Mihir Parmar

Yizhong Wang, Swaroop Mishra, Pegah Alipoormolabashi et al. · allen-ai, amazon-science

How well can NLP models generalize to a variety of unseen tasks when provided with task instructions? To address this question, we first introduce Super-NaturalInstructions, a benchmark of 1,616 diverse NLP tasks and their expert-written instructions. Our collection covers 76 distinct task types, including but not limited to classification, extraction, infilling, sequence tagging, text rewriting, and text composition. This large and diverse collection of tasks enables rigorous benchmarking of cross-task generalization under instructions -- training models to follow instructions on a subset of tasks and evaluating them on the remaining unseen ones. Furthermore, we build Tk-Instruct, a transformer model trained to follow a variety of in-context instructions (plain language task definitions or k-shot examples). Our experiments show that Tk-Instruct outperforms existing instruction-following models such as InstructGPT by over 9% on our benchmark despite being an order of magnitude smaller. We further analyze generalization as a function of various scaling parameters, such as the number of observed tasks, the number of instances per task, and model sizes. We hope our dataset and model facilitate future progress towards more general-purpose NLP models.

Mihir Parmar, Palash Goyal, Xin Liu et al.

Recently, decomposing complex problems into simple subtasks--a crucial part of human-like natural planning--to solve the given problem has significantly boosted the performance of large language models (LLMs). However, leveraging such planning structures during post-training to boost the performance of smaller open-source LLMs remains underexplored. Motivated by this, we introduce PLAN-TUNING, a unified post-training framework that (i) distills synthetic task decompositions (termed "planning trajectories") from large-scale LLMs and (ii) fine-tunes smaller models via supervised and reinforcement-learning objectives designed to mimic these planning processes to improve complex reasoning. On GSM8k and the MATH benchmarks, plan-tuned models outperform strong baselines by an average $\sim7\%$. Furthermore, plan-tuned models show better generalization capabilities on out-of-domain datasets, with average $\sim10\%$ and $\sim12\%$ performance improvements on OlympiadBench and AIME 2024, respectively. Our detailed analysis demonstrates how planning trajectories improves complex reasoning capabilities, showing that PLAN-TUNING is an effective strategy for improving task-specific performance of smaller LLMs.

Himanshu Gupta, Shreyas Verma, Ujjwala Anantheswaran et al. · gatech

Multi-modal Large Language Models (MLLMs) exhibit impressive problem-solving abilities in various domains, but their visual comprehension and abstract reasoning skills remain under-evaluated. To this end, we present PolyMATH, a challenging benchmark aimed at evaluating the general cognitive reasoning abilities of MLLMs. PolyMATH comprises 5,000 manually collected high-quality images of cognitive textual and visual challenges across 10 distinct categories, including pattern recognition, spatial reasoning, and relative reasoning. We conducted a comprehensive, and quantitative evaluation of 15 MLLMs using four diverse prompting strategies, including Chain-of-Thought and Step-Back. The best scores achieved on PolyMATH are ~41%, ~36%, and ~27%, obtained by Claude-3.5 Sonnet, GPT-4o and Gemini-1.5 Pro respectively - highlighting the logical and visual complexity of these questions. A further fine-grained error analysis reveals that these models struggle to understand spatial relations and perform drawn-out, high-level reasoning. This is further strengthened by our ablation study estimating MLLM performance when given textual descriptions in place of diagrams. As evidenced by ~4% improvement over textual descriptions as opposed to actual images, we discover that models do not truly comprehend visual diagrams and the spatial information therein, and are thus prone to logical errors. Finally, we evaluate the OpenAI o1 models and find that their performance only matches the human baseline, highlighting the difficulty of the benchmark. The results on PolyMATH highlight the room for improvement in multi-modal reasoning and provide unique insights to guide the development of future MLLMs.

Mihir Parmar, Swaroop Mishra, Mor Geva et al. · deepmind

In recent years, progress in NLU has been driven by benchmarks. These benchmarks are typically collected by crowdsourcing, where annotators write examples based on annotation instructions crafted by dataset creators. In this work, we hypothesize that annotators pick up on patterns in the crowdsourcing instructions, which bias them to write many similar examples that are then over-represented in the collected data. We study this form of bias, termed instruction bias, in 14 recent NLU benchmarks, showing that instruction examples often exhibit concrete patterns, which are propagated by crowdworkers to the collected data. This extends previous work (Geva et al., 2019) and raises a new concern of whether we are modeling the dataset creator's instructions, rather than the task. Through a series of experiments, we show that, indeed, instruction bias can lead to overestimation of model performance, and that models struggle to generalize beyond biases originating in the crowdsourcing instructions. We further analyze the influence of instruction bias in terms of pattern frequency and model size, and derive concrete recommendations for creating future NLU benchmarks.

Pruthvi Patel, Swaroop Mishra, Mihir Parmar et al.

Large Language Models (LMs) have achieved state-of-the-art performance on many Natural Language Processing (NLP) benchmarks. With the growing number of new benchmarks, we build bigger and more complex LMs. However, building new LMs may not be an ideal option owing to the cost, time and environmental impact associated with it. We explore an alternative route: can we modify data by expressing it in terms of the model's strengths, so that a question becomes easier for models to answer? We investigate if humans can decompose a hard question into a set of simpler questions that are relatively easier for models to solve. We analyze a range of datasets involving various forms of reasoning and find that it is indeed possible to significantly improve model performance (24% for GPT3 and 29% for RoBERTa-SQuAD along with a symbolic calculator) via decomposition. Our approach provides a viable option to involve people in NLP research in a meaningful way. Our findings indicate that Human-in-the-loop Question Decomposition (HQD) can potentially provide an alternate path to building large LMs. Code and data is available at https://github.com/Pruthvi98/QuestionDecomposition

Nemika Tyagi, Mihir Parmar, Mohith Kulkarni et al.

Solving grid puzzles involves a significant amount of logical reasoning. Hence, it is a good domain to evaluate the reasoning capability of a model which can then guide us to improve the reasoning ability of models. However, most existing works evaluate only the final predicted answer of a puzzle, without delving into an in-depth analysis of the LLMs' reasoning chains (such as where they falter) or providing any finer metrics to evaluate them. Since LLMs may rely on simple heuristics or artifacts to predict the final answer, it is crucial to evaluate the generated reasoning chain beyond overall correctness measures, for accurately evaluating the reasoning abilities of LLMs. To this end, we first develop GridPuzzle, an evaluation dataset comprising 274 grid-based puzzles with different complexities. Second, we propose a new error taxonomy derived from manual analysis of reasoning chains from LLMs including GPT-4, Claude-3, Gemini, Mistral, and Llama-2. Then, we develop an LLM-based framework for large-scale subjective evaluation (i.e., identifying errors) and an objective metric, PuzzleEval, to evaluate the correctness of reasoning chains. Evaluating reasoning chains from LLMs leads to several interesting findings. We further show that existing prompting methods used for enhancing models' reasoning abilities do not improve performance on GridPuzzle. This highlights the importance of understanding fine-grained errors and presents a challenge for future research to enhance LLMs' puzzle-solving abilities by developing methods that address these errors. Data and source code are available at https://github.com/Mihir3009/GridPuzzle.

Mihir Parmar, Hanieh Deilamsalehy, Franck Dernoncourt et al.

Extractive summarization plays a pivotal role in natural language processing due to its wide-range applications in summarizing diverse content efficiently, while also being faithful to the original content. Despite significant advancement achieved in extractive summarization by Large Language Models (LLMs), these summaries frequently exhibit incoherence. An important aspect of the coherent summary is its readability for intended users. Although there have been many datasets and benchmarks proposed for creating coherent extractive summaries, none of them currently incorporate user intent to improve coherence in extractive summarization. Motivated by this, we propose a systematically created human-annotated dataset consisting of coherent summaries for five publicly available datasets and natural language user feedback, offering valuable insights into how to improve coherence in extractive summaries. We utilize this dataset for aligning LLMs through supervised fine-tuning with natural language human feedback to enhance the coherence of their generated summaries. Preliminary experiments with Falcon-40B and Llama-2-13B show significant performance improvements (~10% Rouge-L) in terms of producing coherent summaries. We further utilize human feedback to benchmark results over instruction-tuned models such as FLAN-T5 which resulted in several interesting findings. Data and source code are available at https://github.com/Mihir3009/Extract-AI.

Man Luo, Shrinidhi Kumbhar, Ming shen et al. · amazon-science

Logical reasoning is fundamental for humans yet presents a substantial challenge in the domain of Artificial Intelligence. Initially, researchers used Knowledge Representation and Reasoning (KR) systems that did not scale and required non-trivial manual effort. Recently, the emergence of large language models (LLMs) has demonstrated the ability to overcome various limitations of formal Knowledge Representation (KR) systems. Consequently, there's a growing interest in using LLMs for logical reasoning via natural language. This work strives to understand the proficiency of LLMs in logical reasoning by offering a brief review of the latest progress in this area; with a focus on the logical reasoning datasets, tasks, and the methods adopted to utilize LLMs for reasoning. To offer a thorough analysis, we have compiled a benchmark titled LogiGLUE. This includes 24 varied datasets encompassing deductive, abductive, and inductive reasoning. Utilizing LogiGLUE as a foundation, we have trained an instruction fine-tuned language model, resulting in LogiT5. We study single-task training, multi-task training, and "chain-of-thought" knowledge distillation fine-tuning technique to assess the performance of model across the different logical reasoning categories. We also assess various LLMs using LogiGLUE, and the findings indicate that LLMs excel most in abductive reasoning, followed by deductive reasoning, while they are least effective at inductive reasoning. We aim to shed light on the capabilities and potential pathways for enhancing logical reasoning proficiency in LLMs, paving the way for more advanced and nuanced developments in this critical field.

Mihir Parmar, Aakanksha Naik, Himanshu Gupta et al.

Many large language models (LLMs) for medicine have largely been evaluated on short texts, and their ability to handle longer sequences such as a complete electronic health record (EHR) has not been systematically explored. Assessing these models on long sequences is crucial since prior work in the general domain has demonstrated performance degradation of LLMs on longer texts. Motivated by this, we introduce LongBoX, a collection of seven medical datasets in text-to-text format, designed to investigate model performance on long sequences. Preliminary experiments reveal that both medical LLMs (e.g., BioGPT) and strong general domain LLMs (e.g., FLAN-T5) struggle on this benchmark. We further evaluate two techniques designed for long-sequence handling: (i) local-global attention, and (ii) Fusion-in-Decoder (FiD). Our results demonstrate mixed results with long-sequence handling - while scores on some datasets increase, there is substantial room for improvement. We hope that LongBoX facilitates the development of more effective long-sequence techniques for the medical domain. Data and source code are available at https://github.com/Mihir3009/LongBoX.

Himanshu Gupta, Kevin Scaria, Ujjwala Anantheswaran et al. · gatech

The rapid advancement of large language models (LLMs) has sparked interest in data synthesis techniques, aiming to generate diverse and high-quality synthetic datasets. However, these synthetic datasets often suffer from a lack of diversity and added noise. In this paper, we present TarGEN, a multi-step prompting strategy for generating high-quality synthetic datasets utilizing a LLM. An advantage of TarGEN is its seedless nature; it does not require specific task instances, broadening its applicability beyond task replication. We augment TarGEN with a method known as self-correction empowering LLMs to rectify inaccurately labeled instances during dataset creation, ensuring reliable labels. To assess our technique's effectiveness, we emulate 8 tasks from the SuperGLUE benchmark and finetune various language models, including encoder-only, encoder-decoder, and decoder-only models on both synthetic and original training sets. Evaluation on the original test set reveals that models trained on datasets generated by TarGEN perform approximately 1-2% points better than those trained on original datasets (82.84% via syn. vs. 81.12% on og. using Flan-T5). When incorporating instruction tuning, the performance increases to 84.54% on synthetic data vs. 81.49% on original data by Flan-T5. A comprehensive analysis of the synthetic dataset compared to the original dataset reveals that the synthetic dataset demonstrates similar or higher levels of dataset complexity and diversity. Furthermore, the synthetic dataset displays a bias level that aligns closely with the original dataset. Finally, when pre-finetuned on our synthetic SuperGLUE dataset, T5-3B yields impressive results on the OpenLLM leaderboard, surpassing the model trained on the Self-Instruct dataset by 4.14% points. We hope that TarGEN can be helpful for quality data generation and reducing the human efforts to create complex benchmarks.

Ayushi Agarwal, Nisarg Patel, Neeraj Varshney et al. · amazon-science

Though state-of-the-art (SOTA) NLP systems have achieved remarkable performance on a variety of language understanding tasks, they primarily focus on questions that have a correct and a definitive answer. However, in real-world applications, users often ask questions that don't have a definitive answer. Incorrectly answering such questions certainly hampers a system's reliability and trustworthiness. Can SOTA models accurately identify such questions and provide a reasonable response? To investigate the above question, we introduce QnotA, a dataset consisting of five different categories of questions that don't have definitive answers. Furthermore, for each QnotA instance, we also provide a corresponding QA instance i.e. an alternate question that ''can be'' answered. With this data, we formulate three evaluation tasks that test a system's ability to 'identify', 'distinguish', and 'justify' QnotA questions. Through comprehensive experiments, we show that even SOTA models including GPT-3 and Flan T5 do not fare well on these tasks and lack considerably behind the human performance baseline. We conduct a thorough analysis which further leads to several interesting findings. Overall, we believe our work and findings will encourage and facilitate further research in this important area and help develop more robust models.

Kirby Kuznia, Swaroop Mishra, Mihir Parmar et al.

Despite the success of large pre-trained language models (LMs) such as Codex, they show below-par performance on the larger and more complicated programming related questions. We show that LMs benefit from the summarized version of complicated questions. Our findings show that superfluous information often present in problem description such as human characters, background stories, and names (which are included to help humans in understanding a task) does not help models in understanding a task. To this extent, we create a meta-dataset from the frequently used APPS dataset and the newly created CodeContests dataset for the program synthesis task. Our meta-dataset consists of human and synthesized summaries of the long and complicated programming questions. Experimental results on Codex show that our proposed approach outperforms baseline by 8.13% on the APPS dataset and 11.88% on the CodeContests dataset on average in terms of strict accuracy. Our analysis shows that summaries significantly improve performance for introductory (9.86%) and interview (11.48%) programming questions. However, it shows improvement by a small margin (~ 2%) for competitive programming questions, implying scope for future research in this direction.

Mihir Parmar, Swaroop Mishra, Mirali Purohit et al.

Single-task models have proven pivotal in solving specific tasks; however, they have limitations in real-world applications where multi-tasking is necessary and domain shifts are exhibited. Recently, instructional prompts have shown significant improvement towards multi-task generalization; however, the effect of instructional prompts and Multi-Task Learning (MTL) has not been systematically studied in the biomedical domain. Motivated by this, this paper explores the impact of instructional prompts for biomedical MTL. We introduce the BoX, a collection of 32 instruction tasks for Biomedical NLP across (X) various categories. Using this meta-dataset, we propose a unified model termed In-BoXBART, that can jointly learn all tasks of the BoX without any task-specific modules. To the best of our knowledge, this is the first attempt to propose a unified model in the biomedical domain and use instructions to achieve generalization across several biomedical tasks. Experimental results indicate that the proposed model: 1) outperforms the single-task baseline by ~3% and multi-task (without instruction) baseline by ~18% on an average, and 2) shows ~23% improvement compared to the single-task baseline in few-shot learning (i.e., 32 instances per task) on an average. Our analysis indicates that there is significant room for improvement across tasks in the BoX, implying the scope for future research direction.

Ravsehaj Singh Puri, Swaroop Mishra, Mihir Parmar et al.

Recently introduced instruction-paradigm empowers non-expert users to leverage NLP resources by defining a new task in natural language. Instruction-tuned models have significantly outperformed multitask learning models (without instruction); however they are far from state-of-the-art task-specific models. Conventional approaches to improve model performance via creating datasets with large number of task instances or architectural changes in the model may not be feasible for non-expert users. However, they can write alternate instructions to represent an instruction task. Is Instruction-augmentation helpful? We augment a subset of tasks in the expanded version of NATURAL INSTRUCTIONS with additional instructions and find that it significantly improves model performance (up to 35%), especially in the low-data regime. Our results indicate that an additional instruction can be equivalent to ~200 data samples on average across tasks.

Man Luo, Sharad Saxena, Swaroop Mishra et al.

Table Question Answering (TQA) is an important but under-explored task. Most of the existing QA datasets are in unstructured text format and only few of them use tables as the context. To the best of our knowledge, none of TQA datasets exist in the biomedical domain where tables are frequently used to present information. In this paper, we first curate a table question answering dataset, BioTABQA, using 22 templates and the context from a biomedical textbook on differential diagnosis. BioTABQA can not only be used to teach a model how to answer questions from tables but also evaluate how a model generalizes to unseen questions, an important scenario for biomedical applications. To achieve the generalization evaluation, we divide the templates into 17 training and 5 cross-task evaluations. Then, we develop two baselines using single and multi-tasks learning on BioTABQA. Furthermore, we explore instructional learning, a recent technique showing impressive generalizing performance. Experimental results show that our instruction-tuned model outperforms single and multi-task baselines on an average by ~23% and ~6% across various evaluation settings, and more importantly, instruction-tuned model outperforms baselines by ~5% on cross-tasks.

Neeraj Varshney, Agneet Chatterjee, Mihir Parmar et al. · amazon-science

Large Language Models (LLMs) have achieved remarkable performance across a wide variety of natural language tasks; however, their large size makes their inference slow and computationally expensive. Focusing on this problem, we propose to instruction tune LLMs with additional explicit losses from the intermediate layers (LITE) and show that it enables these layers to acquire 'good' generation ability without affecting the generation ability of the final layer. We perform 'dynamic confidence-based early exiting' at token level from the intermediate layers which improves the efficiency of text generation without compromising the quality of the generation. We conduct comprehensive experiments by instruction tuning LLaMA-2 models on the Alpaca dataset and holistically evaluate on four different human-instruction test sets. We show that dynamic early exiting achieves consistent and considerable inference computation cost improvements (37.86% for 7B and 46.35% for 13B model) while maintaining the generation quality of the responses. We further conduct a thorough analysis of the results over several important aspects, such as comparing the semantic similarity of the outputs and dissecting the efficiency improvements by comparing the number of tokens generated in the output. In summary, our work contributes to improving the efficiency of LLM inference while maintaining the generation quality, a crucial step en route to enabling their widespread adoption.

Srija Macherla, Man Luo, Mihir Parmar et al.

Dialogue systems for Automatic Differential Diagnosis (ADD) have a wide range of real-life applications. These dialogue systems are promising for providing easy access and reducing medical costs. Building end-to-end ADD dialogue systems requires dialogue training datasets. However, to the best of our knowledge, there is no publicly available ADD dialogue dataset in English (although non-English datasets exist). Driven by this, we introduce MDDial, the first differential diagnosis dialogue dataset in English which can aid to build and evaluate end-to-end ADD dialogue systems. Additionally, earlier studies present the accuracy of diagnosis and symptoms either individually or as a combined weighted score. This method overlooks the connection between the symptoms and the diagnosis. We introduce a unified score for the ADD system that takes into account the interplay between symptoms and diagnosis. This score also indicates the system's reliability. To the end, we train two moderate-size of language models on MDDial. Our experiments suggest that while these language models can perform well on many natural language understanding tasks, including dialogue tasks in the general domain, they struggle to relate relevant symptoms and disease and thus have poor performance on MDDial. MDDial will be released publicly to aid the study of ADD dialogue research.

Rui Meng, Bhavana Dalvi Mishra, Jiefeng Chen et al.

Autonomous research agents produce competitive solutions and professional-looking manuscripts, yet their outputs contain verifiability failures undetectable by surface-level evaluation: fabricated citations, unreproducible scores, and method descriptions that diverge from the implementation. We address this through three contributions. First, Chain-of-Evidence (CoE), a verifiability framework requiring every claim to be traceable to its evidence source. Second, ScientistOne, an end-to-end autonomous research system that maintains evidence chains by construction throughout literature review, solution discovery, and paper writing. Third, CoE Audit, a post-hoc audit whose four integrity checks -- score verification, specification violation, reference verification, and method-code alignment -- apply uniformly to all systems. Across 75 papers spanning five systems and five frontier research tasks, every baseline exhibits at least one systematic failure mode: hallucinated reference rates reach 21%, score verification passes in as few as 42% of papers, and method-code alignment ranges from 20% to 80%. ScientistOne achieves zero hallucinated references (0/337), perfect score verification (12/12), and the highest method-code alignment (14/15), while matching or exceeding human expert performance on all five tasks. ScientistOne further generalizes to six additional tasks spanning medical imaging, fine-grained recognition, 3D perception, and language modeling, achieving state-of-the-art on Parameter Golf and gold medals on MLE-Bench tasks where baselines fail entirely.

Ruiyao Xu, Mihir Parmar, Tiankai Yang et al.

Learning from preference-based feedback has become an effective approach for aligning LLMs across diverse tasks. However, high-quality human-annotated preference data remains expensive and scarce. Existing methods address this challenge through either self-rewarding, which scales by using purely AI-generated labels but risks unreliability, or active learning, which ensures quality through oracle annotation but cannot fully leverage unlabeled data. In this paper, we present CoAct, a novel framework that synergistically combines self-rewarding and active learning through strategic human-AI collaboration. CoAct leverages self-consistency to identify both reliable self-labeled data and samples that require oracle verification. Additionally, oracle feedback guides the model to generate new instructions within its solvable capability. Evaluated on three reasoning benchmarks across two model families, CoAct achieves average improvements of +13.25% on GSM8K, +8.19% on MATH, and +13.16% on WebInstruct, consistently outperforming all baselines.

Md Atik Ahamed, Mihir Parmar, Palash Goyal et al.

We introduce TFRBench, the first benchmark designed to evaluate the reasoning capabilities of forecasting systems. Traditionally, time-series forecasting has been evaluated solely on numerical accuracy, treating foundation models as ``black boxes.'' Unlike existing benchmarks, TFRBench provides a protocol for evaluating the reasoning generated by forecasting systems--specifically their analysis of cross-channel dependencies, trends, and external events. To enable this, we propose a systematic multi-agent framework that utilizes an iterative verification loop to synthesize numerically grounded reasoning traces. Spanning ten datasets across five domains, our evaluation confirms that this reasoning is causally effective; useful for evaluation; and prompting LLMs with our generated traces significantly improves forecasting accuracy compared to direct numerical prediction (e.g., avg. $\sim40.2\%\to56.6\%)$, validating the quality of our reasoning. Conversely, benchmarking experiments reveal that off-the-shelf LLMs consistently struggle with both reasoning (lower LLM-as-a-Judge scores) and numerical forecasting, frequently failing to capture domain-specific dynamics. TFRBench thus establishes a new standard for interpretable, reasoning-based evaluation in time-series forecasting. Our benchmark is available at: https://tfrbench.github.io

Mihir Parmar, Nisarg Patel, Neeraj Varshney et al. · amazon-science

Recently developed large language models (LLMs) have been shown to perform remarkably well on a wide range of language understanding tasks. But, can they really "reason" over the natural language? This question has been receiving significant research attention and many reasoning skills such as commonsense, numerical, and qualitative have been studied. However, the crucial skill pertaining to 'logical reasoning' has remained underexplored. Existing work investigating this reasoning ability of LLMs has focused only on a couple of inference rules (such as modus ponens and modus tollens) of propositional and first-order logic. Addressing the above limitation, we comprehensively evaluate the logical reasoning ability of LLMs on 25 different reasoning patterns spanning over propositional, first-order, and non-monotonic logics. To enable systematic evaluation, we introduce LogicBench, a natural language question-answering dataset focusing on the use of a single inference rule. We conduct detailed analysis with a range of LLMs such as GPT-4, ChatGPT, Gemini, Llama-2, and Mistral using chain-of-thought prompting. Experimental results show that existing LLMs do not fare well on LogicBench; especially, they struggle with instances involving complex reasoning and negations. Furthermore, they sometimes overlook contextual information necessary for reasoning to arrive at the correct conclusion. We believe that our work and findings facilitate future research for evaluating and enhancing the logical reasoning ability of LLMs. Data and code are available at https://github.com/Mihir3009/LogicBench.

Palash Goyal, Mihir Parmar, Yiwen Song et al.

Automated peer review has evolved from simple text classification to structured feedback generation. However, current state-of-the-art systems still struggle with "surface-level" critiques: they excel at summarizing content but often fail to accurately assess novelty and significance or identify deep methodological flaws because they evaluate papers in a vacuum, lacking the external context a human expert possesses. In this paper, we introduce ScholarPeer, a search-enabled multi-agent framework designed to emulate the cognitive processes of a senior researcher. ScholarPeer employs a dual-stream process of context acquisition and active verification. It dynamically constructs a domain narrative using a historian agent, identifies missing comparisons via a baseline scout, and verifies claims through a multi-aspect Q&A engine, grounding the critique in live web-scale literature. We evaluate ScholarPeer on DeepReview-13K and the results demonstrate that ScholarPeer achieves significant win-rates against state-of-the-art approaches in side-by-side evaluations and reduces the gap to human-level diversity.

Ishani Mondal, Yiwen Song, Mihir Parmar et al.

Long-form visual storytelling requires maintaining continuity across shots, including consistent characters, stable environments, and smooth scene transitions. While existing generative models can produce strong individual frames, they fail to preserve such continuity, leading to appearance changes, inconsistent backgrounds, and abrupt scene shifts. We introduce CANVAS (Continuity-Aware Narratives via Visual Agentic Storyboarding), a multi-agent framework that explicitly plans visual continuity in multi-shot narratives. CANVAS enforces coherence through character continuity, persistent background anchors, and location-aware scene planning for smooth transitions within the same setting We evaluate CANVAS on two storyboard generation benchmarks ST-BENCH and ViStoryBench and introduce a new challenging benchmark HardContinuityBench for long-range narrative consistency. CANVAS consistently outperforms the best-performing baseline, improving background continuity by 21.6%, character consistency by 9.6% and props consistency by 7.6%.

Sarkar Snigdha Sarathi Das, Palash Goyal, Mihir Parmar et al.

Pre-trained Time Series Foundational Models (TSFMs) represent a significant advance, capable of forecasting diverse time series with complex characteristics, including varied seasonalities, trends, and long-range dependencies. Despite their primary goal of universal time series forecasting, their efficacy is far from uniform; divergent training protocols and data sources cause individual TSFMs to exhibit highly variable performance across different forecasting tasks, domains, and horizons. Leveraging this complementary expertise by arbitrating existing TSFM outputs presents a compelling strategy, yet this remains a largely unexplored area of research. In this paper, we conduct a thorough examination of how different TSFMs exhibit specialized performance profiles across various forecasting settings, and how we can effectively leverage this behavior in arbitration between different time series models. We specifically analyze how factors such as model selection and forecast horizon distribution can influence the efficacy of arbitration strategies. Based on this analysis, we propose Synapse, a novel arbitration framework for TSFMs. Synapse is designed to dynamically leverage a pool of TSFMs, assign and adjust predictive weights based on their relative, context-dependent performance, and construct a robust forecast distribution by adaptively sampling from the output quantiles of constituent models. Experimental results demonstrate that Synapse consistently outperforms other popular ensembling techniques as well as individual TSFMs, demonstrating Synapse's efficacy in time series forecasting.

Aswin RRV, Jacob Dineen, Divij Handa et al.

Recent advances in test-time scaling have led to the emergence of thinking LLMs that exhibit self-reflective behaviors and multi-step reasoning. While RL drives this self-improvement paradigm, a recent study (Gandhi et al., 2025) shows that RL alone does not truly instill these new reasoning abilities - it merely draws out behaviors already present in the base models. This raises a question: How can we train the models that don't exhibit such thinking behavior to develop it in the first place? To this end, we propose ThinkTuning, a GRPO-based interactive training approach where we augment the rollouts of a student model with the guidance from a teacher model. A simple idea from classroom practice inspires our method: a teacher poses a problem, lets the student try an answer, then gives corrective feedback -- enough to point the mind in the right direction and then show the solution. Each piece of feedback reshapes the student's thoughts, leading them to arrive at the correct solution. Similarly, we find that this type of implicit supervision through feedback from a teacher model of the same size improves the reasoning capabilities of the student model. In particular, on average, our method shows a 3.85% improvement over zero-shot baselines across benchmarks, and on MATH-500, AIME and GPQA-Diamond it shows 2.08%, 2.23% and 3.99% improvements over the vanilla-GRPO baseline. Source code is available at https://github.com/3rdAT/ThinkTuning.

Alireza Salemi, Mihir Parmar, Palash Goyal et al.

The rapid advancement of Large Language Models (LLMs) has opened new opportunities in data science, yet their practical deployment is often constrained by the challenge of discovering relevant data within large heterogeneous data lakes. Existing methods struggle with this: single-agent systems are quickly overwhelmed by large, heterogeneous files in the large data lakes, while multi-agent systems designed based on a master-slave paradigm depend on a rigid central controller for task allocation that requires precise knowledge of each sub-agent's capabilities. To address these limitations, we propose a novel multi-agent communication paradigm inspired by the blackboard architecture for traditional AI models. In this framework, a central agent posts requests to a shared blackboard, and autonomous subordinate agents -- either responsible for a partition of the data lake or general information retrieval -- volunteer to respond based on their capabilities. This design improves scalability and flexibility by eliminating the need for a central coordinator to have prior knowledge of all sub-agents' expertise. We evaluate our method on three benchmarks that require explicit data discovery: KramaBench and modified versions of DS-Bench and DA-Code to incorporate data discovery. Experimental results demonstrate that the blackboard architecture substantially outperforms baselines, including RAG and the master-slave multi-agent paradigm, achieving between 13% to 57% relative improvement in end-to-end task success and up to a 9% relative gain in F1 score for data discovery over the best-performing baselines across both proprietary and open-source LLMs. Our findings establish the blackboard paradigm as a scalable and generalizable communication framework for multi-agent systems.

Sarkar Snigdha Sarathi Das, Palash Goyal, Mihir Parmar et al.

Time series forecasting is not just numerical extrapolation, but often requires reasoning with unstructured contextual data such as news or events. While specialized Time Series Foundation Models (TSFMs) excel at forecasting based on numerical patterns, they remain unaware to real-world textual signals. Conversely, while LLMs are emerging as zero-shot forecasters, their performance remains uneven across domains and contextual grounding. To bridge this gap, we introduce Nexus, a multi-agent forecasting framework that decomposes prediction into specialized stages: isolating macro-level and micro-level temporal fluctuations, and integrating contextual information when available before synthesizing a final forecast. This decomposition enables Nexus to adapt from seasonal signals to volatile, event-driven information without relying on external statistical anchors or monolithic prompting. We show that current-generation LLMs possess substantially stronger intrinsic forecasting ability than previously recognized, depending critically on how numerical and contextual reasoning are organized. Evaluated on data strictly succeeding LLM knowledge cutoffs spanning Zillow real estate metrics and volatile stock market equities, Nexus consistently matches or outperforms state-of-the-art TSFMs and strong LLM baselines. Beyond numerical accuracy, Nexus produces high-quality reasoning traces that explicitly show the fundamental drivers behind each forecast. Our results establish that real-world forecasting is an agentic reasoning problem extending well beyond only sequence modeling.

Minbeom Kim, Lesly Miculicich, Bhavana Dalvi Mishra et al.

As AI agents move from chat interfaces to systems that read private data, call tools, and execute multi-step workflows, guardrails become a last line of defense against concrete deployment harms. In these settings, guardrail failures are no longer merely answer-quality errors: they can leak secrets, authorize unsafe actions, or block legitimate work. The hardest failures are often contextual: whether an action is acceptable depends on local privacy norms, organizational policies, and user expectations that resist pre-deployment specification. This creates a practical gap: guardrails must adapt to their own operating environments, yet deployment feedback is typically limited to sparse, noisy user-reported failures, and repeated fine-tuning is often impractical. To address this gap, we propose LiSA (Lifelong Safety Adaptation), a conservative policy induction framework that improves a fixed base guardrail through structured memory. LiSA converts occasional failures into reusable policy abstractions so that sparse reports can generalize beyond individual cases, adds conflict-aware local rules to prevent overgeneralization in mixed-label contexts, and applies evidence-aware confidence gating via a posterior lower bound, so that memory reuse scales with accumulated evidence rather than empirical accuracy alone. Across PrivacyLens+, ConFaide+, and AgentHarm, LiSA consistently outperforms strong memory-based baselines under sparse feedback, remains robust under noisy user feedback even at 20% label-flip rates, and pushes the latency--performance frontier beyond backbone model scaling. Ultimately, LiSA offers a practical path to secure AI agents against the unpredictable long tail of real-world edge risks.

Nisarg Patel, Mohith Kulkarni, Mihir Parmar et al.

As Large Language Models (LLMs) continue to exhibit remarkable performance in natural language understanding tasks, there is a crucial need to measure their ability for human-like multi-step logical reasoning. Existing logical reasoning evaluation benchmarks often focus primarily on simplistic single-step or multi-step reasoning with a limited set of inference rules. Furthermore, the lack of datasets for evaluating non-monotonic reasoning represents a crucial gap since it aligns more closely with human-like reasoning. To address these limitations, we propose Multi-LogiEval, a comprehensive evaluation dataset encompassing multi-step logical reasoning with various inference rules and depths. Multi-LogiEval covers three logic types--propositional, first-order, and non-monotonic--consisting of more than 30 inference rules and more than 60 of their combinations with various depths. Leveraging this dataset, we conduct evaluations on a range of LLMs including GPT-4, ChatGPT, Gemini-Pro, Yi, Orca, and Mistral, employing a zero-shot chain-of-thought. Experimental results show that there is a significant drop in the performance of LLMs as the reasoning steps/depth increases (average accuracy of ~68% at depth-1 to ~43% at depth-5). We further conduct a thorough investigation of reasoning chains generated by LLMs which reveals several important findings. We believe that Multi-LogiEval facilitates future research for evaluating and enhancing the logical reasoning ability of LLMs. Data is available at https://github.com/Mihir3009/Multi-LogiEval.

Neeraj Varshney, Mihir Parmar, Nisarg Patel et al.

Pre-training on large corpora of text enables the language models to acquire a vast amount of factual and commonsense knowledge which allows them to achieve remarkable performance on a variety of language understanding tasks. They typically acquire this knowledge by learning from the pre-training text and capturing certain patterns from it. However, real-world settings often present scenarios that do not abide by these patterns i.e. scenarios that break the common assumptions. Can state-of-the-art NLP models correctly reason over the contexts of such scenarios? Addressing the above question, in this paper, we investigate the ability of models to correctly reason over contexts that break the common assumptions. To this end, we first systematically create evaluation data in which each data instance consists of (a) a common assumption, (b) a context that follows the assumption, (c) a context that breaks the assumption, and (d) questions based on the contexts. Then, through evaluations on multiple models including GPT-3 and Flan T5, we show that while doing fairly well on contexts that follow the common assumptions, the models struggle to correctly reason over contexts that break those assumptions. Specifically, the performance gap is as high as 20% absolute points. Furthermore, we thoroughly analyze these results revealing several interesting findings. We believe our work and findings will encourage and facilitate further research in developing more robust models that can also reliably reason over contexts that break the common assumptions. Data is available at \url{https://github.com/nrjvarshney/break_the_common_assumptions}.

Md Atik Ahamed, Mihir Parmar, Palash Goyal et al.

Time Series Foundation Models (TSFMs) excel at numerical forecasting but operate as black boxes lacking qualitative reasoning. Conversely, applying LLMs directly to temporal data introduces a modality gap: text tokenizers fragment continuous numerical values, degrading mathematical relationships and exploding sequence lengths, leading to computational overhead. To resolve this, we introduce STRIDE (Strategic Time-series Reasoning Injected via Distilled Embeddings), a novel framework natively integrating LLM reasoning into the continuous embedding space of TSFMs. Instead of discrete tokens, STRIDE distills reasoning traces into a lightweight LLM, dynamically projecting its mean-pooled hidden states as a cross-modal prior into the target numerical encoder. The architecture is jointly optimized using cross-entropy and quantile losses. Evaluations demonstrate STRIDE establishes state-of-the-art numerical forecasting on GIFT-Eval (0.674 MASE, 0.454 CRPS) compared to TSFMs and exhibits superior in-domain and out-of-domain numerical as well as reasoning performance on TFRBench. Specifically, STRIDE acts as a plug-and-play enhancement, consistently improving diverse TSFMs (e.g., Chronos-2, Timer-S1) across various LLM configurations. Thus, injecting semantic reasoning as a continuous prior equips TSFMs with human-interpretable reasoning while fundamentally improving predictive accuracy.

Mingtian Tan, Mihir Parmar, Palash Goyal et al.

Large Language Models (LLMs) are increasingly applied to forecasting. To evaluate this capability while mitigating pre-training data contamination, several living benchmarks have been proposed. However, existing benchmarks either lack the multidimensional events essential for accurate forecasting due to data scarcity, or focus on relatively closed environments. To assess the predictive capabilities of LLMs in complex, real-world scenarios, we propose LEAF, the first living benchmark for event-augmented forecasting tasks, including future event probabilities, trend and time series forecasting. LEAF utilizes a recursive retrieval agent system paired with dual-agent cross-validation to provide comprehensive and relevant auxiliary text for forecasting. Evaluating state-of-the-art proprietary and open-weight LLMs, we find that these models can leverage signals extracted from complex events to enhance predictive performance. In the stock domain, we find that LLMs achieve better performance on equities they confidently identify as more predictable. Furthermore, the events demonstrate a strong correlation with the target equities. To this end, LEAF provides a necessary, dynamically updating testbed to continuously track and drive progress in event-driven forecasting tasks.

Aswin RRV, Jacob Dineen, Divij Handa et al.

The effectiveness of Reinforcement Learning (RL) in Large Language Models (LLMs) depends on the nature and diversity of the data used before and during RL. In particular, reasoning problems can often be approached in multiple ways that rely on different forms of reasoning, and exposure to only a limited range of such approaches in the training data may limit the effectiveness of RL. Motivated by this, we investigate using diverse self-generated data during mid-training as an intermediate step before RL training. Specifically, we adopt a bootstrapped data-generation framework guided by George Polya's problem-solving approaches for generating multiple variants of correct answers for each question in the training data, and then perform fine-tuning. We first provide a theoretical perspective on how mid-training on such data improves RL and explain how policy-gradient updates can incentivize combining multiple approaches. We then empirically demonstrate that RL-trained models initialized with our mid-training data achieve consistent improvements across various mathematical reasoning benchmarks and other OOD tasks like code generation and narrative reasoning. Overall, our investigative study shows that a language model learning multiple problem-solving approaches, through self-generated data helps subsequent RL.

Chien Van Nguyen, Xuan Shen, Ryan Aponte et al.

Small Language Models (SLMs) have become increasingly important due to their efficiency and performance to perform various language tasks with minimal computational resources, making them ideal for various settings including on-device, mobile, edge devices, among many others. In this article, we present a comprehensive survey on SLMs, focusing on their architectures, training techniques, and model compression techniques. We propose a novel taxonomy for categorizing the methods used to optimize SLMs, including model compression, pruning, and quantization techniques. We summarize the benchmark datasets that are useful for benchmarking SLMs along with the evaluation metrics commonly used. Additionally, we highlight key open challenges that remain to be addressed. Our survey aims to serve as a valuable resource for researchers and practitioners interested in developing and deploying small yet efficient language models.

Mihir Parmar, Xin Liu, Palash Goyal et al.

Recent agent frameworks and inference-time algorithms often struggle with complex planning problems due to limitations in verifying generated plans or reasoning and varying complexity of instances within a single task. Many existing methods for these tasks either perform task-level verification without considering constraints or apply inference-time algorithms without adapting to instance-level complexity. To address these limitations, we propose PlanGEN, a model-agnostic and easily scalable agent framework with three key components: constraint, verification, and selection agents. Specifically, our approach proposes constraint-guided iterative verification to enhance performance of inference-time algorithms--Best of N, Tree-of-Thought, and REBASE. In PlanGEN framework, the selection agent optimizes algorithm choice based on instance complexity, ensuring better adaptability to complex planning problems. Experimental results demonstrate significant improvements over the strongest baseline across multiple benchmarks, achieving state-of-the-art results on NATURAL PLAN ($\sim$8%$\uparrow$), OlympiadBench ($\sim$4%$\uparrow$), DocFinQA ($\sim$7%$\uparrow$), and GPQA ($\sim$1%$\uparrow$). Our key finding highlights that constraint-guided iterative verification improves inference-time algorithms, and adaptive selection further boosts performance on complex planning and reasoning problems.

Venkatesh Mishra, Bimsara Pathiraja, Mihir Parmar et al.

Reasoning abilities of LLMs have been a key focus in recent years. One challenging reasoning domain with interesting nuances is legal reasoning, which requires careful application of rules, and precedents while balancing deductive and analogical reasoning, and conflicts between rules. Although there have been a few works on using LLMs for legal reasoning, their focus has been on overall accuracy. In this paper, we dig deeper to do a step-by-step analysis and figure out where they commit errors. We use the college-level Multiple Choice Question-Answering (MCQA) task from the \textit{Civil Procedure} dataset and propose a new error taxonomy derived from initial manual analysis of reasoning chains with respect to several LLMs, including two objective measures: soundness and correctness scores. We then develop an LLM-based automated evaluation framework to identify reasoning errors and evaluate the performance of LLMs. The computation of soundness and correctness on the dataset using the auto-evaluator framework reveals several interesting insights. Furthermore, we show that incorporating the error taxonomy as feedback in popular prompting techniques marginally increases LLM performance. Our work will also serve as an evaluation framework that can be used in detailed error analysis of reasoning chains for logic-intensive complex tasks.

Souradeep Mukhopadhyay, Rishabh Baral, Nimeesh Mahajan et al.

Large language models (LLMs) such as GPT, Gemini, and Claude often appear adept at solving classic logic puzzles--but how much genuine reasoning underlies their answers? Recent evidence suggests that these models frequently rely on memorized templates rather than reasoning from first principles. When puzzles are slightly modified, their performance collapses, revealing a striking fragility. In particular, we asked: Have LLMs addressed these issues? To what extent? How about perturbations to other puzzles? Is there a general way of reformulating the prompt so that the models do better? To examine these things systematically, we introduce PHANTOM RECALL, a benchmark comprising 25 well-known logic puzzles and 149 carefully designed perturbations that preserve reasoning structure but alter superficial details and solutions. We evaluate eleven leading LLMs and identify a recurring failure mode--phantom recall--where models confidently reproduce memorized solutions or spurious rationales that no longer fit the altered scenario. To probe and mitigate this issue, we contribute three tools: (i) an automated logical-equivalence judge to detect reasoning mismatches, (ii) a taxonomy of fine-grained reasoning error categories, and (iii) a prompting-based mitigation framework guided by these categories. Despite near-perfect accuracy on unmodified puzzles, models significantly underperform humans on perturbed ones, exhibiting both phantom recall and over-elaboration. Our findings reveal a crucial limitation: LLMs often fail to re-reason when contextual cues shift--highlighting the gap between linguistic fluency and logical understanding.

Divij Handa, Mihir Parmar, Aswin RRV et al.

Repeated Sampling (RS) is a simple inference-time algorithm that has been shown to improve model performance on complex tasks. Although it is an effective way of scaling inference time, it often struggles to generate diverse solution candidates, frequently relying on the same underlying approach to solve the problem and thus producing redundant samples. To address this limitation, we propose a new inference algorithm, GuidedSampling, which decouples the exploration and generation phases during inference, increasing diversity of generated candidate solutions. The exploration phase identifies multiple concepts that can be utilized to solve the problem, while the generation phase applies a specific concept to provide final solution candidates. We first define the theoretical bounds of GuidedSampling and then empirically demonstrate that it improves the performance of base model at pass@50 by on an average ~21.6% across various benchmarks compared to RS. Furthermore, models trained on trajectories of GuidedSampling exhibit substantial performance improvements at pass@5 by on an average ~9.7%, compared to models trained on traditional RS. Additionally, models trained with GuidedSampling increases the average number of concepts per instance (1.67 -> 3.03), yielding a diverse set of candidates than traditional RS.

Lesly Miculicich, Mihir Parmar, Hamid Palangi et al.

The deployment of autonomous AI agents in sensitive domains, such as healthcare, introduces critical risks to safety, security, and privacy. These agents may deviate from user objectives, violate data handling policies, or be compromised by adversarial attacks. Mitigating these dangers necessitates a mechanism to formally guarantee that an agent's actions adhere to predefined safety constraints, a challenge that existing systems do not fully address. We introduce VeriGuard, a novel framework that provides formal safety guarantees for LLM-based agents through a dual-stage architecture designed for robust and verifiable correctness. The initial offline stage involves a comprehensive validation process. It begins by clarifying user intent to establish precise safety specifications. VeriGuard then synthesizes a behavioral policy and subjects it to both testing and formal verification to prove its compliance with these specifications. This iterative process refines the policy until it is deemed correct. Subsequently, the second stage provides online action monitoring, where VeriGuard operates as a runtime monitor to validate each proposed agent action against the pre-verified policy before execution. This separation of the exhaustive offline validation from the lightweight online monitoring allows formal guarantees to be practically applied, providing a robust safeguard that substantially improves the trustworthiness of LLM agents.

Ujjwala Anantheswaran, Himanshu Gupta, Mihir Parmar et al.

Event detection refers to identifying event occurrences in a text and comprises of two subtasks; event identification and classification. We present EDM3, a novel approach for Event Detection that formulates three generative tasks: identification, classification, and combined detection. We show that EDM3 helps to learn transferable knowledge that can be leveraged to perform Event Detection and its subtasks concurrently, mitigating the error propagation inherent in pipelined approaches. Unlike previous dataset- or domain-specific approaches, EDM3 utilizes the existing knowledge of language models, allowing it to be trained over any classification schema. We evaluate EDM3 on multiple event detection datasets: RAMS, WikiEvents, MAVEN, and MLEE, showing that EDM3 outperforms 1) single-task performance by 8.4% on average and 2) multi-task performance without instructional prompts by 2.4% on average. We obtain SOTA results on RAMS (71.3% vs. 65.1% F-1) and competitive performance on other datasets. We analyze our approach to demonstrate its efficacy in low-resource and multi-sentence settings. We also show the effectiveness of this approach on non-standard event configurations such as multi-word and multi-class event triggers. Overall, our results show that EDM3 is a promising approach for Event Detection that has the potential for real-world applications.

Mihir Parmar, Mathew Halm, Michael Posa

Frictional contact has been extensively studied as the core underlying behavior of legged locomotion and manipulation, and its nearly-discontinuous nature makes planning and control difficult even when an accurate model of the robot is available. Here, we present empirical evidence that learning an accurate model in the first place can be confounded by contact, as modern deep learning approaches are not designed to capture this non-smoothness. We isolate the effects of contact's non-smoothness by varying the mechanical stiffness of a compliant contact simulator. Even for a simple system, we find that stiffness alone dramatically degrades training processes, generalization, and data-efficiency. Our results raise serious questions about simulated testing environments which do not accurately reflect the stiffness of rigid robotic hardware. Significant additional investigation will be necessary to fully understand and mitigate these effects, and we suggest several avenues for future study.

Mihir Parmar, Ashwin Karthik Ambalavanan, Hong Guan et al.

Here we proposed an approach to analyze text classification methods based on the presence or absence of task-specific terms (and their synonyms) in the text. We applied this approach to study six different transfer-learning and unsupervised methods for screening articles relevant to COVID-19 vaccines and therapeutics. The analysis revealed that while a BERT model trained on search-engine results generally performed well, it miss-classified relevant abstracts that did not contain task-specific terms. We used this insight to create a more effective unsupervised ensemble.