end GAN workflow tailor-made for super-resolution, offers improved performance while maintaining low inference time. The efficacy of SuRGe is substantiated by its superior performance compared to 18 state-of-the-art contenders on 10 benchmark datasets.

Biomedical queries have become increasingly prevalent in web searches, reflecting the growing interest in accessing biomedical literature. Despite recent research on large-language models (LLMs) motivated by endeavours to attain generalized intelligence, their efficacy in replacing task and domain-specific natural language understanding approaches remains questionable. In this paper, we address this question by conducting a comprehensive empirical evaluation of intent detection and named entity recognition (NER) tasks from biomedical text. We show that Supervised Fine Tuned approaches are still relevant and more effective than general-purpose LLMs. Biomedical transformer models such as PubMedBERT can surpass ChatGPT on NER task with only 5 supervised examples., Comment: Accepted to CL4Health LREC-COLING 2024

Dialog summarization has become increasingly important in managing and comprehending large-scale conversations across various domains. This task presents unique challenges in capturing the key points, context, and nuances of multi-turn long conversations for summarization. It is worth noting that the summarization techniques may vary based on specific requirements such as in a shopping-chatbot scenario, the dialog summary helps to learn user preferences, whereas in the case of a customer call center, the summary may involve the problem attributes that a user specified, and the final resolution provided. This work emphasizes the significance of creating coherent and contextually rich summaries for effective communication in various applications. We explore current state-of-the-art approaches for long dialog summarization in different domains and benchmark metrics based evaluations show that one single model does not perform well across various areas for distinct summarization tasks.

Choosing the right person for the right job makes the personnel interview process a cognitively demanding task. Psychometric tests, followed by an interview, have often been used to aid the process although such mechanisms have their limitations. While psychometric tests suffer from faking or social desirability of responses, the interview process depends on the way the responses are analyzed by the interviewers. We propose the use of behaviometry as an assistive tool to facilitate an objective assessment of the interviewee without increasing the cognitive load of the interviewer. Behaviometry is a relatively little explored field of study in the selection process, that utilizes inimitable behavioral characteristics like facial expressions, vocalization patterns, pupillary reactions, proximal behavior, body language, etc. The method analyzes thin slices of behavior and provides unbiased information about the interviewee. The current study proposes the methodology behind this tool to capture facial expressions, in terms of facial asymmetry and micro-expressions. Hemi-facial composites using a structural similarity index was used to develop a progressive time graph of facial asymmetry, as a test case. A frame-by-frame analysis was performed on three YouTube video samples, where Structural similarity index (SSID) scores of 75% and more showed behavioral congruence. The research utilizes open-source computer vision algorithms and libraries (python-opencv and dlib) to formulate the procedure for analysis of the facial asymmetry.

Recent advances in Vision Transformers (ViT) and Stable Diffusion (SD) models with their ability to capture rich semantic features of the image have been used for image correspondence tasks on natural images. In this paper, we examine the ability of a variety of pre-trained ViT (DINO, DINOv2, SAM, CLIP) and SD models, trained exclusively on natural images, for solving the correspondence problems on medical images. While many works have made a case for in-domain training, we show that the models trained on natural images can offer good performance on medical images across different modalities (CT,MR,Ultrasound) sourced from various manufacturers, over multiple anatomical regions (brain, thorax, abdomen, extremities), and on wide variety of tasks. Further, we leverage the correspondence with respect to a template image to prompt a Segment Anything (SAM) model to arrive at single shot segmentation, achieving dice range of 62%-90% across tasks, using just one image as reference. We also show that our single-shot method outperforms the recently proposed few-shot segmentation method - UniverSeg (Dice range 47%-80%) on most of the semantic segmentation tasks(six out of seven) across medical imaging modalities., Comment: Accepted at NeurIPS 2023 R0-FoMo Workshop

Customer feedback is invaluable to companies as they refine their products. Monitoring customer feedback can be automated with Aspect Level Sentiment Classification (ALSC) which allows us to analyse specific aspects of the products in reviews. Large Language Models (LLMs) are the heart of many state-of-the-art ALSC solutions, but they perform poorly in some scenarios requiring Coreference Resolution (CR). In this work, we propose a framework to improve an LLM's performance on CR-containing reviews by fine tuning on highly inferential tasks. We show that the performance improvement is likely attributed to the improved model CR ability. We also release a new dataset that focuses on CR in ALSC., Comment: Work done up till December 2022

Novel intent class detection is an important problem in real world scenario for conversational agents for continuous interaction. Several research works have been done to detect novel intents in a mono-lingual (primarily English) texts and images. But, current systems lack an end-to-end universal framework to detect novel intents across various different languages with less human annotation effort for mis-classified and system rejected samples. This paper proposes NIDAL (Novel Intent Detection and Active Learning based classification), a semi-supervised framework to detect novel intents while reducing human annotation cost. Empirical results on various benchmark datasets demonstrate that this system outperforms the baseline methods by more than 10% margin for accuracy and macro-F1. The system achieves this while maintaining overall annotation cost to be just ~6-10% of the unlabeled data available to the system., Comment: AAAI 2023 Student Abstract

Scarcity of data and technological limitations for resource-poor languages in developing countries like India poses a threat to the development of sophisticated NLU systems for healthcare. To assess the current status of various state-of-the-art language models in healthcare, this paper studies the problem by initially proposing two different Healthcare datasets, Indian Healthcare Query Intent-WebMD and 1mg (IHQID-WebMD and IHQID-1mg) and one real world Indian hospital query data in English and multiple Indic languages (Hindi, Bengali, Tamil, Telugu, Marathi and Gujarati) which are annotated with the query intents as well as entities. Our aim is to detect query intents and extract corresponding entities. We perform extensive experiments on a set of models in various realistic settings and explore two scenarios based on the access to English data only (less costly) and access to target language data (more expensive). We analyze context specific practical relevancy through empirical analysis. The results, expressed in terms of overall F1 score show that our approach is practically useful to identify intents and entities.

While developing perception based deep learning models, the benefit of synthetic data is enormous. However, performance of networks trained with synthetic data for certain computer vision tasks degrade significantly when tested on real world data due to the domain gap between them. One of the popular solutions in bridging this gap between synthetic and actual world data is to frame it as a domain adaptation task. In this paper, we propose and evaluate novel ways for the betterment of such approaches. In particular we build upon the method of UNIT-GAN. In normal GAN training for the task of domain translation, pairing of images from both the domains (viz, real and synthetic) is done randomly. We propose a novel method to efficiently incorporate semantic supervision into this pair selection, which helps in boosting the performance of the model along with improving the visual quality of such transformed images. We illustrate our empirical findings on Cityscapes \cite{cityscapes} and challenging synthetic dataset Synscapes. Though the findings are reported on the base network of UNIT-GAN, they can be easily extended to any other similar network.

As language models are increasingly included in human-facing machine learning tools, bias against demographic subgroups has gained attention. We propose FineDeb, a two-phase debiasing framework for language models that starts with contextual debiasing of embeddings learned by pretrained language models. The model is then fine-tuned on a language modeling objective. Our results show that FineDeb offers stronger debiasing in comparison to other methods which often result in models as biased as the original language model. Our framework is generalizable for demographics with multiple classes, and we demonstrate its effectiveness through extensive experiments and comparisons with state of the art techniques. We release our code and data on GitHub., Comment: Poster presentation at AAAI 2023: The Workshop on Artificial Intelligence for Social Good 2023 (https://amulyayadav.github.io/AI4SG2023/)

For a group of pedestrians without any spatial boundaries, the methods of density estimation is a wide area of research. Besides, there is a specific difficulty when the density along one given pedestrian trajectory is needed in order to plot an `individual-based' fundamental diagram. We illustrate why several methods become ill-defined in this case. We then turn to the widely used Voronoi-cell based density estimate. We show that for a typical situation of crossing flows of pedestrians, Voronoi method has to be adapted to the small sample size. We conclude with general remarks about the meaning of density measurements in such context., Comment: 8 pages, 4 figures. Accepted for Traffic and Granular Flow 2022

In real world scenarios, out-of-distribution (OOD) datasets may have a large distributional shift from training datasets. This phenomena generally occurs when a trained classifier is deployed on varying dynamic environments, which causes a significant drop in performance. To tackle this issue, we are proposing an end-to-end deep multi-task network in this work. Observing a strong relationship between rotation prediction (self-supervised) accuracy and semantic classification accuracy on OOD tasks, we introduce an additional auxiliary classification head in our multi-task network along with semantic classification and rotation prediction head. To observe the influence of this addition classifier in improving the rotation prediction head, our proposed learning method is framed into bi-level optimisation problem where the upper-level is trained to update the parameters for semantic classification and rotation prediction head. In the lower-level optimisation, only the auxiliary classification head is updated through semantic classification head by fixing the parameters of the semantic classification head. The proposed method has been validated through three unseen OOD datasets where it exhibits a clear improvement in semantic classification accuracy than other two baseline methods. Our code is available on GitHub \url{https://github.com/harshita-555/OSSL}, Comment: The shorter version is accepted at the 29th IEEE International Conference on Image Processing (IEEE ICIP 2022)