Showing 1–50 of 80 results for author: Agarwal, D

Support Tokens, Stability Margins, and a New Foundation for Robust LLMs

Authors: Deepak Agarwal, Dhyey Dharmendrakumar Mavani, Suyash Gupta, Karthik Sethuraman, Tejas Dharamsi

Abstract: Self-attention is usually described as a flexible, content-adaptive way to mix a token with information from its past. We reinterpret causal self-attention transformers, the backbone of modern foundation models, within a probabilistic framework, much as classical PCA is extended to probabilistic PCA. This reformulation reveals a key structural consequence of the underlying change of variables: a b… ▽ More Self-attention is usually described as a flexible, content-adaptive way to mix a token with information from its past. We reinterpret causal self-attention transformers, the backbone of modern foundation models, within a probabilistic framework, much as classical PCA is extended to probabilistic PCA. This reformulation reveals a key structural consequence of the underlying change of variables: a barrier constraint emerges on the parameters of self-attention. The resulting geometry exposes a degeneracy boundary where the attention-induced mapping becomes locally ill-conditioned, yielding a stability-margin interpretation analogous to the margin in support vector machines. This, in turn, naturally gives rise to the concept of support tokens. We further show that causal transformers define a consistent stochastic process over infinite token sequences, providing a rigorous probabilistic foundation for sequence modeling. Building on this view, we derive a Bayesian MAP training objective that requires only a minimal modification to standard LLM training: adding a smooth log-barrier penalty to the usual cross-entropy loss. Empirically, the resulting training objective improves robustness to input perturbations and sharpens the margin geometry of the learned representations without sacrificing out-of-sample accuracy. △ Less

Submitted 21 March, 2026; v1 submitted 25 February, 2026; originally announced February 2026.

Comments: 45 pages, 9 figures

ACM Class: I.2.7; G.3; G.4

Understanding Frechet Speech Distance for Synthetic Speech Quality Evaluation

Authors: June-Woo Kim, Dhruv Agarwal, Federica Cerina

Abstract: Objective evaluation of synthetic speech quality remains a critical challenge. Human listening tests are the gold standard, but costly and impractical at scale. Fréchet Distance has emerged as a promising alternative, yet its reliability depends heavily on the choice of embeddings and experimental settings. In this work, we comprehensively evaluate Fréchet Speech Distance (FSD) and its variant Spe… ▽ More Objective evaluation of synthetic speech quality remains a critical challenge. Human listening tests are the gold standard, but costly and impractical at scale. Fréchet Distance has emerged as a promising alternative, yet its reliability depends heavily on the choice of embeddings and experimental settings. In this work, we comprehensively evaluate Fréchet Speech Distance (FSD) and its variant Speech Maximum Mean Discrepancy (SMMD) under varied embeddings and conditions. We further incorporate human listening evaluations alongside TTS intelligibility and synthetic-trained ASR WER to validate the perceptual relevance of these metrics. Our findings show that WavLM Base+ features yield the most stable alignment with human ratings. While FSD and SMMD cannot fully replace subjective evaluation, we show that they can serve as complementary, cost-efficient, and reproducible measures, particularly useful when large-scale or direct listening assessments are infeasible. Code is available at https://github.com/kaen2891/FrechetSpeechDistance. △ Less

Submitted 29 January, 2026; originally announced January 2026.

Comments: accepted to ICASSP 2026

Reinforcement Learning for Self-Healing Material Systems

Authors: Maitreyi Chatterjee, Devansh Agarwal, Biplab Chatterjee

Abstract: The transition to autonomous material systems necessitates adaptive control methodologies to maximize structural longevity. This study frames the self-healing process as a Reinforcement Learning (RL) problem within a Markov Decision Process (MDP), enabling agents to autonomously derive optimal policies that efficiently balance structural integrity maintenance against finite resource consumption. A… ▽ More The transition to autonomous material systems necessitates adaptive control methodologies to maximize structural longevity. This study frames the self-healing process as a Reinforcement Learning (RL) problem within a Markov Decision Process (MDP), enabling agents to autonomously derive optimal policies that efficiently balance structural integrity maintenance against finite resource consumption. A comparative evaluation of discrete-action (Q-learning, DQN) and continuous-action (TD3) agents in a stochastic simulation environment revealed that RL controllers significantly outperform heuristic baselines, achieving near-complete material recovery. Crucially, the TD3 agent utilizing continuous dosage control demonstrated superior convergence speed and stability, underscoring the necessity of fine-grained, proportional actuation in dynamic self-healing applications. △ Less

Submitted 23 November, 2025; originally announced November 2025.

Comments: Accepted to INCOM 2026. This is the camera-ready version

LogSyn: A Few-Shot LLM Framework for Structured Insight Extraction from Unstructured General Aviation Maintenance Logs

Authors: Devansh Agarwal, Maitreyi Chatterjee, Biplab Chatterjee

Abstract: Aircraft maintenance logs hold valuable safety data but remain underused due to their unstructured text format. This paper introduces LogSyn, a framework that uses Large Language Models (LLMs) to convert these logs into structured, machine-readable data. Using few-shot in-context learning on 6,169 records, LogSyn performs Controlled Abstraction Generation (CAG) to summarize problem-resolution narr… ▽ More Aircraft maintenance logs hold valuable safety data but remain underused due to their unstructured text format. This paper introduces LogSyn, a framework that uses Large Language Models (LLMs) to convert these logs into structured, machine-readable data. Using few-shot in-context learning on 6,169 records, LogSyn performs Controlled Abstraction Generation (CAG) to summarize problem-resolution narratives and classify events within a detailed hierarchical ontology. The framework identifies key failure patterns, offering a scalable method for semantic structuring and actionable insight extraction from maintenance logs. This work provides a practical path to improve maintenance workflows and predictive analytics in aviation and related industries. △ Less

Submitted 7 February, 2026; v1 submitted 23 November, 2025; originally announced November 2025.

Comments: Accepted in Proceedings of the 3rd INCOM 2026

Confidential Computing for Cloud Security: Exploring Hardware based Encryption Using Trusted Execution Environments

Authors: Dhruv Deepak Agarwal, Aswani Kumar Cherukuri

Abstract: The growth of cloud computing has revolutionized data processing and storage capacities to another levels of scalability and flexibility. But in the process, it has created a huge challenge of security, especially in terms of safeguarding sensitive data. Classical security practices, including encryption at rest and during transit, fail to protect data in use and expose it to various possible brea… ▽ More The growth of cloud computing has revolutionized data processing and storage capacities to another levels of scalability and flexibility. But in the process, it has created a huge challenge of security, especially in terms of safeguarding sensitive data. Classical security practices, including encryption at rest and during transit, fail to protect data in use and expose it to various possible breaches. In response to this problem , Confidential Computing has been a tool ,seeking to secure data in processing by usage of hardware-based Trusted Execution Environments (TEEs). TEEs, including Intel's Software Guard Extensions (SGX) and ARM's TrustZone, offers protected contexts within the processor, where data is kept confidential ,intact and secure , even with malicious software or compromised operating systems. In this research, we have explored the architecture and security features of TEEs like Intel SGX and ARM TrustZone, and their effectiveness in improving cloud data security. From a thorough literature survey ,we have analyzed the deployment strategies, performance indicators, and practical uses of these TEEs for the same purpose. In addition, we have discussed the issues regarding deployment, possible weaknesses, scalability issues, and integration issues. Our results focuses on the central position of TEEs in strengthening and advancing cloud security infrastructures, pointing towards their ability to create a secure foundation for Confidential Computing. △ Less

Submitted 6 November, 2025; originally announced November 2025.

The Universal Personalizer: Few-Shot Dysarthric Speech Recognition via Meta-Learning

Authors: Dhruuv Agarwal, Harry Zhang, Yang Yu, Quan Wang

Abstract: Personalizing dysarthric ASR is hindered by demanding enrollment collection and per-user training. We propose a hybrid meta-training method for a single model, enabling zero-shot and few-shot on-the-fly personalization via in-context learning (ICL). On Euphonia, it achieves 13.9% Word Error Rate (WER), surpassing speaker-independent baselines (17.5%). On SAP Test-1, our 5.3% WER outperforms the ch… ▽ More Personalizing dysarthric ASR is hindered by demanding enrollment collection and per-user training. We propose a hybrid meta-training method for a single model, enabling zero-shot and few-shot on-the-fly personalization via in-context learning (ICL). On Euphonia, it achieves 13.9% Word Error Rate (WER), surpassing speaker-independent baselines (17.5%). On SAP Test-1, our 5.3% WER outperforms the challenge-winning team (5.97%). On Test-2, our 9.49% trails only the winner (8.11%) but without relying on techniques like offline model-merging or custom audio chunking. Curation yields a 40% WER reduction using random same-speaker examples, validating active personalization. While static text curation fails to beat this baseline, oracle similarity reveals substantial headroom, highlighting dynamic acoustic retrieval as the next frontier. Data ablations confirm rapid low-resource speaker adaptation, establishing the model as a practical personalized solution. △ Less

Submitted 22 February, 2026; v1 submitted 18 September, 2025; originally announced September 2025.

Pothole Detection and Analysis System (PoDAS) for Real Time Data Using Sensor Networks

Authors: Jinesh Mehta, Vinayak Mathur, Dhruv Agarwal, Atish Sharma, Krishna Prakasha

Abstract: Potholes are a major nuisance on the city roads leading to several problems and losses in productivity. Local authorities have cited a lack of geographic localization of these potholes as one of the rate-limiting factors for repairs. This study proposes a novel low-cost wireless sensor-based end-to-end system called PoDAS (Pothole Detection and Analysis System) which can be deployed across major c… ▽ More Potholes are a major nuisance on the city roads leading to several problems and losses in productivity. Local authorities have cited a lack of geographic localization of these potholes as one of the rate-limiting factors for repairs. This study proposes a novel low-cost wireless sensor-based end-to-end system called PoDAS (Pothole Detection and Analysis System) which can be deployed across major cities. We discuss multiple implementation models that can be varied based on the needs of individual cities. Our system uses cross-validation through multiple sensors to achieve higher efficiency than some of the previous models that have been proposed. We also present the results from extensive testing carried out in different environments to ascertain both the efficacy and the efficiency of the proposed system. △ Less

Submitted 14 August, 2025; originally announced August 2025.

Comments: Published in Journal of Engineering and Applied Sciences

Journal ref: Journal of Engineering and Applied Sciences 12(12): 3090-3097, 2017

Semantic Anchoring in Agentic Memory: Leveraging Linguistic Structures for Persistent Conversational Context

Authors: Maitreyi Chatterjee, Devansh Agarwal

Abstract: Large Language Models (LLMs) have demonstrated impressive fluency and task competence in conversational settings. However, their effectiveness in multi-session and long-term interactions is hindered by limited memory persistence. Typical retrieval-augmented generation (RAG) systems store dialogue history as dense vectors, which capture semantic similarity but neglect finer linguistic structures su… ▽ More Large Language Models (LLMs) have demonstrated impressive fluency and task competence in conversational settings. However, their effectiveness in multi-session and long-term interactions is hindered by limited memory persistence. Typical retrieval-augmented generation (RAG) systems store dialogue history as dense vectors, which capture semantic similarity but neglect finer linguistic structures such as syntactic dependencies, discourse relations, and coreference links. We propose Semantic Anchoring, a hybrid agentic memory architecture that enriches vector-based storage with explicit linguistic cues to improve recall of nuanced, context-rich exchanges. Our approach combines dependency parsing, discourse relation tagging, and coreference resolution to create structured memory entries. Experiments on adapted long-term dialogue datasets show that semantic anchoring improves factual recall and discourse coherence by up to 18% over strong RAG baselines. We further conduct ablation studies, human evaluations, and error analysis to assess robustness and interpretability. △ Less

Submitted 18 August, 2025; originally announced August 2025.

Comments: Paper is currently in peer review

MiGrATe: Mixed-Policy GRPO for Adaptation at Test-Time

Authors: Peter Phan, Dhruv Agarwal, Kavitha Srinivas, Horst Samulowitz, Pavan Kapanipathi, Andrew McCallum

Abstract: Large language models (LLMs) are increasingly being applied to black-box optimization tasks, from program synthesis to molecule design. Prior work typically leverages in-context learning to iteratively guide the model towards better solutions. Such methods, however, often struggle to balance exploration of new solution spaces with exploitation of high-reward ones. Recently, test-time training (TTT… ▽ More Large language models (LLMs) are increasingly being applied to black-box optimization tasks, from program synthesis to molecule design. Prior work typically leverages in-context learning to iteratively guide the model towards better solutions. Such methods, however, often struggle to balance exploration of new solution spaces with exploitation of high-reward ones. Recently, test-time training (TTT) with synthetic data has shown promise in improving solution quality. However, the need for hand-crafted training data tailored to each task limits feasibility and scalability across domains. To address this problem, we introduce MiGrATe-a method for online TTT that uses GRPO as a search algorithm to adapt LLMs at inference without requiring external training data. MiGrATe operates via a mixed-policy group construction procedure that combines on-policy sampling with two off-policy data selection techniques: greedy sampling, which selects top-performing past completions, and neighborhood sampling (NS), which generates completions structurally similar to high-reward ones. Together, these components bias the policy gradient towards exploitation of promising regions in solution space, while preserving exploration through on-policy sampling. We evaluate MiGrATe on three challenging domains-word search, molecule optimization, and hypothesis+program induction on the Abstraction and Reasoning Corpus (ARC)-and find that it consistently outperforms both inference-only and TTT baselines, demonstrating the potential of online TTT as a solution for complex search tasks without external supervision. △ Less

Submitted 12 August, 2025; originally announced August 2025.

AutoDiscovery: Open-ended Scientific Discovery via Bayesian Surprise

Authors: Dhruv Agarwal, Bodhisattwa Prasad Majumder, Reece Adamson, Megha Chakravorty, Satvika Reddy Gavireddy, Aditya Parashar, Harshit Surana, Bhavana Dalvi Mishra, Andrew McCallum, Ashish Sabharwal, Peter Clark

Abstract: The promise of autonomous scientific discovery (ASD) hinges not only on answering questions, but also on knowing which questions to ask. Most recent works in ASD explore the use of large language models (LLMs) in goal-driven settings, relying on human-specified research questions to guide hypothesis generation. However, scientific discovery may be accelerated further by allowing the AI system to d… ▽ More The promise of autonomous scientific discovery (ASD) hinges not only on answering questions, but also on knowing which questions to ask. Most recent works in ASD explore the use of large language models (LLMs) in goal-driven settings, relying on human-specified research questions to guide hypothesis generation. However, scientific discovery may be accelerated further by allowing the AI system to drive exploration by its own criteria. The few existing approaches in open-ended ASD select hypotheses based on diversity heuristics or subjective proxies for human interestingness, but the former struggles to meaningfully navigate the typically vast hypothesis space, and the latter suffers from imprecise definitions. This paper presents AutoDiscovery -- a method for open-ended ASD that instead drives scientific exploration using Bayesian surprise. Here, we quantify the epistemic shift from the LLM's prior beliefs about a hypothesis to its posterior beliefs after gathering experimental results. To efficiently explore the space of nested hypotheses, our method employs a Monte Carlo tree search (MCTS) strategy with progressive widening using surprisal as the reward function. We evaluate AutoDiscovery in the setting of data-driven discovery across 21 real-world datasets spanning domains such as biology, economics, finance, and behavioral science. Our results demonstrate that under a fixed budget, AutoDiscovery substantially outperforms competitors by producing 5-29% more discoveries deemed surprising by the LLM. Our human evaluation further reveals that two-thirds of discoveries made by our system are surprising to domain experts as well, suggesting this is an important step towards building open-ended ASD systems. △ Less

Submitted 12 February, 2026; v1 submitted 30 June, 2025; originally announced July 2025.

Comments: Accepted to NeurIPS 2025: https://neurips.cc/virtual/2025/loc/san-diego/poster/116398

Impact of Comments on LLM Comprehension of Legacy Code

Authors: Rock Sabetto, Emily Escamilla, Devesh Agarwal, Sujay Kandwal, Justin F. Brunelle, Scott Rosen, Nitin Naik, Samruddhi Thaker, Eric O. Scott, Jacob Zimmer, Amit Madan, Arun Sridharan, Doug Wendt, Michael Doyle, Christopher Glasz, Jasper Phillips, William Macke, Colin Diggs, Michael Bartholf, Zachary Robin, Paul Ursino

Abstract: Large language models (LLMs) have been increasingly integrated into software engineering and maintenance tasks due to their high performance with software engineering tasks and robust understanding of modern programming languages. However, the ability of LLMs to comprehend code written with legacy languages remains a research gap challenged by real-world legacy systems lacking or containing inaccu… ▽ More Large language models (LLMs) have been increasingly integrated into software engineering and maintenance tasks due to their high performance with software engineering tasks and robust understanding of modern programming languages. However, the ability of LLMs to comprehend code written with legacy languages remains a research gap challenged by real-world legacy systems lacking or containing inaccurate documentation that may impact LLM comprehension. To assess LLM comprehension of legacy languages, there is a need for objective LLM evaluation. In order to objectively measure LLM comprehension of legacy languages, we need an efficient, quantitative evaluation method. We leverage multiple-choice question answering (MCQA), an emerging LLM evaluation methodology, to evaluate LLM comprehension of legacy code and the impact of comment prevalence and inaccurate comments. In this work, we present preliminary findings on the impact of documentation on LLM comprehension of legacy code and outline strategic objectives for future work. △ Less

Submitted 23 April, 2025; originally announced June 2025.

LLM-Synth4KWS: Scalable Automatic Generation and Synthesis of Confusable Data for Custom Keyword Spotting

Authors: Pai Zhu, Quan Wang, Dhruuv Agarwal, Kurt Partridge

Abstract: Custom keyword spotting (KWS) allows detecting user-defined spoken keywords from streaming audio. This is achieved by comparing the embeddings from voice enrollments and input audio. State-of-the-art custom KWS models are typically trained contrastively using utterances whose keywords are randomly sampled from training dataset. These KWS models often struggle with confusing keywords, such as "blue… ▽ More Custom keyword spotting (KWS) allows detecting user-defined spoken keywords from streaming audio. This is achieved by comparing the embeddings from voice enrollments and input audio. State-of-the-art custom KWS models are typically trained contrastively using utterances whose keywords are randomly sampled from training dataset. These KWS models often struggle with confusing keywords, such as "blue" versus "glue". This paper introduces an effective way to augment the training with confusable utterances where keywords are generated and grouped from large language models (LLMs), and speech signals are synthesized with diverse speaking styles from text-to-speech (TTS) engines. To better measure user experience on confusable KWS, we define a new northstar metric using the average area under DET curve from confusable groups (c-AUC). Featuring high scalability and zero labor cost, the proposed method improves AUC by 3.7% and c-AUC by 11.3% on the Speech Commands testing set. △ Less

Submitted 28 May, 2025; originally announced May 2025.

Journal ref: Proc. Interspeech 2025, 2675-2679

Fluent but Foreign: Even Regional LLMs Lack Cultural Alignment

Authors: Dhruv Agarwal, Anya Shukla, Sunayana Sitaram, Aditya Vashistha

Abstract: Large language models (LLMs) are used worldwide, yet exhibit Western cultural tendencies. Many countries are now building ``regional'' or ``sovereign'' LLMs, but it remains unclear whether they reflect local values and practices or merely speak local languages. Using India as a case study, we evaluate six Indic and six global LLMs on two dimensions -- values and practices -- grounded in nationally… ▽ More Large language models (LLMs) are used worldwide, yet exhibit Western cultural tendencies. Many countries are now building ``regional'' or ``sovereign'' LLMs, but it remains unclear whether they reflect local values and practices or merely speak local languages. Using India as a case study, we evaluate six Indic and six global LLMs on two dimensions -- values and practices -- grounded in nationally representative surveys and community-sourced QA datasets. Across tasks, Indic models do not align better with Indian norms than global models; in fact, a U.S. respondent is a closer proxy for Indian values than any Indic model. We further run a user study with 115 Indian users and find that writing suggestions from both global and Indic LLMs introduce Westernized or exoticized writing. Prompting and regional fine-tuning fail to recover alignment and can even degrade existing knowledge. We attribute this to scarce culturally grounded data, especially for pretraining. We position cultural evaluation as a first-class requirement alongside multilingual benchmarks and offer a reusable, community-grounded methodology. We call for native, community-authored corpora and thickxwide evaluations to build truly sovereign LLMs. △ Less

Submitted 22 January, 2026; v1 submitted 24 May, 2025; originally announced May 2025.

Comments: Under review

CRMArena-Pro: Holistic Assessment of LLM Agents Across Diverse Business Scenarios and Interactions

Authors: Kung-Hsiang Huang, Akshara Prabhakar, Onkar Thorat, Divyansh Agarwal, Prafulla Kumar Choubey, Yixin Mao, Silvio Savarese, Caiming Xiong, Chien-Sheng Wu

Abstract: While AI agents hold transformative potential in business, effective performance benchmarking is hindered by the scarcity of public, realistic business data on widely used platforms. Existing benchmarks often lack fidelity in their environments, data, and agent-user interactions, with limited coverage of diverse business scenarios and industries. To address these gaps, we introduce CRMArena-Pro, a… ▽ More While AI agents hold transformative potential in business, effective performance benchmarking is hindered by the scarcity of public, realistic business data on widely used platforms. Existing benchmarks often lack fidelity in their environments, data, and agent-user interactions, with limited coverage of diverse business scenarios and industries. To address these gaps, we introduce CRMArena-Pro, a novel benchmark for holistic, realistic assessment of LLM agents in diverse professional settings. CRMArena-Pro expands on CRMArena with nineteen expert-validated tasks across sales, service, and 'configure, price, and quote' processes, for both Business-to-Business and Business-to-Customer scenarios. It distinctively incorporates multi-turn interactions guided by diverse personas and robust confidentiality awareness assessments. Experiments reveal leading LLM agents achieve only around 58% single-turn success on CRMArena-Pro, with performance dropping significantly to approximately 35% in multi-turn settings. While Workflow Execution proves more tractable for top agents (over 83% single-turn success), other evaluated business skills present greater challenges. Furthermore, agents exhibit near-zero inherent confidentiality awareness; though targeted prompting can improve this, it often compromises task performance. These findings highlight a substantial gap between current LLM capabilities and enterprise demands, underscoring the need for advancements in multi-turn reasoning, confidentiality adherence, and versatile skill acquisition. △ Less

Submitted 24 May, 2025; originally announced May 2025.

GraphemeAug: A Systematic Approach to Synthesized Hard Negative Keyword Spotting Examples

Authors: Harry Zhang, Kurt Partridge, Pai Zhu, Neng Chen, Hyun Jin Park, Dhruuv Agarwal, Quan Wang

Abstract: Spoken Keyword Spotting (KWS) is the task of distinguishing between the presence and absence of a keyword in audio. The accuracy of a KWS model hinges on its ability to correctly classify examples close to the keyword and non-keyword boundary. These boundary examples are often scarce in training data, limiting model performance. In this paper, we propose a method to systematically generate adversa… ▽ More Spoken Keyword Spotting (KWS) is the task of distinguishing between the presence and absence of a keyword in audio. The accuracy of a KWS model hinges on its ability to correctly classify examples close to the keyword and non-keyword boundary. These boundary examples are often scarce in training data, limiting model performance. In this paper, we propose a method to systematically generate adversarial examples close to the decision boundary by making insertion/deletion/substitution edits on the keyword's graphemes. We evaluate this technique on held-out data for a popular keyword and show that the technique improves AUC on a dataset of synthetic hard negatives by 61% while maintaining quality on positives and ambient negative audio data. △ Less

Submitted 24 May, 2025; v1 submitted 20 May, 2025; originally announced May 2025.

Comments: Accepted at Interspeech 2025

Journal ref: Proc. Interspeech 2025, 2680-2684

Visual Feedback of Pattern Separability Improves Myoelectric Decoding Performance of Upper Limb Prostheses

Authors: Ruichen Yang, György M. Lévay, Christopher L. Hunt, Dániel Czeiner, Megan C. Hodgson, Damini Agarwal, Rahul R. Kaliki, Nitish V. Thakor

Abstract: State-of-the-art upper limb myoelectric prostheses often use pattern recognition (PR) control systems that translate electromyography (EMG) signals into desired movements. As prosthesis movement complexity increases, users often struggle to produce sufficiently distinct EMG patterns for reliable classification. Existing training typically involves heuristic, trial-and-error user adjustments to sta… ▽ More State-of-the-art upper limb myoelectric prostheses often use pattern recognition (PR) control systems that translate electromyography (EMG) signals into desired movements. As prosthesis movement complexity increases, users often struggle to produce sufficiently distinct EMG patterns for reliable classification. Existing training typically involves heuristic, trial-and-error user adjustments to static decoder boundaries. Goal: We introduce the Reviewer, a 3D visual interface projecting EMG signals directly into the decoder's classification space, providing intuitive, real-time insight into PR algorithm behavior. This structured feedback reduces cognitive load and fosters mutual, data-driven adaptation between user-generated EMG patterns and decoder boundaries. Methods: A 10-session study with 12 able-bodied participants compared PR performance after motor-based training and updating using the Reviewer versus conventional virtual arm visualization. Performance was assessed using a Fitts law task that involved the aperture of the cursor and the control of orientation. Results: Participants trained with the Reviewer achieved higher completion rates, reduced overshoot, and improved path efficiency and throughput compared to the standard visualization group. Significance: The Reviewer introduces decoder-informed motor training, facilitating immediate and consistent PR-based myoelectric control improvements. By iteratively refining control through real-time feedback, this approach reduces reliance on trial-and-error recalibration, enabling a more adaptive, self-correcting training framework. Conclusion: The 3D visual feedback significantly improves PR control in novice operators through structured training, enabling feedback-driven adaptation and reducing reliance on extensive heuristic adjustments. △ Less

Submitted 15 May, 2025; v1 submitted 14 May, 2025; originally announced May 2025.

Interpretable AI-driven Guidelines for Type 2 Diabetes Treatment from Observational Data

Authors: Dewang Kumar Agarwal, Dimitris J. Bertsimas

Abstract: Objective: Create precise, structured, data-backed guidelines for type 2 diabetes treatment progression, suitable for clinical adoption. Research Design and Methods: Our training cohort was composed of patient (with type 2 diabetes) visits from Boston Medical Center (BMC) from 1998 to 2014. We divide visits into 4 groups based on the patient's treatment regimen before the visit, and further divi… ▽ More Objective: Create precise, structured, data-backed guidelines for type 2 diabetes treatment progression, suitable for clinical adoption. Research Design and Methods: Our training cohort was composed of patient (with type 2 diabetes) visits from Boston Medical Center (BMC) from 1998 to 2014. We divide visits into 4 groups based on the patient's treatment regimen before the visit, and further divide them into subgroups based on the recommended treatment during the visit. Since each subgroup has observational data, which has confounding bias (sicker patients are prescribed more aggressive treatments), we used machine learning and optimization to remove some datapoints so that the remaining data resembles a randomized trial. On each subgroup, we train AI-backed tree-based models to prescribe treatment changes. Once we train these tree models, we manually combine the models for every group to create an end-to-end prescription pipeline for all patients in that group. In this process, we prioritize stepping up to a more aggressive treatment before considering less aggressive options. We tested this pipeline on unseen data from BMC, and an external dataset from Hartford healthcare (type 2 diabetes patient visits from January 2020 to May 2024). Results: The median HbA1c reduction achieved by our pipelines is 0.26% more than what the doctors achieved on the unseen BMC patients. For the Hartford cohort, our pipelines were better by 0.13%. Conclusions: This precise, interpretable, and efficient AI-backed approach to treatment progression in type 2 diabetes is predicted to outperform the current practice and can be deployed to improve patient outcomes. △ Less

Submitted 16 April, 2025; originally announced April 2025.

BingoGuard: LLM Content Moderation Tools with Risk Levels

Authors: Fan Yin, Philippe Laban, Xiangyu Peng, Yilun Zhou, Yixin Mao, Vaibhav Vats, Linnea Ross, Divyansh Agarwal, Caiming Xiong, Chien-Sheng Wu

Abstract: Malicious content generated by large language models (LLMs) can pose varying degrees of harm. Although existing LLM-based moderators can detect harmful content, they struggle to assess risk levels and may miss lower-risk outputs. Accurate risk assessment allows platforms with different safety thresholds to tailor content filtering and rejection. In this paper, we introduce per-topic severity rubri… ▽ More Malicious content generated by large language models (LLMs) can pose varying degrees of harm. Although existing LLM-based moderators can detect harmful content, they struggle to assess risk levels and may miss lower-risk outputs. Accurate risk assessment allows platforms with different safety thresholds to tailor content filtering and rejection. In this paper, we introduce per-topic severity rubrics for 11 harmful topics and build BingoGuard, an LLM-based moderation system designed to predict both binary safety labels and severity levels. To address the lack of annotations on levels of severity, we propose a scalable generate-then-filter framework that first generates responses across different severity levels and then filters out low-quality responses. Using this framework, we create BingoGuardTrain, a training dataset with 54,897 examples covering a variety of topics, response severity, styles, and BingoGuardTest, a test set with 988 examples explicitly labeled based on our severity rubrics that enables fine-grained analysis on model behaviors on different severity levels. Our BingoGuard-8B, trained on BingoGuardTrain, achieves the state-of-the-art performance on several moderation benchmarks, including WildGuardTest and HarmBench, as well as BingoGuardTest, outperforming best public models, WildGuard, by 4.3\%. Our analysis demonstrates that incorporating severity levels into training significantly enhances detection performance and enables the model to effectively gauge the severity of harmful responses. △ Less

Submitted 9 March, 2025; originally announced March 2025.

Comments: 10 pages, 4 figures, 4 tables. ICLR 2025 poster

Steering AI-Driven Personalization of Scientific Text for General Audiences

Authors: Taewook Kim, Dhruv Agarwal, Jordan Ackerman, Manaswi Saha

Abstract: Digital media platforms (e.g., science blogs) offer opportunities to communicate scientific content to general audiences at scale. However, these audiences vary in their scientific expertise, literacy levels, and personal backgrounds, making effective science communication challenging. To address this challenge, we designed TranSlider, an AI-powered tool that generates personalized translations of… ▽ More Digital media platforms (e.g., science blogs) offer opportunities to communicate scientific content to general audiences at scale. However, these audiences vary in their scientific expertise, literacy levels, and personal backgrounds, making effective science communication challenging. To address this challenge, we designed TranSlider, an AI-powered tool that generates personalized translations of scientific text based on individual user profiles (e.g., hobbies, location, and education). Our tool features an interactive slider that allows users to steer the degree of personalization from 0 (weakly relatable) to 100 (strongly relatable), leveraging LLMs to generate the translations with chosen degrees. Through an exploratory study with 15 participants, we investigated both the utility of these AI-personalized translations and how interactive reading features influenced users' understanding and reading experiences. We found that participants who preferred higher degrees of personalization appreciated the relatable and contextual translations, while those who preferred lower degrees valued concise translations with subtle contextualization. Furthermore, participants reported the compounding effect of multiple translations on their understanding of scientific content. Drawing on these findings, we discuss several implications for facilitating science communication and designing steerable interfaces to support human-AI alignment. △ Less

Submitted 9 August, 2025; v1 submitted 15 November, 2024; originally announced November 2024.

Comments: 28 pages, 7 figures, 1 table. Accepted to PACM HCI (CSCW 2025)

Evaluating Cultural and Social Awareness of LLM Web Agents

Authors: Haoyi Qiu, Alexander R. Fabbri, Divyansh Agarwal, Kung-Hsiang Huang, Sarah Tan, Nanyun Peng, Chien-Sheng Wu

Abstract: As large language models (LLMs) expand into performing as agents for real-world applications beyond traditional NLP tasks, evaluating their robustness becomes increasingly important. However, existing benchmarks often overlook critical dimensions like cultural and social awareness. To address these, we introduce CASA, a benchmark designed to assess LLM agents' sensitivity to cultural and social no… ▽ More As large language models (LLMs) expand into performing as agents for real-world applications beyond traditional NLP tasks, evaluating their robustness becomes increasingly important. However, existing benchmarks often overlook critical dimensions like cultural and social awareness. To address these, we introduce CASA, a benchmark designed to assess LLM agents' sensitivity to cultural and social norms across two web-based tasks: online shopping and social discussion forums. Our approach evaluates LLM agents' ability to detect and appropriately respond to norm-violating user queries and observations. Furthermore, we propose a comprehensive evaluation framework that measures awareness coverage, helpfulness in managing user queries, and the violation rate when facing misleading web content. Experiments show that current LLMs perform significantly better in non-agent than in web-based agent environments, with agents achieving less than 10% awareness coverage and over 40% violation rates. To improve performance, we explore two methods: prompting and fine-tuning, and find that combining both methods can offer complementary advantages -- fine-tuning on culture-specific datasets significantly enhances the agents' ability to generalize across different regions, while prompting boosts the agents' ability to navigate complex tasks. These findings highlight the importance of constantly benchmarking LLM agents' cultural and social awareness during the development cycle. △ Less

Submitted 8 March, 2025; v1 submitted 30 October, 2024; originally announced October 2024.

Comments: NAACL 2025 Findings

GE2E-KWS: Generalized End-to-End Training and Evaluation for Zero-shot Keyword Spotting

Authors: Pai Zhu, Jacob W. Bartel, Dhruuv Agarwal, Kurt Partridge, Hyun Jin Park, Quan Wang

Abstract: We propose GE2E-KWS -- a generalized end-to-end training and evaluation framework for customized keyword spotting. Specifically, enrollment utterances are separated and grouped by keywords from the training batch and their embedding centroids are compared to all other test utterance embeddings to compute the loss. This simulates runtime enrollment and verification stages, and improves convergence… ▽ More We propose GE2E-KWS -- a generalized end-to-end training and evaluation framework for customized keyword spotting. Specifically, enrollment utterances are separated and grouped by keywords from the training batch and their embedding centroids are compared to all other test utterance embeddings to compute the loss. This simulates runtime enrollment and verification stages, and improves convergence stability and training speed by optimizing matrix operations compared to SOTA triplet loss approaches. To benchmark different models reliably, we propose an evaluation process that mimics the production environment and compute metrics that directly measure keyword matching accuracy. Trained with GE2E loss, our 419KB quantized conformer model beats a 7.5GB ASR encoder by 23.6% relative AUC, and beats a same size triplet loss model by 60.7% AUC. Our KWS models are natively streamable with low memory footprints, and designed to continuously run on-device with no retraining needed for new keywords (zero-shot). △ Less

Submitted 21 October, 2024; originally announced October 2024.

Comments: 8 pages, 6 figures, 2 tables The paper is accepted in IEEE Spoken Language Technology (SLT) 2024

Journal ref: 2024 IEEE Spoken Language Technology Workshop (SLT)

Sylber: Syllabic Embedding Representation of Speech from Raw Audio

Authors: Cheol Jun Cho, Nicholas Lee, Akshat Gupta, Dhruv Agarwal, Ethan Chen, Alan W Black, Gopala K. Anumanchipalli

Abstract: Syllables are compositional units of spoken language that efficiently structure human speech perception and production. However, current neural speech representations lack such structure, resulting in dense token sequences that are costly to process. To bridge this gap, we propose a new model, Sylber, that produces speech representations with clean and robust syllabic structure. Specifically, we p… ▽ More Syllables are compositional units of spoken language that efficiently structure human speech perception and production. However, current neural speech representations lack such structure, resulting in dense token sequences that are costly to process. To bridge this gap, we propose a new model, Sylber, that produces speech representations with clean and robust syllabic structure. Specifically, we propose a self-supervised learning (SSL) framework that bootstraps syllabic embeddings by distilling from its own initial unsupervised syllabic segmentation. This results in a highly structured representation of speech features, offering three key benefits: 1) a fast, linear-time syllable segmentation algorithm, 2) efficient syllabic tokenization with an average of 4.27 tokens per second, and 3) novel phonological units suited for efficient spoken language modeling. Our proposed segmentation method is highly robust and generalizes to out-of-domain data and unseen languages without any tuning. By training token-to-speech generative models, fully intelligible speech can be reconstructed from Sylber tokens with a significantly lower bitrate than baseline SSL tokens. This suggests that our model effectively compresses speech into a compact sequence of tokens with minimal information loss. Lastly, we demonstrate that categorical perception-a linguistic phenomenon in speech perception-emerges naturally in Sylber, making the embedding space more categorical and sparse than previous speech features and thus supporting the high efficiency of our tokenization. Together, we present a novel SSL approach for representing speech as syllables, with significant potential for efficient speech tokenization and spoken language modeling. △ Less

Submitted 2 March, 2025; v1 submitted 9 October, 2024; originally announced October 2024.

Comments: Accepted at ICLR 2025

ViDAS: Vision-based Danger Assessment and Scoring

Authors: Pranav Gupta, Advith Krishnan, Naman Nanda, Ananth Eswar, Deeksha Agarwal, Pratham Gohil, Pratyush Goel

Abstract: We present a novel dataset aimed at advancing danger analysis and assessment by addressing the challenge of quantifying danger in video content and identifying how human-like a Large Language Model (LLM) evaluator is for the same. This is achieved by compiling a collection of 100 YouTube videos featuring various events. Each video is annotated by human participants who provided danger ratings on a… ▽ More We present a novel dataset aimed at advancing danger analysis and assessment by addressing the challenge of quantifying danger in video content and identifying how human-like a Large Language Model (LLM) evaluator is for the same. This is achieved by compiling a collection of 100 YouTube videos featuring various events. Each video is annotated by human participants who provided danger ratings on a scale from 0 (no danger to humans) to 10 (life-threatening), with precise timestamps indicating moments of heightened danger. Additionally, we leverage LLMs to independently assess the danger levels in these videos using video summaries. We introduce Mean Squared Error (MSE) scores for multimodal meta-evaluation of the alignment between human and LLM danger assessments. Our dataset not only contributes a new resource for danger assessment in video content but also demonstrates the potential of LLMs in achieving human-like evaluations. △ Less

Submitted 1 October, 2024; originally announced October 2024.

Comments: Preprint

AI Suggestions Homogenize Writing Toward Western Styles and Diminish Cultural Nuances

Authors: Dhruv Agarwal, Mor Naaman, Aditya Vashistha

Abstract: Large language models (LLMs) are being increasingly integrated into everyday products and services, such as coding tools and writing assistants. As these embedded AI applications are deployed globally, there is a growing concern that the AI models underlying these applications prioritize Western values. This paper investigates what happens when a Western-centric AI model provides writing suggestio… ▽ More Large language models (LLMs) are being increasingly integrated into everyday products and services, such as coding tools and writing assistants. As these embedded AI applications are deployed globally, there is a growing concern that the AI models underlying these applications prioritize Western values. This paper investigates what happens when a Western-centric AI model provides writing suggestions to users from a different cultural background. We conducted a cross-cultural controlled experiment with 118 participants from India and the United States who completed culturally grounded writing tasks with and without AI suggestions. Our analysis reveals that AI provided greater efficiency gains for Americans compared to Indians. Moreover, AI suggestions led Indian participants to adopt Western writing styles, altering not just what is written but also how it is written. These findings show that Western-centric AI models homogenize writing toward Western norms, diminishing nuances that differentiate cultural expression. △ Less

Submitted 12 March, 2025; v1 submitted 17 September, 2024; originally announced September 2024.

Comments: Accepted at CHI 2025

Adversarial training of Keyword Spotting to Minimize TTS Data Overfitting

Authors: Hyun Jin Park, Dhruuv Agarwal, Neng Chen, Rentao Sun, Kurt Partridge, Justin Chen, Harry Zhang, Pai Zhu, Jacob Bartel, Kyle Kastner, Gary Wang, Andrew Rosenberg, Quan Wang

Abstract: The keyword spotting (KWS) problem requires large amounts of real speech training data to achieve high accuracy across diverse populations. Utilizing large amounts of text-to-speech (TTS) synthesized data can reduce the cost and time associated with KWS development. However, TTS data may contain artifacts not present in real speech, which the KWS model can exploit (overfit), leading to degraded ac… ▽ More The keyword spotting (KWS) problem requires large amounts of real speech training data to achieve high accuracy across diverse populations. Utilizing large amounts of text-to-speech (TTS) synthesized data can reduce the cost and time associated with KWS development. However, TTS data may contain artifacts not present in real speech, which the KWS model can exploit (overfit), leading to degraded accuracy on real speech. To address this issue, we propose applying an adversarial training method to prevent the KWS model from learning TTS-specific features when trained on large amounts of TTS data. Experimental results demonstrate that KWS model accuracy on real speech data can be improved by up to 12% when adversarial loss is used in addition to the original KWS loss. Surprisingly, we also observed that the adversarial setup improves accuracy by up to 8%, even when trained solely on TTS and real negative speech data, without any real positive examples. △ Less

Submitted 19 August, 2024; originally announced August 2024.

Comments: to be published in a Workshop at Interspeech 2024, Synthetic Data's Transformative Role in Foundational Speech Models

Journal ref: Proc. Synthetic Data's Transformative Role in Foundational Speech Models 2024, 86-90

Utilizing TTS Synthesized Data for Efficient Development of Keyword Spotting Model

Authors: Hyun Jin Park, Dhruuv Agarwal, Neng Chen, Rentao Sun, Kurt Partridge, Justin Chen, Harry Zhang, Pai Zhu, Jacob Bartel, Kyle Kastner, Gary Wang, Andrew Rosenberg, Quan Wang

Abstract: This paper explores the use of TTS synthesized training data for KWS (keyword spotting) task while minimizing development cost and time. Keyword spotting models require a huge amount of training data to be accurate, and obtaining such training data can be costly. In the current state of the art, TTS models can generate large amounts of natural-sounding data, which can help reducing cost and time f… ▽ More This paper explores the use of TTS synthesized training data for KWS (keyword spotting) task while minimizing development cost and time. Keyword spotting models require a huge amount of training data to be accurate, and obtaining such training data can be costly. In the current state of the art, TTS models can generate large amounts of natural-sounding data, which can help reducing cost and time for KWS model development. Still, TTS generated data can be lacking diversity compared to real data. To pursue maximizing KWS model accuracy under the constraint of limited resources and current TTS capability, we explored various strategies to mix TTS data and real human speech data, with a focus on minimizing real data use and maximizing diversity of TTS output. Our experimental results indicate that relatively small amounts of real audio data with speaker diversity (100 speakers, 2k utterances) and large amounts of TTS synthesized data can achieve reasonably high accuracy (within 3x error rate of baseline), compared to the baseline (trained with 3.8M real positive utterances). △ Less

Submitted 26 July, 2024; originally announced July 2024.

Comments: to be published in a Workshop at Interspeech 2024, Synthetic Data's Transformative Role in Foundational Speech Models

Journal ref: Proc. Synthetic Data's Transformative Role in Foundational Speech Models 2024, 16-20

Synth4Kws: Synthesized Speech for User Defined Keyword Spotting in Low Resource Environments

Authors: Pai Zhu, Dhruuv Agarwal, Jacob W. Bartel, Kurt Partridge, Hyun Jin Park, Quan Wang

Abstract: One of the challenges in developing a high quality custom keyword spotting (KWS) model is the lengthy and expensive process of collecting training data covering a wide range of languages, phrases and speaking styles. We introduce Synth4Kws - a framework to leverage Text to Speech (TTS) synthesized data for custom KWS in different resource settings. With no real data, we found increasing TTS phrase… ▽ More One of the challenges in developing a high quality custom keyword spotting (KWS) model is the lengthy and expensive process of collecting training data covering a wide range of languages, phrases and speaking styles. We introduce Synth4Kws - a framework to leverage Text to Speech (TTS) synthesized data for custom KWS in different resource settings. With no real data, we found increasing TTS phrase diversity and utterance sampling monotonically improves model performance, as evaluated by EER and AUC metrics over 11k utterances of the speech command dataset. In low resource settings, with 50k real utterances as a baseline, we found using optimal amounts of TTS data can improve EER by 30.1% and AUC by 46.7%. Furthermore, we mix TTS data with varying amounts of real data and interpolate the real data needed to achieve various quality targets. Our experiments are based on English and single word utterances but the findings generalize to i18n languages and other keyword types. △ Less

Submitted 23 July, 2024; originally announced July 2024.

Comments: 5 pages, 5 figures, 2 tables The paper is accepted in Interspeech SynData4GenAI 2024 Workshop - https://syndata4genai.org/#call-for-papers

Journal ref: Proc. Synthetic Data's Transformative Role in Foundational Speech Models 2024, 11-15

DiscoveryBench: Towards Data-Driven Discovery with Large Language Models

Authors: Bodhisattwa Prasad Majumder, Harshit Surana, Dhruv Agarwal, Bhavana Dalvi Mishra, Abhijeetsingh Meena, Aryan Prakhar, Tirth Vora, Tushar Khot, Ashish Sabharwal, Peter Clark

Abstract: Can the rapid advances in code generation, function calling, and data analysis using large language models (LLMs) help automate the search and verification of hypotheses purely from a set of provided datasets? To evaluate this question, we present DiscoveryBench, the first comprehensive benchmark that formalizes the multi-step process of data-driven discovery. The benchmark is designed to systemat… ▽ More Can the rapid advances in code generation, function calling, and data analysis using large language models (LLMs) help automate the search and verification of hypotheses purely from a set of provided datasets? To evaluate this question, we present DiscoveryBench, the first comprehensive benchmark that formalizes the multi-step process of data-driven discovery. The benchmark is designed to systematically assess current model capabilities in discovery tasks and provide a useful resource for improving them. Our benchmark contains 264 tasks collected across 6 diverse domains, such as sociology and engineering, by manually deriving discovery workflows from published papers to approximate the real-world challenges faced by researchers, where each task is defined by a dataset, its metadata, and a discovery goal in natural language. We additionally provide 903 synthetic tasks to conduct controlled evaluations across task complexity. Furthermore, our structured formalism of data-driven discovery enables a facet-based evaluation that provides useful insights into different failure modes. We evaluate several popular LLM-based reasoning frameworks using both open and closed LLMs as baselines on DiscoveryBench and find that even the best system scores only 25%. Our benchmark, thus, illustrates the challenges in autonomous data-driven discovery and serves as a valuable resource for the community to make progress. △ Less

Submitted 1 July, 2024; originally announced July 2024.

Comments: Website: https://github.com/allenai/discoverybench

Coding Speech through Vocal Tract Kinematics

Authors: Cheol Jun Cho, Peter Wu, Tejas S. Prabhune, Dhruv Agarwal, Gopala K. Anumanchipalli

Abstract: Vocal tract articulation is a natural, grounded control space of speech production. The spatiotemporal coordination of articulators combined with the vocal source shapes intelligible speech sounds to enable effective spoken communication. Based on this physiological grounding of speech, we propose a new framework of neural encoding-decoding of speech -- Speech Articulatory Coding (SPARC). SPARC co… ▽ More Vocal tract articulation is a natural, grounded control space of speech production. The spatiotemporal coordination of articulators combined with the vocal source shapes intelligible speech sounds to enable effective spoken communication. Based on this physiological grounding of speech, we propose a new framework of neural encoding-decoding of speech -- Speech Articulatory Coding (SPARC). SPARC comprises an articulatory analysis model that infers articulatory features from speech audio, and an articulatory synthesis model that synthesizes speech audio from articulatory features. The articulatory features are kinematic traces of vocal tract articulators and source features, which are intuitively interpretable and controllable, being the actual physical interface of speech production. An additional speaker identity encoder is jointly trained with the articulatory synthesizer to inform the voice texture of individual speakers. By training on large-scale speech data, we achieve a fully intelligible, high-quality articulatory synthesizer that generalizes to unseen speakers. Furthermore, the speaker embedding is effectively disentangled from articulations, which enables accent-perserving zero-shot voice conversion. To the best of our knowledge, this is the first demonstration of universal, high-performance articulatory inference and synthesis, suggesting the proposed framework as a powerful coding system of speech. △ Less

Submitted 14 December, 2024; v1 submitted 18 June, 2024; originally announced June 2024.

Journal ref: IEEE Journal of Selected Topics in Signal Processing, vol. 18, no. 8, pp. 1427-1440, Dec. 2024

EchoGuide: Active Acoustic Guidance for LLM-Based Eating Event Analysis from Egocentric Videos

Authors: Vineet Parikh, Saif Mahmud, Devansh Agarwal, Ke Li, François Guimbretière, Cheng Zhang

Abstract: Self-recording eating behaviors is a step towards a healthy lifestyle recommended by many health professionals. However, the current practice of manually recording eating activities using paper records or smartphone apps is often unsustainable and inaccurate. Smart glasses have emerged as a promising wearable form factor for tracking eating behaviors, but existing systems primarily identify when e… ▽ More Self-recording eating behaviors is a step towards a healthy lifestyle recommended by many health professionals. However, the current practice of manually recording eating activities using paper records or smartphone apps is often unsustainable and inaccurate. Smart glasses have emerged as a promising wearable form factor for tracking eating behaviors, but existing systems primarily identify when eating occurs without capturing details of the eating activities (E.g., what is being eaten). In this paper, we present EchoGuide, an application and system pipeline that leverages low-power active acoustic sensing to guide head-mounted cameras to capture egocentric videos, enabling efficient and detailed analysis of eating activities. By combining active acoustic sensing for eating detection with video captioning models and large-scale language models for retrieval augmentation, EchoGuide intelligently clips and analyzes videos to create concise, relevant activity records on eating. We evaluated EchoGuide with 9 participants in naturalistic settings involving eating activities, demonstrating high-quality summarization and significant reductions in video data needed, paving the way for practical, scalable eating activity tracking. △ Less

Submitted 31 July, 2024; v1 submitted 15 June, 2024; originally announced June 2024.

Comments: Accepted at ISWC '24

SonicID: User Identification on Smart Glasses with Acoustic Sensing

Authors: Ke Li, Devansh Agarwal, Ruidong Zhang, Vipin Gunda, Tianjun Mo, Saif Mahmud, Boao Chen, François Guimbretière, Cheng Zhang

Abstract: Smart glasses have become more prevalent as they provide an increasing number of applications for users. They store various types of private information or can access it via connections established with other devices. Therefore, there is a growing need for user identification on smart glasses. In this paper, we introduce a low-power and minimally-obtrusive system called SonicID, designed to authen… ▽ More Smart glasses have become more prevalent as they provide an increasing number of applications for users. They store various types of private information or can access it via connections established with other devices. Therefore, there is a growing need for user identification on smart glasses. In this paper, we introduce a low-power and minimally-obtrusive system called SonicID, designed to authenticate users on glasses. SonicID extracts unique biometric information from users by scanning their faces with ultrasonic waves and utilizes this information to distinguish between different users, powered by a customized binary classifier with the ResNet-18 architecture. SonicID can authenticate users by scanning their face for 0.06 seconds. A user study involving 40 participants confirms that SonicID achieves a true positive rate of 97.4%, a false positive rate of 4.3%, and a balanced accuracy of 96.6% using just 1 minute of training data collected for each new user. This performance is relatively consistent across different remounting sessions and days. Given this promising performance, we further discuss the potential applications of SonicID and methods to improve its performance in the future. △ Less

Submitted 24 October, 2024; v1 submitted 12 June, 2024; originally announced June 2024.

Comments: 27 pages, 6 tables, 9 figures

MunchSonic: Tracking Fine-grained Dietary Actions through Active Acoustic Sensing on Eyeglasses

Authors: Saif Mahmud, Devansh Agarwal, Ashwin Ajit, Qikang Liang, Thalia Viranda, Francois Guimbretiere, Cheng Zhang

Abstract: We introduce MunchSonic, an AI-powered active acoustic sensing system integrated into eyeglasses to track fine-grained dietary actions. MunchSonic emits inaudible ultrasonic waves from the eyeglass frame, with the reflected signals capturing detailed positions and movements of body parts, including the mouth, jaw, arms, and hands involved in eating. These signals are processed by a deep learning p… ▽ More We introduce MunchSonic, an AI-powered active acoustic sensing system integrated into eyeglasses to track fine-grained dietary actions. MunchSonic emits inaudible ultrasonic waves from the eyeglass frame, with the reflected signals capturing detailed positions and movements of body parts, including the mouth, jaw, arms, and hands involved in eating. These signals are processed by a deep learning pipeline to classify six actions: hand-to-mouth movements for food intake, chewing, drinking, talking, face-hand touching, and other activities (null). In an unconstrained study with 12 participants, MunchSonic achieved a 93.5% macro F1-score in a user-independent evaluation with a 2-second resolution in tracking these actions, also demonstrating its effectiveness in tracking eating episodes and food intake frequency within those episodes. △ Less

Submitted 2 August, 2024; v1 submitted 31 May, 2024; originally announced May 2024.

Comments: 8 pages, 7 figures

Conversational Agents to Facilitate Deliberation on Harmful Content in WhatsApp Groups

Authors: Dhruv Agarwal, Farhana Shahid, Aditya Vashistha

Abstract: WhatsApp groups have become a hotbed for the propagation of harmful content including misinformation, hate speech, polarizing content, and rumors, especially in Global South countries. Given the platform's end-to-end encryption, moderation responsibilities lie on group admins and members, who rarely contest such content. Another approach is fact-checking, which is unscalable, and can only contest… ▽ More WhatsApp groups have become a hotbed for the propagation of harmful content including misinformation, hate speech, polarizing content, and rumors, especially in Global South countries. Given the platform's end-to-end encryption, moderation responsibilities lie on group admins and members, who rarely contest such content. Another approach is fact-checking, which is unscalable, and can only contest factual content (e.g., misinformation) but not subjective content (e.g., hate speech). Drawing on recent literature, we explore deliberation -- open and inclusive discussion -- as an alternative. We investigate the role of a conversational agent in facilitating deliberation on harmful content in WhatsApp groups. We conducted semi-structured interviews with 21 Indian WhatsApp users, employing a design probe to showcase an example agent. Participants expressed the need for anonymity and recommended AI assistance to reduce the effort required in deliberation. They appreciated the agent's neutrality but pointed out the futility of deliberation in echo chamber groups. Our findings highlight design tensions for such an agent, including privacy versus group dynamics and freedom of speech in private spaces. We discuss the efficacy of deliberation using deliberative theory as a lens, compare deliberation with moderation and fact-checking, and provide design recommendations for future such systems. Ultimately, this work advances CSCW by offering insights into designing deliberative systems for combating harmful content in private group chats on social media. △ Less

Submitted 16 August, 2024; v1 submitted 30 May, 2024; originally announced May 2024.

Comments: Accepted at CSCW 2024

Prompt Leakage effect and defense strategies for multi-turn LLM interactions

Authors: Divyansh Agarwal, Alexander R. Fabbri, Ben Risher, Philippe Laban, Shafiq Joty, Chien-Sheng Wu

Abstract: Prompt leakage poses a compelling security and privacy threat in LLM applications. Leakage of system prompts may compromise intellectual property, and act as adversarial reconnaissance for an attacker. A systematic evaluation of prompt leakage threats and mitigation strategies is lacking, especially for multi-turn LLM interactions. In this paper, we systematically investigate LLM vulnerabilities a… ▽ More Prompt leakage poses a compelling security and privacy threat in LLM applications. Leakage of system prompts may compromise intellectual property, and act as adversarial reconnaissance for an attacker. A systematic evaluation of prompt leakage threats and mitigation strategies is lacking, especially for multi-turn LLM interactions. In this paper, we systematically investigate LLM vulnerabilities against prompt leakage for 10 closed- and open-source LLMs, across four domains. We design a unique threat model which leverages the LLM sycophancy effect and elevates the average attack success rate (ASR) from 17.7% to 86.2% in a multi-turn setting. Our standardized setup further allows dissecting leakage of specific prompt contents such as task instructions and knowledge documents. We measure the mitigation effect of 7 black-box defense strategies, along with finetuning an open-source model to defend against leakage attempts. We present different combination of defenses against our threat model, including a cost analysis. Our study highlights key takeaways for building secure LLM applications and provides directions for research in multi-turn LLM interactions △ Less

Submitted 29 July, 2024; v1 submitted 24 April, 2024; originally announced April 2024.

ActSonic: Recognizing Everyday Activities from Inaudible Acoustic Wave Around the Body

Authors: Saif Mahmud, Vineet Parikh, Qikang Liang, Ke Li, Ruidong Zhang, Ashwin Ajit, Vipin Gunda, Devansh Agarwal, François Guimbretière, Cheng Zhang

Abstract: We present ActSonic, an intelligent, low-power active acoustic sensing system integrated into eyeglasses that can recognize 27 different everyday activities (e.g., eating, drinking, toothbrushing) from inaudible acoustic waves around the body. It requires only a pair of miniature speakers and microphones mounted on each hinge of the eyeglasses to emit ultrasonic waves, creating an acoustic aura ar… ▽ More We present ActSonic, an intelligent, low-power active acoustic sensing system integrated into eyeglasses that can recognize 27 different everyday activities (e.g., eating, drinking, toothbrushing) from inaudible acoustic waves around the body. It requires only a pair of miniature speakers and microphones mounted on each hinge of the eyeglasses to emit ultrasonic waves, creating an acoustic aura around the body. The acoustic signals are reflected based on the position and motion of various body parts, captured by the microphones, and analyzed by a customized self-supervised deep learning framework to infer the performed activities on a remote device such as a mobile phone or cloud server. ActSonic was evaluated in user studies with 19 participants across 19 households to track its efficacy in everyday activity recognition. Without requiring any training data from new users (leave-one-participant-out evaluation), ActSonic detected 27 activities, achieving an average F1-score of 86.6% in fully unconstrained scenarios and 93.4% in prompted settings at participants' homes. △ Less

Submitted 25 November, 2024; v1 submitted 22 April, 2024; originally announced April 2024.

Comments: Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, Volume 8, Issue 4, November 2024, IMWUT/UbiComp 2025

Ring-a-Pose: A Ring for Continuous Hand Pose Tracking

Authors: Tianhong Catherine Yu, Guilin Hu, Ruidong Zhang, Hyunchul Lim, Saif Mahmud, Chi-Jung Lee, Ke Li, Devansh Agarwal, Shuyang Nie, Jinseok Oh, François Guimbretière, Cheng Zhang

Abstract: We present Ring-a-Pose, a single untethered ring that tracks continuous 3D hand poses. Located in the center of the hand, the ring emits an inaudible acoustic signal that each hand pose reflects differently. Ring-a-Pose imposes minimal obtrusions on the hand, unlike multi-ring or glove systems. It is not affected by the choice of clothing that may cover wrist-worn systems. In a series of three use… ▽ More We present Ring-a-Pose, a single untethered ring that tracks continuous 3D hand poses. Located in the center of the hand, the ring emits an inaudible acoustic signal that each hand pose reflects differently. Ring-a-Pose imposes minimal obtrusions on the hand, unlike multi-ring or glove systems. It is not affected by the choice of clothing that may cover wrist-worn systems. In a series of three user studies with a total of 30 participants, we evaluate Ring-a-Pose's performance on pose tracking and micro-finger gesture recognition. Without collecting any training data from a user, Ring-a-Pose tracks continuous hand poses with a joint error of 14.1mm. The joint error decreases to 10.3mm for fine-tuned user-dependent models. Ring-a-Pose recognizes 7-class micro-gestures with a 90.60% and 99.27% accuracy for user-independent and user-dependent models, respectively. Furthermore, the ring exhibits promising performance when worn on any finger. Ring-a-Pose enables the future of smart rings to track and recognize hand poses using relatively low-power acoustic sensing. △ Less

Submitted 11 November, 2024; v1 submitted 19 April, 2024; originally announced April 2024.

GenVideo: One-shot Target-image and Shape Aware Video Editing using T2I Diffusion Models

Authors: Sai Sree Harsha, Ambareesh Revanur, Dhwanit Agarwal, Shradha Agrawal

Abstract: Video editing methods based on diffusion models that rely solely on a text prompt for the edit are hindered by the limited expressive power of text prompts. Thus, incorporating a reference target image as a visual guide becomes desirable for precise control over edit. Also, most existing methods struggle to accurately edit a video when the shape and size of the object in the target image differ fr… ▽ More Video editing methods based on diffusion models that rely solely on a text prompt for the edit are hindered by the limited expressive power of text prompts. Thus, incorporating a reference target image as a visual guide becomes desirable for precise control over edit. Also, most existing methods struggle to accurately edit a video when the shape and size of the object in the target image differ from the source object. To address these challenges, we propose "GenVideo" for editing videos leveraging target-image aware T2I models. Our approach handles edits with target objects of varying shapes and sizes while maintaining the temporal consistency of the edit using our novel target and shape aware InvEdit masks. Further, we propose a novel target-image aware latent noise correction strategy during inference to improve the temporal consistency of the edits. Experimental analyses indicate that GenVideo can effectively handle edits with objects of varying shapes, where existing approaches fail. △ Less

Submitted 18 April, 2024; originally announced April 2024.

Comments: CVPRw 2024

Data-driven Discovery with Large Generative Models

Authors: Bodhisattwa Prasad Majumder, Harshit Surana, Dhruv Agarwal, Sanchaita Hazra, Ashish Sabharwal, Peter Clark

Abstract: With the accumulation of data at an unprecedented rate, its potential to fuel scientific discovery is growing exponentially. This position paper urges the Machine Learning (ML) community to exploit the capabilities of large generative models (LGMs) to develop automated systems for end-to-end data-driven discovery -- a paradigm encompassing the search and verification of hypotheses purely from a se… ▽ More With the accumulation of data at an unprecedented rate, its potential to fuel scientific discovery is growing exponentially. This position paper urges the Machine Learning (ML) community to exploit the capabilities of large generative models (LGMs) to develop automated systems for end-to-end data-driven discovery -- a paradigm encompassing the search and verification of hypotheses purely from a set of provided datasets, without the need for additional data collection or physical experiments. We first outline several desiderata for an ideal data-driven discovery system. Then, through DATAVOYAGER, a proof-of-concept utilizing GPT-4, we demonstrate how LGMs fulfill several of these desiderata -- a feat previously unattainable -- while also highlighting important limitations in the current system that open up opportunities for novel ML research. We contend that achieving accurate, reliable, and robust end-to-end discovery systems solely through the current capabilities of LGMs is challenging. We instead advocate for fail-proof tool integration, along with active user moderation through feedback mechanisms, to foster data-driven scientific discoveries with efficiency and reproducibility. △ Less

Submitted 21 February, 2024; originally announced February 2024.

EchoWrist: Continuous Hand Pose Tracking and Hand-Object Interaction Recognition Using Low-Power Active Acoustic Sensing On a Wristband

Authors: Chi-Jung Lee, Ruidong Zhang, Devansh Agarwal, Tianhong Catherine Yu, Vipin Gunda, Oliver Lopez, James Kim, Sicheng Yin, Boao Dong, Ke Li, Mose Sakashita, Francois Guimbretiere, Cheng Zhang

Abstract: Our hands serve as a fundamental means of interaction with the world around us. Therefore, understanding hand poses and interaction context is critical for human-computer interaction. We present EchoWrist, a low-power wristband that continuously estimates 3D hand pose and recognizes hand-object interactions using active acoustic sensing. EchoWrist is equipped with two speakers emitting inaudible s… ▽ More Our hands serve as a fundamental means of interaction with the world around us. Therefore, understanding hand poses and interaction context is critical for human-computer interaction. We present EchoWrist, a low-power wristband that continuously estimates 3D hand pose and recognizes hand-object interactions using active acoustic sensing. EchoWrist is equipped with two speakers emitting inaudible sound waves toward the hand. These sound waves interact with the hand and its surroundings through reflections and diffractions, carrying rich information about the hand's shape and the objects it interacts with. The information captured by the two microphones goes through a deep learning inference system that recovers hand poses and identifies various everyday hand activities. Results from the two 12-participant user studies show that EchoWrist is effective and efficient at tracking 3D hand poses and recognizing hand-object interactions. Operating at 57.9mW, EchoWrist is able to continuously reconstruct 20 3D hand joints with MJEDE of 4.81mm and recognize 12 naturalistic hand-object interactions with 97.6% accuracy. △ Less

Submitted 29 March, 2024; v1 submitted 30 January, 2024; originally announced January 2024.

One Style Does Not Regulate All: Moderation Practices in Public and Private WhatsApp Groups

Authors: Farhana Shahid, Dhruv Agarwal, Aditya Vashistha

Abstract: WhatsApp is the largest social media platform in the Global South and is a virulent force in global misinformation and political propaganda. Due to end-to-end encryption WhatsApp can barely review any content and mostly rely on volunteer moderation by group admins. Yet, little is known about how WhatsApp group admins manage their groups, what factors and values influence moderation decisions, and… ▽ More WhatsApp is the largest social media platform in the Global South and is a virulent force in global misinformation and political propaganda. Due to end-to-end encryption WhatsApp can barely review any content and mostly rely on volunteer moderation by group admins. Yet, little is known about how WhatsApp group admins manage their groups, what factors and values influence moderation decisions, and what challenges they face while managing their groups. To fill this gap, we interviewed admins of 32 diverse groups and reviewed content from 30 public groups in India and Bangladesh. We observed notable differences in the formation, members' behavior, and moderation of public versus private groups, as well as in how WhatsApp admins operate compared to those on other platforms. We used Baumrind's typology of 'parenting styles' as a lens to examine how admins enact care and control during volunteer moderation. We identified four styles based on how caring and controlling the admins are and discuss design recommendations to help them better manage problematic content in WhatsApp groups. △ Less

Submitted 2 January, 2025; v1 submitted 15 January, 2024; originally announced January 2024.

Active Foundational Models for Fault Diagnosis of Electrical Motors

Authors: Sriram Anbalagan, Sai Shashank GP, Deepesh Agarwal, Balasubramaniam Natarajan, Babji Srinivasan

Abstract: Fault detection and diagnosis of electrical motors are of utmost importance in ensuring the safe and reliable operation of several industrial systems. Detection and diagnosis of faults at the incipient stage allows corrective actions to be taken in order to reduce the severity of faults. The existing data-driven deep learning approaches for machine fault diagnosis rely extensively on huge amounts… ▽ More Fault detection and diagnosis of electrical motors are of utmost importance in ensuring the safe and reliable operation of several industrial systems. Detection and diagnosis of faults at the incipient stage allows corrective actions to be taken in order to reduce the severity of faults. The existing data-driven deep learning approaches for machine fault diagnosis rely extensively on huge amounts of labeled samples, where annotations are expensive and time-consuming. However, a major portion of unlabeled condition monitoring data is not exploited in the training process. To overcome this limitation, we propose a foundational model-based Active Learning framework that utilizes less amount of labeled samples, which are most informative and harnesses a large amount of available unlabeled data by effectively combining Active Learning and Contrastive Self-Supervised Learning techniques. It consists of a transformer network-based backbone model trained using an advanced nearest-neighbor contrastive self-supervised learning method. This approach empowers the backbone to learn improved representations of samples derived from raw, unlabeled vibration data. Subsequently, the backbone can undergo fine-tuning to address a range of downstream tasks, both within the same machines and across different machines. The effectiveness of the proposed methodology has been assessed through the fine-tuning of the backbone for multiple target tasks using three distinct machine-bearing fault datasets. The experimental evaluation demonstrates a superior performance as compared to existing state-of-the-art fault diagnosis methods with less amount of labeled data. △ Less

Submitted 26 November, 2023; originally announced November 2023.

Comments: 30 pages, 2 figures, 7 tables

Eye Disease Prediction using Ensemble Learning and Attention on OCT Scans

Authors: Gauri Naik, Nandini Narvekar, Dimple Agarwal, Nishita Nandanwar, Himangi Pande

Abstract: Eye diseases have posed significant challenges for decades, but advancements in technology have opened new avenues for their detection and treatment. Machine learning and deep learning algorithms have become instrumental in this domain, particularly when combined with Optical Coherent Technology (OCT) imaging. We propose a novel method for efficient detection of eye diseases from OCT images. Our t… ▽ More Eye diseases have posed significant challenges for decades, but advancements in technology have opened new avenues for their detection and treatment. Machine learning and deep learning algorithms have become instrumental in this domain, particularly when combined with Optical Coherent Technology (OCT) imaging. We propose a novel method for efficient detection of eye diseases from OCT images. Our technique enables the classification of patients into disease free (normal eyes) or affected by specific conditions such as Choroidal Neovascularization (CNV), Diabetic Macular Edema (DME), or Drusen. In this work, we introduce an end to end web application that utilizes machine learning and deep learning techniques for efficient eye disease prediction. The application allows patients to submit their raw OCT scanned images, which undergo segmentation using a trained custom UNet model. The segmented images are then fed into an ensemble model, comprising InceptionV3 and Xception networks, enhanced with a self attention layer. This self attention approach leverages the feature maps of individual models to achieve improved classification accuracy. The ensemble model's output is aggregated to predict and classify various eye diseases. Extensive experimentation and optimization have been conducted to ensure the application's efficiency and optimal performance. Our results demonstrate the effectiveness of the proposed approach in accurate eye disease prediction. The developed web application holds significant potential for early detection and timely intervention, thereby contributing to improved eye healthcare outcomes. △ Less

Submitted 26 November, 2023; originally announced November 2023.

Comments: Full paper accepted at FICC (Springer) 2024

Bring Your Own KG: Self-Supervised Program Synthesis for Zero-Shot KGQA

Authors: Dhruv Agarwal, Rajarshi Das, Sopan Khosla, Rashmi Gangadharaiah

Abstract: We present BYOKG, a universal question-answering (QA) system that can operate on any knowledge graph (KG), requires no human-annotated training data, and can be ready to use within a day -- attributes that are out-of-scope for current KGQA systems. BYOKG draws inspiration from the remarkable ability of humans to comprehend information present in an unseen KG through exploration -- starting at rand… ▽ More We present BYOKG, a universal question-answering (QA) system that can operate on any knowledge graph (KG), requires no human-annotated training data, and can be ready to use within a day -- attributes that are out-of-scope for current KGQA systems. BYOKG draws inspiration from the remarkable ability of humans to comprehend information present in an unseen KG through exploration -- starting at random nodes, inspecting the labels of adjacent nodes and edges, and combining them with their prior world knowledge. In BYOKG, exploration leverages an LLM-backed symbolic agent that generates a diverse set of query-program exemplars, which are then used to ground a retrieval-augmented reasoning procedure to predict programs for arbitrary questions. BYOKG is effective over both small- and large-scale graphs, showing dramatic gains in QA accuracy over a zero-shot baseline of 27.89 and 58.02 F1 on GrailQA and MetaQA, respectively. On GrailQA, we further show that our unsupervised BYOKG outperforms a supervised in-context learning method, demonstrating the effectiveness of exploration. Lastly, we find that performance of BYOKG reliably improves with continued exploration as well as improvements in the base LLM, notably outperforming a state-of-the-art fine-tuned model by 7.08 F1 on a sub-sampled zero-shot split of GrailQA. △ Less

Submitted 21 May, 2024; v1 submitted 13 November, 2023; originally announced November 2023.

Art or Artifice? Large Language Models and the False Promise of Creativity

Authors: Tuhin Chakrabarty, Philippe Laban, Divyansh Agarwal, Smaranda Muresan, Chien-Sheng Wu

Abstract: Researchers have argued that large language models (LLMs) exhibit high-quality writing capabilities from blogs to stories. However, evaluating objectively the creativity of a piece of writing is challenging. Inspired by the Torrance Test of Creative Thinking (TTCT), which measures creativity as a process, we use the Consensual Assessment Technique [3] and propose the Torrance Test of Creative Writ… ▽ More Researchers have argued that large language models (LLMs) exhibit high-quality writing capabilities from blogs to stories. However, evaluating objectively the creativity of a piece of writing is challenging. Inspired by the Torrance Test of Creative Thinking (TTCT), which measures creativity as a process, we use the Consensual Assessment Technique [3] and propose the Torrance Test of Creative Writing (TTCW) to evaluate creativity as a product. TTCW consists of 14 binary tests organized into the original dimensions of Fluency, Flexibility, Originality, and Elaboration. We recruit 10 creative writers and implement a human assessment of 48 stories written either by professional authors or LLMs using TTCW. Our analysis shows that LLM-generated stories pass 3-10X less TTCW tests than stories written by professionals. In addition, we explore the use of LLMs as assessors to automate the TTCW evaluation, revealing that none of the LLMs positively correlate with the expert assessments. △ Less

Submitted 8 March, 2024; v1 submitted 25 September, 2023; originally announced September 2023.

Comments: ACM CHI 2024

Foundational Models for Fault Diagnosis of Electrical Motors

Authors: Sriram Anbalagan, Deepesh Agarwal, Balasubramaniam Natarajan, Babji Srinivasan

Abstract: A majority of recent advancements related to the fault diagnosis of electrical motors are based on the assumption that training and testing data are drawn from the same distribution. However, the data distribution can vary across different operating conditions during real-world operating scenarios of electrical motors. Consequently, this assumption limits the practical implementation of existing s… ▽ More A majority of recent advancements related to the fault diagnosis of electrical motors are based on the assumption that training and testing data are drawn from the same distribution. However, the data distribution can vary across different operating conditions during real-world operating scenarios of electrical motors. Consequently, this assumption limits the practical implementation of existing studies for fault diagnosis, as they rely on fully labelled training data spanning all operating conditions and assume a consistent distribution. This is because obtaining a large number of labelled samples for several machines across different fault cases and operating scenarios may be unfeasible. In order to overcome the aforementioned limitations, this work proposes a framework to develop a foundational model for fault diagnosis of electrical motors. It involves building a neural network-based backbone to learn high-level features using self-supervised learning, and then fine-tuning the backbone to achieve specific objectives. The primary advantage of such an approach is that the backbone can be fine-tuned to achieve a wide variety of target tasks using very less amount of training data as compared to traditional supervised learning methodologies. The empirical evaluation demonstrates the effectiveness of the proposed approach by obtaining more than 90\% classification accuracy by fine-tuning the backbone not only across different types of fault scenarios or operating conditions, but also across different machines. This illustrates the promising potential of the proposed approach for cross-machine fault diagnosis tasks in real-world applications. △ Less

Submitted 31 July, 2023; originally announced July 2023.

Comments: 7 pages, 1 figure, 5 tables, submitted to IEEE PESGRE 2023

SPLAL: Similarity-based pseudo-labeling with alignment loss for semi-supervised medical image classification

Authors: Md Junaid Mahmood, Pranaw Raj, Divyansh Agarwal, Suruchi Kumari, Pravendra Singh

Abstract: Medical image classification is a challenging task due to the scarcity of labeled samples and class imbalance caused by the high variance in disease prevalence. Semi-supervised learning (SSL) methods can mitigate these challenges by leveraging both labeled and unlabeled data. However, SSL methods for medical image classification need to address two key challenges: (1) estimating reliable pseudo-la… ▽ More Medical image classification is a challenging task due to the scarcity of labeled samples and class imbalance caused by the high variance in disease prevalence. Semi-supervised learning (SSL) methods can mitigate these challenges by leveraging both labeled and unlabeled data. However, SSL methods for medical image classification need to address two key challenges: (1) estimating reliable pseudo-labels for the images in the unlabeled dataset and (2) reducing biases caused by class imbalance. In this paper, we propose a novel SSL approach, SPLAL, that effectively addresses these challenges. SPLAL leverages class prototypes and a weighted combination of classifiers to predict reliable pseudo-labels over a subset of unlabeled images. Additionally, we introduce alignment loss to mitigate model biases toward majority classes. To evaluate the performance of our proposed approach, we conduct experiments on two publicly available medical image classification benchmark datasets: the skin lesion classification (ISIC 2018) and the blood cell classification dataset (BCCD). The experimental results empirically demonstrate that our approach outperforms several state-of-the-art SSL methods over various evaluation metrics. Specifically, our proposed approach achieves a significant improvement over the state-of-the-art approach on the ISIC 2018 dataset in both Accuracy and F1 score, with relative margins of 2.24\% and 11.40\%, respectively. Finally, we conduct extensive ablation experiments to examine the contribution of different components of our approach, validating its effectiveness. △ Less

Submitted 10 July, 2023; originally announced July 2023.

Comments: Under Review

Machine Reading Comprehension using Case-based Reasoning

Authors: Dung Thai, Dhruv Agarwal, Mudit Chaudhary, Wenlong Zhao, Rajarshi Das, Manzil Zaheer, Jay-Yoon Lee, Hannaneh Hajishirzi, Andrew McCallum

Abstract: We present an accurate and interpretable method for answer extraction in machine reading comprehension that is reminiscent of case-based reasoning (CBR) from classical AI. Our method (CBR-MRC) builds upon the hypothesis that contextualized answers to similar questions share semantic similarities with each other. Given a test question, CBR-MRC first retrieves a set of similar cases from a nonparame… ▽ More We present an accurate and interpretable method for answer extraction in machine reading comprehension that is reminiscent of case-based reasoning (CBR) from classical AI. Our method (CBR-MRC) builds upon the hypothesis that contextualized answers to similar questions share semantic similarities with each other. Given a test question, CBR-MRC first retrieves a set of similar cases from a nonparametric memory and then predicts an answer by selecting the span in the test context that is most similar to the contextualized representations of answers in the retrieved cases. The semi-parametric nature of our approach allows it to attribute a prediction to the specific set of evidence cases, making it a desirable choice for building reliable and debuggable QA systems. We show that CBR-MRC provides high accuracy comparable with large reader models and outperforms baselines by 11.5 and 8.4 EM on NaturalQuestions and NewsQA, respectively. Further, we demonstrate the ability of CBR-MRC in identifying not just the correct answer tokens but also the span with the most relevant supporting evidence. Lastly, we observe that contexts for certain question types show higher lexical diversity than others and find that CBR-MRC is robust to these variations while performance using fully-parametric methods drops. △ Less

Submitted 5 December, 2023; v1 submitted 24 May, 2023; originally announced May 2023.

Comments: 9 pages, 2 figures

LLMs as Factual Reasoners: Insights from Existing Benchmarks and Beyond

Authors: Philippe Laban, Wojciech Kryściński, Divyansh Agarwal, Alexander R. Fabbri, Caiming Xiong, Shafiq Joty, Chien-Sheng Wu

Abstract: With the recent appearance of LLMs in practical settings, having methods that can effectively detect factual inconsistencies is crucial to reduce the propagation of misinformation and improve trust in model outputs. When testing on existing factual consistency benchmarks, we find that a few large language models (LLMs) perform competitively on classification benchmarks for factual inconsistency de… ▽ More With the recent appearance of LLMs in practical settings, having methods that can effectively detect factual inconsistencies is crucial to reduce the propagation of misinformation and improve trust in model outputs. When testing on existing factual consistency benchmarks, we find that a few large language models (LLMs) perform competitively on classification benchmarks for factual inconsistency detection compared to traditional non-LLM methods. However, a closer analysis reveals that most LLMs fail on more complex formulations of the task and exposes issues with existing evaluation benchmarks, affecting evaluation precision. To address this, we propose a new protocol for inconsistency detection benchmark creation and implement it in a 10-domain benchmark called SummEdits. This new benchmark is 20 times more cost-effective per sample than previous benchmarks and highly reproducible, as we estimate inter-annotator agreement at about 0.9. Most LLMs struggle on SummEdits, with performance close to random chance. The best-performing model, GPT-4, is still 8\% below estimated human performance, highlighting the gaps in LLMs' ability to reason about facts and detect inconsistencies when they occur. △ Less

Submitted 23 May, 2023; originally announced May 2023.

CoralStyleCLIP: Co-optimized Region and Layer Selection for Image Editing

Authors: Ambareesh Revanur, Debraj Basu, Shradha Agrawal, Dhwanit Agarwal, Deepak Pai

Abstract: Edit fidelity is a significant issue in open-world controllable generative image editing. Recently, CLIP-based approaches have traded off simplicity to alleviate these problems by introducing spatial attention in a handpicked layer of a StyleGAN. In this paper, we propose CoralStyleCLIP, which incorporates a multi-layer attention-guided blending strategy in the feature space of StyleGAN2 for obtai… ▽ More Edit fidelity is a significant issue in open-world controllable generative image editing. Recently, CLIP-based approaches have traded off simplicity to alleviate these problems by introducing spatial attention in a handpicked layer of a StyleGAN. In this paper, we propose CoralStyleCLIP, which incorporates a multi-layer attention-guided blending strategy in the feature space of StyleGAN2 for obtaining high-fidelity edits. We propose multiple forms of our co-optimized region and layer selection strategy to demonstrate the variation of time complexity with the quality of edits over different architectural intricacies while preserving simplicity. We conduct extensive experimental analysis and benchmark our method against state-of-the-art CLIP-based methods. Our findings suggest that CoralStyleCLIP results in high-quality edits while preserving the ease of use. △ Less

Submitted 8 March, 2023; originally announced March 2023.

Comments: CVPR 2023

AugTriever: Unsupervised Dense Retrieval and Domain Adaptation by Scalable Data Augmentation

Authors: Rui Meng, Ye Liu, Semih Yavuz, Divyansh Agarwal, Lifu Tu, Ning Yu, Jianguo Zhang, Meghana Bhat, Yingbo Zhou

Abstract: Dense retrievers have made significant strides in text retrieval and open-domain question answering. However, most of these achievements have relied heavily on extensive human-annotated supervision. In this study, we aim to develop unsupervised methods for improving dense retrieval models. We propose two approaches that enable annotation-free and scalable training by creating pseudo querydocument… ▽ More Dense retrievers have made significant strides in text retrieval and open-domain question answering. However, most of these achievements have relied heavily on extensive human-annotated supervision. In this study, we aim to develop unsupervised methods for improving dense retrieval models. We propose two approaches that enable annotation-free and scalable training by creating pseudo querydocument pairs: query extraction and transferred query generation. The query extraction method involves selecting salient spans from the original document to generate pseudo queries. On the other hand, the transferred query generation method utilizes generation models trained for other NLP tasks, such as summarization, to produce pseudo queries. Through extensive experimentation, we demonstrate that models trained using these augmentation methods can achieve comparable, if not better, performance than multiple strong dense baselines. Moreover, combining these strategies leads to further improvements, resulting in superior performance of unsupervised dense retrieval, unsupervised domain adaptation and supervised finetuning, benchmarked on both BEIR and ODQA datasets. Code and datasets are publicly available at https://github.com/salesforce/AugTriever. △ Less

Submitted 29 October, 2024; v1 submitted 17 December, 2022; originally announced December 2022.

Comments: DCAI24, October 25, 2024, Boise, ID