Tirthankar Ghosal

Abhishek Kumar, Isha Motiyani, Tilak Kasturi et al.

Engineering diagrams pose a distinct challenge for vision-language models: unlike natural images or general documents, they encode information through dense spatial layouts, domain-specific symbols, and cross-references between visual callouts and structured parts tables. Despite their centrality to service, repair, and design workflows, there is no public benchmark for measuring VLM capabilities in this domain; existing datasets primarily focus on flowcharts, scientific figures, or business documents. To address this gap, we introduce Enginuity, the first open dataset and benchmark for evaluating VLMs on complex engineering diagrams. We define two tasks over a corpus of U.S. military service and repair manuals: structured parts-table extraction (Task 1) and free-form visual diagram question answering (VQA)(Task 2) for benchmarking. We evaluate four frontier VLMs (GPT-5.2 Chat, Claude Opus 4.7, Gemma 4, Qwen3-VL-32B-Instruct) under zero-shot and chain-of-thought prompting. On Task 1, models reach Recall@all of 0.61-0.87 but Token F1pen of only 0.03-0.18, exposing a systematic gap between part identification and description fidelity. Task 2 reveals a consistent factual-reasoning gap across all models. A supporting analysis shows that token-overlap metrics under-report model capability on technical descriptions by 2-6x relative to semantic similarity, motivating LLM-as-judge calibration for domain-specific evaluation. We release the dataset, annotations, evaluation harness, and per-sample model outputs to support a reproducible study of VLM capability on engineering content.

Sanjay Das, Ran Elgedawy, Ethan Seefried et al.

Operational safety in high-stakes domains such as industrial process control, autonomous, and safety-critical systems, demand reliable hazard identification. While large language models (LLMs) have shown promise in automating safety analysis tasks, single-turn, monolithic inference is brittle: it lacks the self-correction, deliberation, and contextual refinement that safety engineers apply iteratively. In this paper, we introduce HAZDIAL, a framework that investigates whether structured agentic dialogue-multi-agent, multi-turn interactions improves the quality of NLP- based hazard identification over single-pass baselines. We systematically compare two dialogue modalities: adversarial debate and constructive discussion, and propose an algorithm-based agentic interaction optimization. We evaluate all configurations against a curated golden dataset using standard classification metrics (accuracy, precision, recall, F1) and novel dialogue metrics. This work advances the intersection of dialogue systems, multi-agent reasoning, and AI safety, providing an empirical evidence for dialogue-driven hazard analysis.

Shashwat Sourav, Tanjin. He, Maria K. Y. Chan et al.

AI co-scientists are increasingly used for scientific discovery, but current evaluations still do not test them on a key task: moving from a concrete scientific or technological problem to a plausible, mechanism-grounded solution hypothesis. This gap is especially important in materials science and, in particular, battery research, where a useful proposal must identify the relevant failure mode, propose a credible intervention, and explain why that intervention should improve the target property. We introduce Matter to Mechanism, a benchmark for evaluating AI co-scientists on problem-to-hypothesis reasoning in materials science, with a focus on battery materials research. The benchmark contains 2,645 instances derived from scientific publications. Each instance includes a structured problem statement, a candidate solution hypothesis, an explicit reasoning trace, and domain-grounded annotations such as material system, component, failure mode, intervention, mechanism, target property, and claimed outcome. We also introduce a metric suite that measures reasoning fidelity, problem alignment, mechanistic specificity, novelty, plausibility, and problem decomposition quality, and combine them into a composite score. Using this framework, we evaluate several AI co-scientist systems and show that Matter to Mechanism reveals interpretable system differences that are only partially recovered by standard text-similarity metrics. We further show through adversarial stress tests that the aggregate score is more stable than individual metric dimensions under superficial gaming attacks.

Shashwat Sourav, Viktoriia Baibakova, Sanjay Das et al.

Knowledge graphs (KGs) can provide structured scientific context to language models, but it remains unclear which graph facts actually shape the generated hypotheses. We study KG-guided hypothesis generation for battery materials across Mistral-7B, Llama-3.1-70B, and Gemini 2.5 Flash. We perturb local KGs by varying density, ontology richness, topology, and control structure, and evaluate outputs with both provided-graph and fixed-reference metrics. Across models, KG utility is selective and model-dependent: graph context changes outputs, but no-KG outputs also recover substantial graph content from model priors. Compact top-k subgraphs often approximate full-KG behavior, including when claimed-outcome triples are held out. At the same time, compression is not unique to one semantic ranking rule, random and topology-based subsets can also recover much of the signal. These results support a redundancy-aware Compressive KG hypothesis: useful KG signal is often recoverable from compact, scientifically structured subgraphs rather than requiring the full local graph.

Yuan-Sen Ting, Tuan Dung Nguyen, Tirthankar Ghosal et al.

We present a comprehensive evaluation of proprietary and open-weights large language models using the first astronomy-specific benchmarking dataset. This dataset comprises 4,425 multiple-choice questions curated from the Annual Review of Astronomy and Astrophysics, covering a broad range of astrophysical topics. Our analysis examines model performance across various astronomical subfields and assesses response calibration, crucial for potential deployment in research environments. Claude-3.5-Sonnet outperforms competitors by up to 4.6 percentage points, achieving 85.0% accuracy. For proprietary models, we observed a universal reduction in cost every 3-to-12 months to achieve similar score in this particular astronomy benchmark. open-weights models have rapidly improved, with LLaMA-3-70b (80.6%) and Qwen-2-72b (77.7%) now competing with some of the best proprietary models. We identify performance variations across topics, with non-English-focused models generally struggling more in exoplanet-related fields, stellar astrophysics, and instrumentation related questions. These challenges likely stem from less abundant training data, limited historical context, and rapid recent developments in these areas. This pattern is observed across both open-weights and proprietary models, with regional dependencies evident, highlighting the impact of training data diversity on model performance in specialized scientific domains. Top-performing models demonstrate well-calibrated confidence, with correlations above 0.9 between confidence and correctness, though they tend to be slightly underconfident. The development for fast, low-cost inference of open-weights models presents new opportunities for affordable deployment in astronomy. The rapid progress observed suggests that LLM-driven research in astronomy may become feasible in the near future.

Sandeep Kumar, Tirthankar Ghosal, Asif Ekbal

To this date, the efficacy of the scientific publishing enterprise fundamentally rests on the strength of the peer review process. The journal editor or the conference chair primarily relies on the expert reviewers' assessment, identify points of agreement and disagreement and try to reach a consensus to make a fair and informed decision on whether to accept or reject a paper. However, with the escalating number of submissions requiring review, especially in top-tier Artificial Intelligence (AI) conferences, the editor/chair, among many other works, invests a significant, sometimes stressful effort to mitigate reviewer disagreements. Here in this work, we introduce a novel task of automatically identifying contradictions among reviewers on a given article. To this end, we introduce ContraSciView, a comprehensive review-pair contradiction dataset on around 8.5k papers (with around 28k review pairs containing nearly 50k review pair comments) from the open review-based ICLR and NeurIPS conferences. We further propose a baseline model that detects contradictory statements from the review pairs. To the best of our knowledge, we make the first attempt to identify disagreements among peer reviewers automatically. We make our dataset and code public for further investigations.

Sandeep Kumar, Tirthankar Ghosal, Vinayak Goyal et al.

The widespread adoption of Large Language Models (LLMs) and publicly available ChatGPT have marked a significant turning point in the integration of Artificial Intelligence (AI) into people's everyday lives. This study examines the ability of Large Language Models (LLMs) to generate future research ideas from scientific papers. Unlike tasks such as summarization or translation, idea generation lacks a clearly defined reference set or structure, making manual evaluation the default standard. However, human evaluation in this setting is extremely challenging ie: it requires substantial domain expertise, contextual understanding of the paper, and awareness of the current research landscape. This makes it time-consuming, costly, and fundamentally non-scalable, particularly as new LLMs are being released at a rapid pace. Currently, there is no automated evaluation metric specifically designed for this task. To address this gap, we propose two automated evaluation metrics: Idea Alignment Score (IAScore) and Idea Distinctness Index. We further conducted human evaluation to assess the novelty, relevance, and feasibility of the generated future research ideas. This investigation offers insights into the evolving role of LLMs in idea generation, highlighting both its capability and limitations. Our work contributes to the ongoing efforts in evaluating and utilizing language models for generating future research ideas. We make our datasets and codes publicly available

Jinchu Li, Yuan-Sen Ting, Alberto Accomazzi et al.

We construct a concept-object knowledge graph from the full astro-ph corpus through July 2025. Using an automated pipeline, we extract named astrophysical objects from OCR-processed papers, resolve them to SIMBAD identifiers, and link them to scientific concepts annotated in the source corpus. We then test whether historical graph structure can forecast new concept-object associations before they appear in print. Because the concepts are derived from clustering and therefore overlap semantically, we apply an inference-time concept-similarity smoothing step uniformly to all methods. Across four temporal cutoffs on a physically meaningful subset of concepts, an implicit-feedback matrix factorization model (alternating least squares, ALS) with smoothing outperforms the strongest neighborhood baseline (KNN using text-embedding concept similarity) by 16.8% on NDCG@100 (0.144 vs 0.123) and 19.8% on Recall@100 (0.175 vs 0.146), and exceeds the best recency heuristic by 96% and 88%, respectively. These results indicate that historical literature encodes predictive structure not captured by global heuristics or local neighborhood voting, suggesting a path toward tools that could help triage follow-up targets for scarce telescope time.

Charles O'Neill, Tirthankar Ghosal, Roberta Răileanu et al.

Generating novel and creative scientific hypotheses is a cornerstone in achieving Artificial General Intelligence. Large language and reasoning models have the potential to aid in the systematic creation, selection, and validation of scientifically informed hypotheses. However, current foundation models often struggle to produce scientific ideas that are both novel and feasible. One reason is the lack of a dedicated dataset that frames Scientific Hypothesis Generation (SHG) as a Natural Language Generation (NLG) task. In this paper, we introduce HypoGen, the first dataset of approximately 5500 structured problem-hypothesis pairs extracted from top-tier computer science conferences structured with a Bit-Flip-Spark schema, where the Bit is the conventional assumption, the Spark is the key insight or conceptual leap, and the Flip is the resulting counterproposal. HypoGen uniquely integrates an explicit Chain-of-Reasoning component that reflects the intellectual process from Bit to Flip. We demonstrate that framing hypothesis generation as conditional language modelling, with the model fine-tuned on Bit-Flip-Spark and the Chain-of-Reasoning (and where, at inference, we only provide the Bit), leads to improvements in the overall quality of the hypotheses. Our evaluation employs automated metrics and LLM judge rankings for overall quality assessment. We show that by fine-tuning on our HypoGen dataset we improve the novelty, feasibility, and overall quality of the generated hypotheses. The HypoGen dataset is publicly available at huggingface.co/datasets/UniverseTBD/hypogen-dr1.

Sandeep Kumar, Mohit Sahu, Vardhan Gacche et al.

The integrity of the peer-review process is vital for maintaining scientific rigor and trust within the academic community. With the steady increase in the usage of large language models (LLMs) like ChatGPT in academic writing, there is a growing concern that AI-generated texts could compromise scientific publishing, including peer-reviews. Previous works have focused on generic AI-generated text detection or have presented an approach for estimating the fraction of peer-reviews that can be AI-generated. Our focus here is to solve a real-world problem by assisting the editor or chair in determining whether a review is written by ChatGPT or not. To address this, we introduce the Term Frequency (TF) model, which posits that AI often repeats tokens, and the Review Regeneration (RR) model, which is based on the idea that ChatGPT generates similar outputs upon re-prompting. We stress test these detectors against token attack and paraphrasing. Finally, we propose an effective defensive strategy to reduce the effect of paraphrasing on our models. Our findings suggest both our proposed methods perform better than the other AI text detectors. Our RR model is more robust, although our TF model performs better than the RR model without any attacks. We make our code, dataset, and model public.

Ran Elgedawy, Sanjay Das, Ethan Seefried et al.

Operational safety at mission-critical work sites is a top priority given the complex and hazardous nature of daily tasks. This paper presents the Human-Agent Risk Navigation and Event Safety System (HARNESS), a modular AI framework designed to forecast hazardous events and analyze operational risks in U.S. Department of Energy (DOE) environments. HARNESS integrates Large Language Models (LLMs) with structured work data, historical event retrieval, and risk analysis to proactively identify potential hazards. A human-in-the-loop mechanism allows subject matter experts (SMEs) to refine predictions, creating an adaptive learning loop that enhances performance over time. By combining SME collaboration with iterative agentic reasoning, HARNESS improves the reliability and efficiency of predictive safety systems. Preliminary deployment shows promising results, with future work focusing on quantitative evaluation of accuracy, SME agreement, and decision latency reduction.

Maria Mahbub, Vanessa Lama, Sanjay Das et al.

High-Risk Property (HRP) classification is critical at U.S. Department of Energy (DOE) sites, where inventories include sensitive and often dual-use equipment. Compliance must track evolving rules designated by various export control policies to make transparent and auditable decisions. Traditional expert-only workflows are time-consuming, backlog-prone, and struggle to keep pace with shifting regulatory boundaries. We demo ORCHID, a modular agentic system for HRP classification that pairs retrieval-augmented generation (RAG) with human oversight to produce policy-based outputs that can be audited. Small cooperating agents, retrieval, description refiner, classifier, validator, and feedback logger, coordinate via agent-to-agent messaging and invoke tools through the Model Context Protocol (MCP) for model-agnostic on-premise operation. The interface follows an Item to Evidence to Decision loop with step-by-step reasoning, on-policy citations, and append-only audit bundles (run-cards, prompts, evidence). In preliminary tests on real HRP cases, ORCHID improves accuracy and traceability over a non-agentic baseline while deferring uncertain items to Subject Matter Experts (SMEs). The demonstration shows single item submission, grounded citations, SME feedback capture, and exportable audit artifacts, illustrating a practical path to trustworthy LLM assistance in sensitive DOE compliance workflows.

Rui Pan, Tuan Dung Nguyen, Hardik Arora et al.

Continual pretraining of large language models on domain-specific data has been proposed to enhance performance on downstream tasks. In astronomy, the previous absence of astronomy-focused benchmarks has hindered objective evaluation of these specialized LLM models. Leveraging a recent initiative to curate high-quality astronomical MCQs, this study aims to quantitatively assess specialized LLMs in astronomy. We find that the previously released AstroLLaMA series, based on LLaMA-2-7B, underperforms compared to the base model. We demonstrate that this performance degradation can be partially mitigated by utilizing high-quality data for continual pretraining, such as summarized text from arXiv. Despite the observed catastrophic forgetting in smaller models, our results indicate that continual pretraining on the 70B model can yield significant improvements. However, the current supervised fine-tuning dataset still constrains the performance of instruct models. In conjunction with this study, we introduce a new set of models, AstroLLaMA-3-8B and AstroLLaMA-2-70B, building upon the previous AstroLLaMA series.

Ashish Rana, Deepanshu Khanna, Tirthankar Ghosal et al.

Exponential growth in digital information outlets and the race to publish has made scientific misinformation more prevalent than ever. However, the task to fact-verify a given scientific claim is not straightforward even for researchers. Scientific claim verification requires in-depth knowledge and great labor from domain experts to substantiate supporting and refuting evidence from credible scientific sources. The SciFact dataset and corresponding task provide a benchmarking leaderboard to the community to develop automatic scientific claim verification systems via extracting and assimilating relevant evidence rationales from source abstracts. In this work, we propose a modular approach that sequentially carries out binary classification for every prediction subtask as in the SciFact leaderboard. Our simple classifier-based approach uses reduced abstract representations to retrieve relevant abstracts. These are further used to train the relevant rationale-selection model. Finally, we carry out two-step stance predictions that first differentiate non-relevant rationales and then identify supporting or refuting rationales for a given claim. Experimentally, our system RerrFact with no fine-tuning, simple design, and a fraction of model parameters fairs competitively on the leaderboard against large-scale, modular, and joint modeling approaches. We make our codebase available at https://github.com/ashishrana160796/RerrFact.

Ethan Seefried, Ran Eldegaway, Sanjay Das et al.

Decades of engineering drawings and technical records remain locked in legacy archives with inconsistent or missing metadata, making retrieval difficult and often manual. We present Blueprint, a layout-aware multimodal retrieval system designed for large-scale engineering repositories. Blueprint detects canonical drawing regions, applies region-restricted VLM-based OCR, normalizes identifiers (e.g., DWG, part, facility), and fuses lexical and dense retrieval with a lightweight region-level reranker. Deployed on ~770k unlabeled files, it automatically produces structured metadata suitable for cross-facility search. We evaluate Blueprint on a 5k-file benchmark with 350 expert-curated queries using pooled, graded (0/1/2) relevance judgments. Blueprint delivers a 10.1% absolute gain in Success@3 and an 18.9% relative improvement in nDCG@3 over the strongest vision-language baseline}, consistently outperforming across vision, text, and multimodal intents. Oracle ablations reveal substantial headroom under perfect region detection and OCR. We release all queries, runs, annotations, and code to facilitate reproducible evaluation on legacy engineering archives.

Agada Joseph Oche, Ademola Glory Folashade, Tirthankar Ghosal et al.

Retrieval-Augmented Generation (RAG) represents a major advancement in natural language processing (NLP), combining large language models (LLMs) with information retrieval systems to enhance factual grounding, accuracy, and contextual relevance. This paper presents a comprehensive systematic review of RAG, tracing its evolution from early developments in open domain question answering to recent state-of-the-art implementations across diverse applications. The review begins by outlining the motivations behind RAG, particularly its ability to mitigate hallucinations and outdated knowledge in parametric models. Core technical components-retrieval mechanisms, sequence-to-sequence generation models, and fusion strategies are examined in detail. A year-by-year analysis highlights key milestones and research trends, providing insight into RAG's rapid growth. The paper further explores the deployment of RAG in enterprise systems, addressing practical challenges related to retrieval of proprietary data, security, and scalability. A comparative evaluation of RAG implementations is conducted, benchmarking performance on retrieval accuracy, generation fluency, latency, and computational efficiency. Persistent challenges such as retrieval quality, privacy concerns, and integration overhead are critically assessed. Finally, the review highlights emerging solutions, including hybrid retrieval approaches, privacy-preserving techniques, optimized fusion strategies, and agentic RAG architectures. These innovations point toward a future of more reliable, efficient, and context-aware knowledge-intensive NLP systems.

Atilla Kaan Alkan, Shashwat Sourav, Maja Jablonska et al.

Hypothesis generation is a fundamental step in scientific discovery, yet it is increasingly challenged by information overload and disciplinary fragmentation. Recent advances in Large Language Models (LLMs) have sparked growing interest in their potential to enhance and automate this process. This paper presents a comprehensive survey of hypothesis generation with LLMs by (i) reviewing existing methods, from simple prompting techniques to more complex frameworks, and proposing a taxonomy that categorizes these approaches; (ii) analyzing techniques for improving hypothesis quality, such as novelty boosting and structured reasoning; (iii) providing an overview of evaluation strategies; and (iv) discussing key challenges and future directions, including multimodal integration and human-AI collaboration. Our survey aims to serve as a reference for researchers exploring LLMs for hypothesis generation.

Tijmen de Haan, Yuan-Sen Ting, Tirthankar Ghosal et al.

General-purpose large language models, despite their broad capabilities, often struggle with specialized domain knowledge, a limitation particularly pronounced in more accessible, lower-parameter versions. This gap hinders their deployment as effective agents in demanding fields such as astronomy. Building on our prior work with AstroSage-8B, this study introduces AstroSage-70B, a significantly larger and more advanced domain-specialized natural-language AI assistant. It is designed for research and education across astronomy, astrophysics, space science, astroparticle physics, cosmology, and astronomical instrumentation. Developed from the Llama-3.1-70B foundation, AstroSage-70B underwent extensive continued pre-training on a vast corpus of astronomical literature, followed by supervised fine-tuning and model merging. Beyond its 70-billion parameter scale, this model incorporates refined datasets, judiciously chosen learning hyperparameters, and improved training procedures, achieving state-of-the-art performance on complex astronomical tasks. Notably, we integrated reasoning chains into the SFT dataset, enabling AstroSage-70B to either answer the user query immediately, or first emit a human-readable thought process. Evaluated on the AstroMLab-1 benchmark -- comprising 4,425 questions from literature withheld during training -- AstroSage-70B achieves state-of-the-art performance. It surpasses all other tested open-weight and proprietary models, including leading systems like o3, Gemini-2.5-Pro, Claude-3.7-Sonnet, Deepseek-R1, and Qwen-3-235B, even those with API costs two orders of magnitude higher. This work demonstrates that domain specialization, when applied to large-scale models, can enable them to outperform generalist counterparts in specialized knowledge areas like astronomy, thereby advancing the frontier of AI capabilities in the field.

Ethan Seefried, Prahitha Movva, Naga Harshita Marupaka et al.

We propose Enginuity - the first open, large-scale, multi-domain engineering diagram dataset with comprehensive structural annotations designed for automated diagram parsing. By capturing hierarchical component relationships, connections, and semantic elements across diverse engineering domains, our proposed dataset would enable multimodal large language models to address critical downstream tasks including structured diagram parsing, cross-modal information retrieval, and AI-assisted engineering simulation. Enginuity would be transformative for AI for Scientific Discovery by enabling artificial intelligence systems to comprehend and manipulate the visual-structural knowledge embedded in engineering diagrams, breaking down a fundamental barrier that currently prevents AI from fully participating in scientific workflows where diagram interpretation, technical drawing analysis, and visual reasoning are essential for hypothesis generation, experimental design, and discovery.

Kartik Shinde, Laurent Besacier, Ondrej Bojar et al.

This paper presents the third edition of AutoMin, a shared task on automatic meeting summarization into minutes. In 2025, AutoMin featured the main task of minuting, the creation of structured meeting minutes, as well as a new task: question answering (QA) based on meeting transcripts. The minuting task covered two languages, English and Czech, and two domains: project meetings and European Parliament sessions. The QA task focused solely on project meetings and was available in two settings: monolingual QA in English, and cross-lingual QA, where questions were asked and answered in Czech based on English meetings. Participation in 2025 was more limited compared to previous years, with only one team joining the minuting task and two teams participating in QA. However, as organizers, we included multiple baseline systems to enable a comprehensive evaluation of current (2025) large language models (LLMs) on both tasks.

Ahmad Maroof Karimi, Woong Shin, Jesse Hines et al.

We present EPIC, an AI-driven platform designed to augment operational data analytics. EPIC employs a hierarchical multi-agent architecture where a top-level large language model provides query processing, reasoning and synthesis capabilities. These capabilities orchestrate three specialized low-level agents for information retrieval, descriptive analytics, and predictive analytics. This architecture enables EPIC to perform HPC operational analytics on multi-modal data, including text, images, and tabular formats, dynamically and iteratively. EPIC addresses the limitations of existing HPC operational analytics approaches, which rely on static methods that struggle to adapt to evolving analytics tasks and stakeholder demands. Through extensive evaluations on the Frontier HPC system, we demonstrate that EPIC effectively handles complex queries. Using descriptive analytics as a use case, fine-tuned smaller models outperform large state-of-the-art foundation models, achieving up to 26% higher accuracy. Additionally, we achieved 19x savings in LLM operational costs compared to proprietary solutions by employing a hybrid approach that combines large foundational models with fine-tuned local open-weight models.

Jan Philip Wahle, Nischal Ashok, Terry Ruas et al.

A drastic rise in potentially life-threatening misinformation has been a by-product of the COVID-19 pandemic. Computational support to identify false information within the massive body of data on the topic is crucial to prevent harm. Researchers proposed many methods for flagging online misinformation related to COVID-19. However, these methods predominantly target specific content types (e.g., news) or platforms (e.g., Twitter). The methods' capabilities to generalize were largely unclear so far. We evaluate fifteen Transformer-based models on five COVID-19 misinformation datasets that include social media posts, news articles, and scientific papers to fill this gap. We show tokenizers and models tailored to COVID-19 data do not provide a significant advantage over general-purpose ones. Our study provides a realistic assessment of models for detecting COVID-19 misinformation. We expect that evaluating a broad spectrum of datasets and models will benefit future research in developing misinformation detection systems.

Tirthankar Ghosal, Amitra Salam, Swati Tiwari et al.

Detecting novelty of an entire document is an Artificial Intelligence (AI) frontier problem that has widespread NLP applications, such as extractive document summarization, tracking development of news events, predicting impact of scholarly articles, etc. Important though the problem is, we are unaware of any benchmark document level data that correctly addresses the evaluation of automatic novelty detection techniques in a classification framework. To bridge this gap, we present here a resource for benchmarking the techniques for document level novelty detection. We create the resource via event-specific crawling of news documents across several domains in a periodic manner. We release the annotated corpus with necessary statistics and show its use with a developed system for the problem in concern.