Scientific Discovery

In this report, we present LOGOS (Language Of Generative Objects in Science), a scientific generative language model that unifies heterogeneous tasks across the natural sciences within a single autoregressive framework based on a shared scientific grammar. It encodes diverse scientific objects and their spatial interactions as token sequences over a common vocabulary. By representing spatial contact and constraint patterns as discrete tokens, the model captures complex structural interactions in a purely sequential manner, without relying on explicit coordinates or geometric neural networks. This unified representation enables a wide range of downstream tasks to be formulated consistently as next-token prediction in the same grammar space, creating strong alignment between continued multi-domain pre-training and downstream objectives. Across diverse tasks, LOGOS consistently matches or outperforms domain-specific baselines, providing preliminary evidence for the feasibility of "one model fits all" in the natural sciences. We train LOGOS models at different scales (1B, 3B, and 8B parameters) and find a consistent positive correlation between model size and performance. This suggests that the future of AI for Science (AI4S) may not lie in building an independent technical stack that is separated from large language models (LLMs). Instead, it may depend on deeply aligning scientific foundation models with LLMs through shared architectures, shared training paradigms, and shared inference infrastructure, so that LLMs can truly become a new entry point for AI4S. We release the model weights and associated resources to facilitate further research.

This work investigates the ability of large language models (LLMs) to generate mathematical equations from scientific texts. Prior work faces challenges in unstructured grounding, multi-equation dependency, and humanaligned evaluation. To this end, we construct a dataset of AI research papers, pairing contextual passages with ground-truth equations and variable descriptions. We develop an explainable equation generation workflow and evaluate it across diverse open- and closed-source LLM backbones. We introduce an evaluation protocol combining automatic metrics, LLM-based rubrics, and human judgments to assess accuracy, explainability, and human-LLM alignment. Results indicate that LLMs perform moderately on lexical- and syntactic-based similarity, while struggling with semantic accuracy. Comparisons between LLM-based evaluations and human judgments reveal limited alignment, highlighting challenges in using LLMs to assess equation quality. These findings offer insights for improving equation generation models and developing more reliable evaluation methods for scientific text. We provide code and data for reproducibility.

Neural networks are used as generative surrogate models for scientific discovery, which are trainable approximations of scientific simulations. These models enable users to replace time-consuming numerical simulations with learned alternatives, providing quick solutions. However, high-fidelity generative surrogate models require massive training datasets, which can create storage and I/O challenges. Lossy compression is a promising way to reduce this burden, but compression errors may affect the model quality in subtle ways, making it challenging to quantify their impact. In this work, we examine how lossy compression of training data impacts the quality of generative surrogate models. We begin by characterizing the uncertainty inherent in training neural networks, showing that identical training configurations can produce different models. By exploiting this variability, we propose a method to estimate how much compression-induced error a surrogate model can tolerate without affecting its accuracy. Evaluation of two application simulations demonstrates that our approach significantly reduces memory/storage requirements and speeds up training while producing high-quality surrogate models. These results show that lossy compression saves data storage up to 23.7x and 39x with negligible impact on the quality of the surrogate model. Meanwhile, reducing the size of the training data set also enhances the data loading speed and reduces the training time by up to 3x.

Frontier scientific reasoning remains a major challenge for large language models (LLMs), where even the strongest commercial systems fall short of expert-level performance. A closer look at model behavior reveals substantial complementarity that single-model evaluation hides: different frontier models excel on different question types, and no single model captures the full picture. We present SciOrch, a framework that trains a lightweight 8B model to orchestrate frontier LLMs for scientific reasoning. The orchestrator decomposes each question, delegates sub-problems to selected commercial models through API calls, and synthesizes a final answer. Training such an orchestrator is fundamentally harder than conventional agentic RL: each action triggers an API call that is expensive in both dollar cost and latency, making standard online rollouts infeasible. We address this with MCTS-based approach, producing diverse orchestration trajectories, extracting per-node single-turn samples, and optimizing the orchestrator with GRPO-style training. On a 240-question test set spanning SGI-Reasoning and Scientists' First Exam, SciOrch reaches 56.66% average accuracy, outperforming the strongest single commercial model by 3.74% and the strongest multi-agent baseline by 3.33%. It also attains the best accuracy on both SGI and SFE with less than half the API cost of typical multi-agent methods.

Qualitative coding is central to social science, but expert annotation is difficult to scale. LLMs offer a possible extension, yet require careful validation when the target construct is interpretive, theoretically loaded, and only indirectly expressed. We study this problem in a difficult case: detecting whether authors treat Bayesian models as descriptions of mental and neural mechanisms (realism) or as useful mathematical tools (instrumentalism). Our method combines a theory-driven codebook, expert-coded reference annotations, a diagnostic-gated prompt-optimization search yielding a shared zero-shot prompt for three frontier LLMs (GPT-5.1, Claude Sonnet 4.6, Gemini 3 Pro Preview), and multi-rater reliability analysis. The final prompt achieved a held-out combined reliability score of 0.76 (harmonic mean of ICC = 0.79 and $α$ = 0.74), with all diagnostics satisfied. Deployed on 6,858 quotes from 210 articles, the three LLMs reached substantial quote-level agreement (ICC = 0.80; $α$ = 0.76; combined = 0.78) and near-perfect article-level rank stability ($r$ = 0.96-0.97 across rater pairs). The corpus was predominantly weakly realist, but article-level stances were rarely uniform: only 1.4% of articles used a single band, while 59.5% spanned four or more. Low-level perception/motor articles scored 8.8 Realism points higher than high-level cognition articles ($p < .001$, $d = 0.60$), quantifying a long-held qualitative intuition. We present this as an expert-led case study; the framework is intended to generalize to similar theoretically demanding tasks, not to all qualitative analysis.

Scientific discovery relies on large-scale hypothesis testing. However, the capacity to identify true discoveries while controlling false discovery faces major challenges: obtaining relevant reference data (the null distribution) is resource-intensive, leaving finite-data uncertainty, and the procedure should account for the inherent structure in the hypothesis space, when such structure exists. Here, we present a framework for controlling the false discovery rate both when each hypothesis is evidenced only by a finite count of null draws, leaving its p-value uncertain, and when the hypothesis space carries arbitrary structure, requiring only that the structure be represented through a suitable reproducing kernel. We present two decision rules that are both robust to structural mis-specification, yet offer a distinct trade-off between exact FDR control and statistical power. The first rule guarantees exact FDR control; the second maximizes power by adapting mirror-statistic control into count space, utilizing an analytical framework to assess FDR control when exact mirror symmetry is relaxed. Furthermore, the tractability gained by the RKHS framework allows us to directly investigate finite-data uncertainties, which we leverage to suggest a policy for the efficient allocation of null distribution samples.

High-throughput experimentation can evaluate many reaction conditions, yet combinatorial condition spaces still exceed the available experiment budget. This makes experiment selection a sequential decision problem: each new condition must be chosen from limited observations before its outcome is known. LLMs can express task-specific selection logic. A direct recommendation, however, is neither a persistent executable object that can be validated and revised nor an independently auditable decision rule. We introduce CARE, a reference-conditioned controller that separates program synthesis from experiment selection. An LLM writes an executable scoring program that ranks the remaining conditions, while a non-LLM reference policy supplies a numerical candidate and support summary. CARE forms an optional alternative from the program, applies a reference-conditioned intervention gate to compare it with the reference, and records the decision before the selected outcome is revealed. Each new outcome updates controller state and can trigger retention, revision, or regeneration of the active program. This outcome-guided program evolution changes the scoring logic without updating the LLM parameters. In matched offline replay with 30 seeds on eight reaction-optimization tasks, CARE attains the lowest normalized regret, the highest normalized best-so-far AUC, and the highest Top-1% Success@15 among the evaluated methods. These results support using a scoring program written by an LLM as one component of a reference-conditioned optimizer rather than as a standalone experiment selector.

Error-bounded lossy compression is a fundamental technique for managing the rapidly growing volumes of scientific data produced by modern simulations and observational instruments. Most state-of-the-art-compressors follow a prediction-residual paradigm, where compression effectiveness depends on the quality of the predictor: more accurate predictions generate smaller residuals that are easier to compress. This observation raises a question: can modern machine learning models serve as superior predictors for scientific data compression? Answering this question directly is challenging because developing compression-specific ML predictors requires substantial resources. Instead, we leverage the climate domain where highly accurate pretrained weather forecasting foundation models already exist, making them an ideal testbed. We present a framework that integrates spatial and temporal deep learning models into a conventional error-bounded compression pipeline. The framework supports auto-regressive forecasting models and avoids error accumulation. Using ERA5 climate data as a representative large-scale scientific dataset, we evaluate three distinct ML predictors: a VAEformer-based codec (CRA5), a graph neural network forecaster (GraphCast), and a vision-transformer forecaster (Aurora), against the state-of-the-art compressor SZ3.1 under identical quantization and entropy-coding backends. Our evaluation over approximately 1.7 TB of data reveals a surprising result: although ML predictors generate more accurate predictions and can improve reconstruction quality by up to 91% while achieving up to 9.6x higher compression ratios for highly predictable variables, they do not improve overall dataset-level compression ratio. We show that prediction accuracy alone is insufficient: the spatial structure of the resulting residuals plays a decisive role in entropy coding efficiency.

Current LLM-based research agents have advanced through agent orchestration, yet largely overlook scientific knowledge orchestration. Existing works often reduce papers to abstracts, surface mentions, and flat \texttt{cites} edges, omitting key entities, claims, evidence, mechanisms, and method lineages essential for scientific reasoning. To this end, we introduce \textbf{Agents-K1}, an end-to-end knowledge orchestration pipeline that converts raw documents into agent-native scientific knowledge graphs. Agents-K1 integrates three components under a unifying theoretical foundation: a multimodal parser whose five-module schema captures entities, multimodal evidence, citations, and typed inter-entity relations across the full paper rather than abstracts alone; a 4B information-extraction backbone trained with GRPO under a rule-based reward; and a graphanything CLI, a tri-source agent interface that unifies web search, multimodal graph retrieval, and cross-document traversal. On top of this, we process 2.46 million scientific papers across six subjects to produce \textbf{Scholar-KG}, of which we release a one-million-paper subset, and the full Scholar-KG is accessible via the SCP link below. The same pipeline can be extended to general-domain corpora and to schema-conformant data synthesis. Extensive experiments demonstrate that Agents-K1 achieves superior performance in scientific information extraction, knowledge graph construction, and multi-hop scientific reasoning.

Scientific laboratories increasingly rely on AI systems to reason about experiments, but the physical act of doing science remains largely outside their reach. AI can help read literature, generate hypotheses, and plan protocols, yet the execution of those protocols at the bench still requires a human operator. Vision-Language-Action (VLA) models provide one possible interface between written protocols and robot execution, but existing policies are trained mostly on household and tabletop demonstrations and rarely encounter the instruments, transparent liquids, or fixed protocol workflows found in scientific laboratories. Closing this gap requires both laboratory-specific supervision and a unified learning framework that can accommodate the diverse robot embodiments used to execute experimental protocols. We therefore identify data and embodiment as central bottlenecks alongside model design. To address the data side, we build RoboGenesis, a simulation-based workflow and data engine that composes configured laboratory workflows from atomic skills, validates and filters rollouts, and exports structured demonstrations across supported robot profiles. On the policy side, we present LabVLA, trained with a two-stage recipe: FAST action token pretraining first makes the Qwen3-VL-4B-Instruct backbone action aware before any continuous control is learned, and flow matching posttraining then attaches a DiT action expert under knowledge insulation. On the LabUtopia benchmark, LabVLA achieves the highest average success rate among all evaluated baselines under both in-distribution and out-of-distribution settings.

Current discussions of AI in scientific discovery are often dominated by two visible capabilities: search over existing knowledge and execution through optimization, simulation, and automation. Both are important, but neither fully captures the central act of discovery: the formation and evolution of models. This paper proposes a three-layer view of AI in discovery. Layer 1 is search and retrieval by large language models. Layer 2, as the main innovation of this paper, is model formation through qualitative reasoning: the capacity to recognize when a current framework is structurally inadequate and to understand the problem within a broader representational space, not through trial and error, but through structural insight into what is missing and where it can be found. Layer 3 is execution, optimization, and refinement. The main claim is that Layer 2 is both the most important and the least developed. Search without model formation remains confined to inherited frameworks, while execution without conceptual revision only amplifies an existing formulation. We illustrate Layer 2 reasoning through three case studies: S. S. Chern's intrinsic proof of the Gauss-Bonnet theorem, the resolution of the Nesterov Accelerated Gradient convergence problem via Lyapunov functions, and the autonomous disproof of the Erdos unit distance conjecture by OpenAI in 2026. Each case exhibits the same structural signature: a framework that had become inadequate, a missing conceptual object, and a resolution found in an unexpected neighboring field.

Novelty is a crucial metric for assessing the quality of academic papers. Scholars strive to highlight the novel aspects of their work, particularly in the title, abstract, and introduction. Peer review, serving as the gatekeeper of scientific rigor, rigorously evaluates the novelty of papers, yet a cognitive gap may exist between author self-promotion and reviewer evaluation. To investigate this, we analyzed 15,328 academic papers published in Nature Communications from 2016 to 2021, along with their peer-review comments. We found that both reviewers and authors emphasize result-oriented innovation, with reviewers adopting a more comprehensive evaluation perspective. Furthermore, by examining promotional intensity against inherent paper novelty, we found that its effect depends on the paper's actual innovation level. Highly innovative papers benefit from stronger promotional language, receiving more positive evaluations. We also found that promotional language significantly correlates with reviewer disagreement on novelty specifically for papers of moderate innovativeness, whereas it has negligible impact for papers with either very high or very low novelty. This reveals how promotional language operates most prominently in the gray area of academic evaluation.

Learning functional relationships from noisy data is a central problem in scientific inference. Spectral methods approximate unknown functions by expanding them in a basis and estimating the corresponding coefficients from data, but the stability of these coefficients under noise remains poorly understood. Here we study supervised regression with additive label noise using sparse spectral representations across multiple bases and dimensions. We show that noise induces a predictable drift in the learned coefficient vector whose magnitude depends on the effective number of active spectral modes. After whitening the empirical feature geometry, we derive a closed-form expression for the overlap between noisy and noiseless coefficient vectors, revealing a universal degradation curve governed by a single intrinsic noise scale. Numerical experiments across Fourier, Legendre, Bessel, and Haar bases confirm the theoretical prediction. The results demonstrate that spectral learning exhibits a fundamental noise threshold beyond which coefficient estimates become unstable, placing intrinsic limits on recovering functional structure from noisy data.

Three paradigmatic forms of inference recur across scientific reasoning: deduction, induction, and causal abduction. Reliably evaluating LLMs on these in scientific settings is currently out of reach: scientific benchmarks built on human annotations are costly and lack mechanistic ground truth, while synthetic logical-reasoning benchmarks do not resemble real scientific documents. We introduce SciR, a benchmark that combines multi-paradigm reasoning with controllable scientific rendering, anchored on three paradigmatic scientific problems. Tasks are generated from formal objects (deduction tree, inductive rule hypothesis, causal graph) to guarantee verifiable answers, then rendered into multi-document scientific discourse via per-track domain-tuned genres. The construction lets us independently vary two difficulty axes: how hard it is to extract the key information needed for inference, and how hard the principled inference itself is. We test six models. Both axes hurt every model, and their effects compound. The rendering even hurts neurosymbolic pipelines, which hand inference to a verified solver. The two axes yield a per-model extraction-vs-inference profile: for instance, reasoning models like deepseek-r1 mostly surpass non-reasoning instruct models on the inference axis. To our knowledge, SciR is the first multi-paradigm scientific-reasoning benchmark with parametric control on both extraction and inference difficulty.

AI agents are increasingly being developed to accelerate scientific discovery, yet their practical capabilities in real research settings remain poorly understood. Existing benchmarks for AI agents rarely capture the complexity, heterogeneity, and extended reasoning required by scientific work, whereas benchmarks for scientific tasks often reduce research to static, direct problems and provide limited support for interactive evaluation. Here, we introduce SciAgentArena, a systematic benchmark for evaluating AI agents in real-world scientific research scenarios drawn from emerging needs across multiple domains. SciAgentArena comprises approximately 200 tasks with stepwise verification and an interactive, agent-agnostic environment for assessing diverse AI agents. Using this benchmark, we find that current agents can contribute effectively to well-specified data-analysis workflows, particularly when the task structure and evaluation criteria are clear. However, their performance remains uneven across scientific contexts: agents struggle to generate genuinely novel insights, sustain self-directed exploration, and formulate robust solutions for open-ended research questions. We further characterize common failure modes across agents and identify opportunities for improving their reliability, autonomy, and scientific reasoning. Together, SciAgentArena provides a practical framework for measuring progress in AI agents for science and for guiding the design of future agents capable of addressing complex scientific challenges. Full codes, tasks, and datasets can be accessed via this link: https://sciagentarena.github.io/.

Scientific figures compress complex pipelines into a single canvas, yet understanding them requires paper-grounded, step-by-step narration aligned with visual highlights a capability missing from current video generation systems and benchmarks. To address this, we introduce paper-grounded figure-to-video generation: generating narrated, region-grounded walkthrough videos from a figure and its paper. We propose MINARD (Multimodal Interpretation of Narrated Architecture via Region Decomposition), a pipeline that generates paper-grounded narrations and sequentially grounds them to figure regions. We also release FigTalk, a benchmark with new sequential and component-level grounding metrics derived. On FigTalk, MINARD generates humanlike, paper-faithful narrations and outperforms narration-conditioned figure spatial grounding compared to existing approaches in both automatic and human evaluation

High-stakes clinical use of large vision-language models (LVLMs) requires reasoning that is grounded in visual evidence and clinical knowledge, not just correct final answers. We introduce OpenMedReason, a large-scale, open multimodal medical reasoning corpus comprising approximately 450K image-question-answer instances whose reasoning traces are primarily derived from curated biomedical, human-authored scientific articles. OpenMedReason provides high-fidelity supervision beyond synthetic chains of thought, covering diverse medical domain vision modalities such as radiological scans, microscopic images, visible light photographs, charts, and others. We complement it with OpenMedReason-Bench, a held-out benchmark that allows fine-grained evaluation of LVLMs along three complementary axes of capability, including perception, medical knowledge, and rationale, enabling diagnostic evaluation beyond final-answer accuracy. OpenMedReason is a rich training resource that exhibits its effectiveness in both supervised fine-tuning (SFT) and reinforcement-based alignment. Training with OpenMedReason yields a 20% average improvement in VQA accuracy over the base model and achieves performance within 4.2% of the strongest comparable-scale medical LVLMs. Fine-grained performance analysis confirms that the gains are not concentrated in any single axis: OpenMedReason improves perception, medical knowledge, and rationale jointly, and its reasoning traces are preferred over those of the base model in 86.1% of pairwise comparisons. We release the code and dataset at huggingface.co/datasets/neginb/OpenMedReason.

The Coupled Model Intercomparison Project Phase 6 (CMIP6) has generated thousands of peer-reviewed publications documenting model configurations, evaluation procedures, emergent constraints, and projection uncertainties. As the community transitions toward CMIP7, efficiently extracting and operationalizing this unstructured knowledge alongside live data analysis represents a critical bottleneck. Here we present CMIP-Forge, a hybrid retrieval-augmented generation (RAG) and autonomous analysis system that bridges the gap between scientific literature and Earth System Grid Federation (ESGF) data archives. The system pairs a curated corpus of 6,581 CMIP6-related open-access publications (101,828 indexed chunks) with an agentic pipeline in which a tool-augmented worker plans and executes Python workflows over live climate data, while a panel of independent reviewer models audits its methodology end to end. CMIP-Forge introduces a multi-layered Defense-in-Depth architecture that enforces physical and methodological invariants through executable mechanisms: Abstract Syntax Tree (AST) static analysis, audited scientific primitives, and an autonomous adversarial peer-review protocol. We demonstrate the system's capabilities through end-to-end autonomous research pipelines spanning atmospheric teleconnections, ocean dynamics, regional extremes, and global warming projections. An agentic analysis system grounded in peer-reviewed literature, constrained by automated code guardrails, and audited by an independent adversarial review loop can complete complex climate-research workflows autonomously. The same experiments expose concrete failure modes of the review loop (sycophantic regression, REVISE verdicts that are never resolved, and the submission of stub code for review), each diagnosable from the immutable telemetry and provenance record released with the article.

LLMs are increasingly used to generate and judge scientific ideas. This makes novelty evaluation a central problem. Full idea evaluation is difficult because it often requires judging a method, its feasibility, and its empirical promise. We therefore study a cleaner upstream object: the research question (RQ). RQ generation is a prerequisite for scientific ideation, and RQs can be compared against questions pursued in real papers. We introduce RQ-Bench, a benchmark built from recent arXiv papers. For each paper, we reconstruct author-anchored RQs from its cited background, gaps, and contributions. These RQs are not the only valid questions for the same background. They are author-anchored reference points for testing novelty judgments. We evaluate model-generated RQs with standalone LLM judging, comparative LLM judging, and human expert evaluation. LLM judges consistently rate model-generated RQs as highly novel, producing a novelty mirage; in comparative evaluations, this preference becomes even stronger. Domain experts, however, reach the opposite conclusion and prefer the author-anchored reference questions. We further find that many generated RQs are narrow or source-bound, a dimension that LLM judges often miss unless explicitly tested. Overall, the contradictory novelty evaluations between LLM judges and human experts raise a serious concern about the reliability of using LLMs to assess the scientific novelty of research questions.

Open-ended scientific discovery asks agents to move beyond executing analyses for predefined questions. Across multiple rounds of exploration, a discovery agent must decide which phenomena warrant investigation while avoiding overinterpretation, where emerging claims exceed the evidential scope of the analyses. This creates an evidence-calibration problem: the exploration trajectory must be coupled with claim status so that evidence can guide both what to investigate next and what can be claimed. We introduce \textsc{StatefulDiscovery}, a discovery framework that externalizes investigation state and uses it to coordinate frontier selection, evidence acquisition, and claim adjudication. We evaluate \textsc{StatefulDiscovery} across 40 real-data discovery tasks. Compared with several baselines, \textsc{StatefulDiscovery} produces more claims overall judged to be both well-supported and high-value. Ablations indicate distinct roles for structured hypotheses, local adjudication, frontier control and persistent states. Together, these results suggest that explicit discovery state can couple exploration with evidence-calibrated claim formation. Our code is released at \href{https://github.com/SUSTech-GenAI/StatefulDiscovery.git}{https://github.com/SUSTech-GenAI/StatefulDiscovery.git}.

Scientific AI agents increasingly retrieve evidence, reason across sources, and synthesize conclusions used in consequential decisions. Yet, their ability to do so in high-stakes domains such as health remains unclear. We introduce SciConBench, a large-scale live benchmark of 9.11K questions and expert-written conclusions from systematic reviews to evaluate open-domain scientific conclusion synthesis. The benchmark draws on an expert-validated automated evaluation pipeline that decomposes conclusions into atomic facts and measures correctness and comprehensiveness via factual precision and recall. To mitigate data leakage, we further introduce SciConHarness, a clean-room evaluation harness that equips agents with controlled web interaction to ensure valid measurement. Evaluating 8 frontier models and deep research agents, we find that factual quality remains low: under clean-room settings, the best agent achieves only a factual F1 of 0.337. Our clean-room setting consistently reduces performance relative to unconstrained evaluation, suggesting that leakage inflates estimates of models' true synthesis capabilities. Finally, we audit consumer-facing agents (e.g., Google AI Overview, OpenEvidence) and find they frequently generate incomplete and sometimes contradictory conclusions, even when the ground-truth answer is available. Overall, our results show that reliable synthesis of scientific conclusions remains an open challenge, and that clean-room evaluation is essential for assessing open-domain AI agents.

Scientific discovery is often a collective process: researchers share partial results, inspect failed attempts, and build on each other's ideas over long time horizons. Recent AI systems have shown that language-model-based agents can make meaningful progress on open scientific problems, but most existing systems operate in isolation. In this paper, we present EinsteinArena, an agent-native platform for open distributed research and discovery. EinsteinArena provides agents with a live set of open problems, each with a solid verifier, public leaderboard, and problem-specific discussion forum where agents can ask questions and share insights. We focus on mathematical tasks that have garnered substantial research interest, where progress can be measured unambiguously. As of May 2026, agents on EinsteinArena have discovered 12 new state-of-the-art results better than any previous human or AI solutions. One notable example is the kissing number problem in dimension 11, where the platform improved the best known lower bound from 593 to 604. This advance did not come from a single agent or isolated run. Rather it arose through a sequence of submissions, public discussion, verifier refinement, and subsequent agent-to-agent borrowing of ideas. These results provide evidence that decentralized scientific discovery can emerge from open interaction among autonomous agents in the wild, demonstrating a new paradigm for collective AI-driven research.

Accurate posterior estimation is central to scientific inference, as uncertainties determine what can be reliably learned from observational data. While Markov chain Monte Carlo methods provide asymptotic convergence guarantees, they are computationally demanding in high-dimensional settings. Neural network-based generative models for entire discretized 3D fields enable fast amortized inference but often lack convergence guarantees and principled accuracy assessment. Using Hamiltonian Monte Carlo to obtain reference posterior samples, we conduct a controlled field-level evaluation of an implicit generative model (Stochastic Interpolants) and an explicit likelihood-based model (GLOW normalizing flows). This comparison, unavailable in typical applications, enables the detection of posterior geometry failures that standard metrics cannot capture. As a case study, we consider the cosmological inverse problem of inferring cosmic initial conditions from present-day large-scale structure. To match the precision of modern cosmological data, this problem increasingly relies on complex, non-linear, and non-differentiable simulators, which are incompatible with gradient-based inference frameworks. Generative models offer a route to address these challenges, provided their inferred posteriors are reliable. In this work, we show that matching posterior means, marginal distributions, or achieving high cross-correlation does not imply correct uncertainty structure, as revealed by posterior variance fields and sample-based evaluations. Through this work, we aim to raise awareness of the challenges of uncertainty estimation in high-dimensional field-level settings, highlighting the importance of careful design and validation of neural generative approaches for scientific applications.

Differential equations play a critical role in scientific discovery because they provide a mathematical framework to describe the behaviour of physical phenomena. As a promising alternative to traditional first principles, data-driven differential equation discovery has attracted increasing attention for its ability to infer governing laws directly from experimental or simulated data, especially when the underlying physics is unclear. However, the field has expanded rapidly along diverse methodological directions, particularly with the emergence of AI-based approaches, and still lacks a clear organizing perspective. In this Review, we propose a problem-oriented perspective on data-driven differential equation discovery. We first introduce a two-dimensional phase diagram of equation discoverability, where discovery problems are organized according to structural complexity and coefficient complexity. This phase diagram shows how the field has moved from the discovery of sparse equations with simple coefficients toward more complex governing laws with richer structures and more flexible parameterizations. It also clarifies why different methodological families succeed or fail in different problem settings. We then present the representation-evaluation-optimization (REO) framework as a fundamental abstraction of the discovery process. By identifying the core problems of equation discovery that persist across algorithmic variations, REO shifts the discussion from individual algorithms to the fundamental principles that determine discoverability. We connect these perspectives to applications across physics and adjacent sciences, and argue that the next challenge is not merely recovering equations, but using them to revise existing theories, distil mechanisms and form new scientific concepts.

AI Scientist agents are often evaluated as if capability were mainly a function of model quality, prompting, or reasoning scaffolds. We test a different hypothesis in drug-asset valuation: for knowledge-intensive scientific decisions, the limiting factor is often the evidence substrate the agent can access. We run a controlled three-arm ablation on a production valuation agent: A is a plain web-only LLM analyst, B adds public structured tools plus a 14-dimension valuation playbook, verifier, objectivity policy and red-team, and C adds the proprietary Noah AI corpus of curated pipeline, trial and deal intelligence. Across a 13-asset stratified benchmark, B improves calibration and audit discipline: tier-in-range accuracy rises from 0.80 to 0.89 and objectivity from 3.16 to 3.30. But B does not remove the factual ceiling. Under capability-superset accounting, A and B recover only 0.25 and 0.38 of the curated gold competitive record, while C recovers 0.96; on the curated long-tail subset, C reaches 0.93 vs. 0.26/0.30. Raw blind-panel decision quality is similar for A and B (7.01 vs. 6.96), so we introduce completeness-aware decision utility: informed decision-quality = decision-quality x gold-coverage. On this metric, C reaches 7.43 vs. 1.76/2.57 for A/B. Even a perfect non-proprietary-data report would be capped at 3.83 by B's coverage. The result is not that reasoning scaffolds are unimportant; they improve calibration and discipline. Rather, proprietary evidence sets the upper bound of what the AI Scientist can know and therefore decide.

Modern artificial intelligence excels at prediction but cannot explain. From large language models to AI-for-science systems, today's machines answer what by recombining patterns already present in the human literature, yet they cannot reason out why a phenomenon must arise from underlying principles even though explanation, not prediction, lies at the heart of scientific discovery. Here we ask whether the structure of scientific explanation can be operationalized to guide how a machine generates hypotheses. We introduce DN-Hypo-Pipeline, a hypothesis-generation framework that adopts a layered, explanation-theoretic scaffold: Hempel's deductive-nomological (DN) model supplies the output form and deductive validity of a hypothesis, Salmon's causal-process account supplies an organizing constraint on where to search for the governing laws, and Armstrong's view of laws as relations between universals supplies the bridge from a phenomenon's constituent processes to the laws that may be associated with it. Rather than searching the space of what has been written, the framework searches the space of what principles govern a phenomenon: given an explanandum, it abstracts the universals instantiated in the phenomenon's formation process, retrieves the laws relating those universals, and deductively reconstructs a new, testable explanation. Evaluated in data-science modeling and judged by both LLMs and human experts, hypotheses generated through this principled reasoning significantly outperform those from direct prompting. Crucially, we translated the two highest-scoring hypotheses into novel algorithms one that reduces the Transformer's theoretical complexity with only minimal performance loss, and another that achieves competitive accuracy with substantially fewer parameters.

Bold projections that artificial intelligence will accelerate scientific discovery have raced ahead of evidence from working scientists, and the field still lacks large-scale, scientist-in-the-loop tests of these claims. Here we mount the largest such evaluation to date and map what AI cannot yet do for science. We invited authors of 121,640 recent preprints across biology, medicine, chemistry, and the social sciences to judge ideas that large language models (LLMs) generated from the context and puzzles of their own papers. 6,749 scientists returned 25,139 sets of ratings on novelty, empirical feasibility, probability of being true, and favorability of adoption. Three patterns emerge. First, non-reasoning LLMs collapse into a narrow "hivemind" of similar ideas; reasoning models roam a wider hypothesis space, yet no model class spontaneously proposes null hypotheses -- a move humans make more freely. Second, scientists reward ideas that resemble their own and prize probability over novelty, though social scientists tolerate risk more readily than life scientists. Senior social scientists are the harshest critics, and their skepticism is well-earned: LLMs falter most in pluralistic fields like the social sciences that demand context-aware interpretation and evolving theories. Third, automated evaluators on which the community currently relies -- LLM-as-a-judge, artificial metrics, and even state-of-the-art (SOTA) models -- agree only weakly with expert judgment, and retrieval augmentation and scientist persona prompting yield only marginal gains. A Qwen3-14B reward model we post-trained on human ratings captures field taste nuances, beats SOTA models by up to 27%, and closes the gap to the inter-rater consistency of independent peer reviewers. For all the hype, today's scientific AI still represents a collaborator whose imagination, outputs and judgment benefit from human grounding.

Scientific machine learning is limited less by model size than by the data it is trained on. Observational data records what happened but not why; template synthetic data has a known generating process but only for the simulator's template, not the case a user faces. We argue a third option is now operationally feasible: instrumented data, in which every datum carries the mechanistic model that produced it, an explicit uncertainty over that model, and an executable family of counterfactuals. Verification-and-validation (V&V) instrumented image-to-simulation pipelines are one realisation: a sensor observation becomes a fully specified, solver-backed simulation with explicit, editable parameters and a propagated aleatoric/epistemic uncertainty. The substrate is case-specific, mechanistically supervised, and supports causal interventions through Pearl's do-operator. Near-term consequences for validation, auditing, and surrogate training span computational biology, climate, materials, fluid mechanics, and medical imaging; a longer-term, falsifiable implication concerns foundation models for scientific reasoning.

Evaluations of large language models (LLMs) in scientific information seeking tasks have become increasingly use-centric, such as conducting live or multi-turn evaluations with real users. These evaluations still assume a single, static chat interface, but as models are integrated into new interfaces, evaluations must shift to incorporate interface-specific criteria. We propose a new evaluation framework based on a formative study with $16$ participants that tests models' ability to generate multiple responses to one query that differ along an interpretable axis of language (language complexity), inspired by direct manipulation interfaces from human-centered design literature. We evaluate GPT-5.1, GPT-5 mini, Claude Sonnet 4.5 + Thinking, and DeepSeek-V3.1 by generating 5 responses at different levels of language complexity for $98$ scientific queries. While models vary complexity across responses, most changes remain inconsistent, with the best performing model (Claude Sonnet 4.5) only shifting reliable complexity measures in the correct direction $46\%$ of the time. Our findings hold with increased sample size and alternative complexity levels.

We present the knowledge manifold: a Riemannian geometric space in which a corpus of documents is arranged according to semantic positional relationships derived from character n-gram TF-IDF representations. The framework proceeds in five tightly coupled stages. First, each document is converted to a character-level n-gram TF-IDF vector (4-7 grams, up to 250,000 features, L2-normalized) and embedded in a two-dimensional knowledge map via constrained stress minimization with repulsion, variance, and centering regularizers. Second, knowledge at an arbitrary query point is estimated through Smoothed Particle Hydrodynamics (SPH) interpolation using a cubic-spline kernel, yielding an interpolated TF-IDF feature vector that can be linguistically characterized. Third, directional knowledge gradients at 0, 45, and 90 degrees are computed from the SPH interpolation map, and pairwise directional similarity is quantified via inner product and cosine similarity. Fourth, a Gaussian Process Regression (GPR) model, with a Constant x RBF + White kernel fitted on a 10-dimensional SVD projection, provides a Bayesian posterior mean, uncertainty estimate, and per-document contribution rate at the query point. Fifth, geodesics in the knowledge space are obtained by minimizing a discrete Riemannian path energy derived from the SPH-induced metric tensor, using L-BFGS-B with seven deterministic initial-path candidates. We apply the formulation to a corpus of 20 papers in fiber-reinforced composite materials and aerospace structural mechanics, showing that the semantic map recovers meaningful research clusters, geodesic paths reveal natural conceptual bridges between distant topics, and SPH/GPR interpolation enables the generation of virtual knowledge: hypothetical paper abstracts describing unstudied but geometrically predicted research directions.

Recent advances in agentic visualization have enabled the translation of natural language into executable scientific visualization (SciVis) workflows. While general-purpose coding agents show strong capabilities, they often lack the tool-specific expertise required for SciVis tasks. In this work, we present SciVisAgentSkills, a collection of reusable agent skills that augment coding agents for scientific data analysis and visualization by encoding environment assumptions, tool usage patterns, and domain heuristics across scientific tools such as ParaView, napari, VMD, and TTK. We evaluate these skills on Codex and Claude Code using SciVisAgentBench, a benchmark of 108 expert-designed multi-step tasks. Results show that agent skills improve mean task scores across the evaluated suites, with token-efficiency benefits that depend on the agent harness and tool setting. These findings highlight the importance of structured procedural knowledge for enabling reliable, long-horizon SciVis workflows, while also showing that skills should be studied alongside the execution harness that loads and applies them. The skills are available at https://github.com/KuangshiAi/SciVisAgentSkills.

Lossy compression is essential for massive spatiotemporal data from scientific simulations. Learned compressors can achieve high compression ratios at moderate accuracy targets, but their aggregate reconstruction losses do not guarantee accuracy for each block. Existing Guaranteed Autoencoder (GAE) methods add a per-block residual correction by retaining SVD/PCA-style coefficients until the target is met. This works at moderate tolerances, but in the high-fidelity regime with block-level NRMSE from 10^-6 to 10^-4, the number of retained coefficients grows quickly and the correction stream dominates the total rate. We propose a residual-centric view: the learned residual is structurally different from the original scientific field and should be coded with a representation designed for that residual. We introduce two residual coders. LBRC is a deterministic, training-free pipeline that adaptively quantizes the learned residual to the target NRMSE and losslessly encodes the resulting integer residual using 3D Lorenzo differencing, zigzag mapping, bit-plane coding, and entropy coding. NGLR adds a causal neural predictor that outputs a normalized bias for an integer-rounded Lorenzo prediction in the same deterministic integer pipeline, reducing the entropy of the remaining residual code while preserving deterministic decoding. The predictor weights are serialized and counted in the bitstream. Across E3SM, JHTDB, and ERA5 at block-level NRMSE targets from 10^-6 to 10^-4, LBRC improves compression ratio over GAE by 30-60% and is broadly competitive with SZ. NGLR adds a further 10-40% over LBRC and outperforms SZ in the evaluated high-fidelity regime. These results show that residual representations tailored to learned-compressor residuals can preserve the advantage of learned compression when global residual correction becomes rate-dominant.

What would it mean to have a scientific understanding of AI? Models are not static objects: they are snapshots of time-evolving processes shaped by data, objectives, architectures, and optimization dynamics. Yet much of AI research treats models as fixed artifacts, analyzing behaviors after training rather than asking why they emerge. This position paper argues that a science of AI must move beyond post-hoc fixes and study the training dynamics that produce model behavior. Such a science should support progressively stronger forms of understanding: predicting outcomes from early training signals, intervening when trajectories go wrong, and ultimately designing training procedures that more reliably produce desired properties. Scaling laws have made prediction routine for loss; the challenge is extending this success to capabilities, biases, robustness, and safety-relevant behaviors. We articulate requirements for such theories grounded in the history and philosophy of science, examine progress in mechanistic interpretability, fairness, memorization, and simplicity bias, and identify concrete open problems.

We present Archi, an open-source, end-to-end framework for scientific collaborations that combines the systematic ingestion and organization of heterogeneous data sources with the deployment of configurable, private, and extensible agents that retrieve and reason over them. An instance of Archi has been deployed for the Computing Operations team of the CMS experiment at CERN's LHC since February 2026 as a support agent for technical operators, offering retrieval and analysis capabilities by combining documentation, historical data, and live monitoring systems. We evaluate the system on operator feedback and a question set collected from production usage, graded by human and automated panels. The system proves effective at operational tasks, resolving real-world queries posed by CMS operators. We also observe that locally-hosted, open-weight models perform competitively, enabling fully private management of sensitive data.

Large language models (LLMs) are increasingly deployed as autonomous agents in scientific tasks. Yet whether these systems can effectively engage in forms of inductive reasoning relevant to scientific discovery remains an open question. In this work, we introduce FALSIFYBENCH, an evaluation framework for hypothesis-driven reasoning inspired by the classic Wason 2-4-6 task, in which agents must discover hidden semantic properties by iteratively proposing examples and receiving feedback. This task captures key elements of scientific reasoning: hypothesis generation, evidence gathering, and belief revision in response to both confirming and disconfirming evidence. Our evaluation of 12 LLMs across model families and scales shows that reasoning models are generally stronger scientific reasoners than instruction-tuned models, although no model comes close to optimal performance. The primary driver of success is the capacity for negative testing: models that actively seek to falsify their hypotheses consistently outperform those that primarily seek confirmation. Moreover, a fine-grained turn-level analysis, neglected in previous work, reveals that failure is tied to identifiable patterns in how models navigate the hypothesis space.

While Process Reward Models (PRMs) have achieved remarkable success in mathematical reasoning, their application in complex scientific domains-such as biology, chemistry, and physics remains largely unexplored. Scientific problems demand not only logical rigor but also factual consistency and the precise usage of domain-specific tools, areas where current models often suffer from hallucinations and lack of verification. In this paper, we first construct SCIPRM70K, a large-scale dataset featuring Chain-of-Tool trajectories that explicitly interleave reasoning with the execution of scientific tools. Building upon this, we train an efficient reward model called Sci-PRM to provide fine-grained supervision on tool selection, execution accuracy, and result interpretation at each step in one inference. Experiments demonstrate that Sci-PRM significantly enhances foundation models in two key aspects: (1) it enables effective test-time scaling via Best-of-N selection; and (2) when integrated into Reinforcement Learning, it serves as a dense reward signal that mitigates the critical issue of advantage disappearance, allowing the model to break through existing performance ceilings.

Scientific simulators are increasingly being integrated into LLM-driven systems for high-stakes simulation-driven decision-making. However, existing frameworks primarily use LLMs to generate, calibrate, or execute simulators, treating them as black-box interfaces rather than as structured mechanistic systems that can be reasoned about. As a result, current approaches lack the ability to identify, represent, and reason about the assumptions and mechanisms underlying simulator behavior, limiting transparency, auditability, and decision justification. We introduce MechSim, a mechanism-grounded neuro-symbolic reasoning framework for executable scientific simulators. Unlike prior neuro-symbolic approaches that primarily reason over static symbolic structures, MechSim enables LLM agents to reason about the mechanisms, assumptions, and execution behavior of scientific simulators. Our framework represents simulators through a shared structured schema capturing assumptions, variables, mechanism dependencies, and execution traces. On top of this representation, LLM agents operate as constrained reasoning engines that generate structured, evidence-grounded explanations linking simulator outcomes to their underlying mechanisms. We evaluate our approach across multiple high-stakes domains and show that it improves mechanism-level explanation quality, simulator analysis, and downstream decision-making reliability.

This article situates large language models (LLMs) within the longer history of computational approaches to concept analysis in the history, philosophy, and sociology of science (HPSS). We examine what LLMs add to existing methods, how they inherit longstanding problems, and review recent case studies that employ them. In the first part, we reconstruct computational conceptual history before LLMs by bringing together three strands of work: early digital methods in HPSS, distributional approaches from digital history and related research, and lexical semantic change detection. We provide an overview of the main challenges and opportunities, focusing on corpus construction, operationalization and modelling choices, and evaluation and interpretation. In the second part, we turn to the era of LLMs, starting with a short introduction to LLMs before reviewing LLM-based work on lexical semantic change detection and relevant case studies in HPSS. We then revisit the earlier methodological questions, showing how issues of corpus construction, model choice and training data, operationalization trade-offs, and evaluation and interpretation play out in LLM-based workflows.

Modeling temporal evolution is important to analyzing and reasoning about scientific phenomena, yet most machine learning methods provide deterministic forward predictions that overlook multiple plausible outcomes and rarely support backward reasoning, limiting their usefulness in practical scientific workflows. We present a framework that integrates diffusion-based generative modeling with interactive visual analytics for scientific exploration. We introduce DiffUNet^2, a conditional diffusion model that enables bidirectional, any-to-any generation across time and captures distributions of plausible system evolutions. Built upon the model, our interactive system supports branching timeline exploration, user-guided state editing, and probability-space navigation, enabling scientists to actively explore alternative hypotheses rather than passively observe predictions. We evaluate the model on 5 datasets across different scientific domains to validate its predictive accuracy and probability-space ensemble quality. In collaboration with domain experts, we demonstrate the effectiveness of our approach in supporting practical scientific temporal data analysis workflows. By integrating modeling and visual interaction, our approach enables scientists to interactively explore system dynamics, transforming generative models into tools for hypothesis-driven scientific analysis.

Understanding and anticipating scientific change requires models that distinguish between endogenous consolidation and exogenous diffusion of scientific concepts. Using the quantum computing subtree of concepts in OpenAlex, we construct a temporally resolved concept co-occurrence network and track each concept pair through its upstream citation lineage and downstream diffusion. We train LightGBM models on distributional and diversity-aware features to predict four outcomes: endogenous reinforcement, exogenous diffusion, their ratio, and diffusion entropy. After controlling for overall publication growth of the scientific body, endogenous reinforcement proves largely unpredictable in the primary quantum-computing benchmark. In contrast, exogenous diffusion and entropy are strongly predictable ($R^2$ up to $0.78à) and are driven by upstream heterogeneity, citation breadth, and distributional dispersion, as shown by SHAP analyses; replications on robotics, advanced materials, and neuro implants confirm that exogenous diffusion remains the top-ranked target across fields ($R^2_test \sim 0.60-0.87$), while endogenous predictability rises markedly in neuro implants (R^2_test = 0.83), indicating that the quantum-computing asymmetry does not generalise uniformly. Case studies reveal that sharp entropy increases coincide with the opening of new conceptual frontiers, while entropy collapses signal technological convergence or paradigm displacement. These results demonstrate that conceptual diffusion is governed by stable structural regularities embedded in semantic and citation environments. By identifying early diversity-based signals of cross-domain uptake, the approach provides a scalable foundation for anticipatory scientometrics, technology foresight, and innovation-oriented policy analysis in rapidly evolving research fields.

We introduce an explainable machine-learning approach that forecasts the structural precursors of scientific breakthroughs -- the emergence and intensification of links between research concepts -- by modelling how OpenAlex concept networks evolve over time. Using 59 semantic and topological features, a two-stage LightGBM model jointly predicts the formation and the future weight of concept pairs, adding a regression stage that quantifies expected intensity to prior link-existence forecasts. Relative to the state of the art, the approach improves accuracy and explainability at once: comparative validation across four technology and biomedical domains yields ROC-AUC in [0.954, 0.967] at all horizons without re-tuning, exceeding the roughly 0.90 of prior models, while every forecast rests on structural, auditable features rather than opaque embeddings. Classification performance is high (AUC about 0.95) and regression remains stable (RMSLE 0.45 to 0.6 over one to five years). Feature attribution shows that structural factors -- particularly Adamic-Adar similarity and degree-based Hadamard measures -- consistently drive accuracy, suggesting that breakthrough-relevant recombinations emerge in tightly connected sub-networks. Two expert-anchored cases, quantum annealing and AI-enabled quantum architectures, show the model surfacing technological convergence consistent with expert expectations. We then outline a three-layer decision architecture -- detection, expert translation, institutional integration -- that turns these forecasts into evidence-based research strategy and policy, anchored in open data and explainable features.

Does scientific publishing reward the quality of ideas or the advantage of connections? The question is universal to prestige-driven science, yet it has resisted decades of study because a paper's quality could not be gauged ahead of its publication fate without using that fate as the yardstick. We break this constraint by measuring a paper's idea quality directly from its text, before publication, using a discipline-trained LLM evaluator that scores the idea without seeing author names or outcomes. Using economics as a case study, we combine this text-legible idea-quality score with an execution-quality rubric, a connection index, an author-ability index, and an off-the-shelf language-model text score to estimate a five-input production function for journal placement across 6,208 economics working papers. The inputs are not rivals but a sequence along the ladder of prestige. Execution sets a meritocratic floor and is the largest input overall. Text-legible idea quality grades the rungs in between. Connections set a favoritism ceiling that bites mainly near the apex, the most selective journals. Connections work through two additive channels: connected authors write papers that score higher, and at equal scores their papers are still more likely to place better. Yet this advantage is bounded. Connections raise the odds of every rung without making the apex the typical outcome for ordinary ideas, and even the highest-scoring papers face real friction reaching the visible journal ladder. The result nests, rather than chooses between, the meritocracy and network accounts of how science is published.

Autonomous science is moving from demonstration to infrastructure. Large language model agents now plan experiments, and self-driving laboratories execute them. Yet every such system rebuilds the link between the reasoning agent and the physical instrument from scratch, against fragmented vendor SDKs and standards built for deterministic software clients rather than probabilistic, goal-directed agents. Recent agent-interoperability protocols clarify two of the three edges of an agentic ecosystem (Anthropic's Model Context Protocol (MCP) standardizes the agent-to-tool edge, and Google's Agent2Agent (A2A) the agent-to-agent edge), but neither models the agent-to-instrument edge, where operations are stateful, safety-critical, exclusively owned, physically embodied, and produce measurements with units, calibration, and uncertainty. We present the Lab Agent Protocol (LAP), a protocol design that fills this gap. LAP retains A2A's peer-to-peer, discovery-first, task-lifecycle structure and adds four physical-world primitives: (i) the InstrumentCard, a signed capability and physical-limit description; (ii) first-class reservation for exclusive instrument and sample locking; (iii) a safety-fence handshake with operator-confirmation tokens cryptographically bound to a specific task and its parameters, gating hazardous and irreversible operations; and (iv) a MeasurementResult schema that makes every result physically typed (QUDT/UCUM), calibration-anchored, uncertainty-bearing, and reproducible by construction. We specify roles, a six-layer architecture, the JSON-RPC method set, the task and safety state machines, the error model, and cross-laboratory federation, and walk a closed-loop autonomous campaign through the protocol end-to-end. LAP is transport-compatible with the A2A/MCP ecosystem and encapsulates rather than replaces existing device standards such as SiLA 2 and OPC-UA.

Scientific images function as critical evidence in research communication, yet their integrity faces unprecedented threats from AI-generated content that introduces subtle but consequential errors. Existing evaluation paradigms prove inadequate: perceptual quality metrics poorly correlate with scientific validity, while language models lack domain-specific verification capabilities. To address this gap, we propose the \textbf{S}cientific \textbf{I}mage \textbf{U}tility and \textbf{U}pgradability \textbf{A}ssessment (\textbf{SIU$^2$A}) framework, which introduces two complementary dimensions for scientific image evaluation. \textbf{Utility} encompasses \textit{error detection} (identifying scientific inaccuracies) and \textit{correction feasibility} (assessing whether errors can be reliably repaired). \textbf{Upgradability} measures the quality of correction. We categorize scientific image corruption into four fundamental types: Detail Distortion, Incompleteness, False Content, and Entity Confusion. Based on this taxonomy, we construct SIU$^2$A-Benchmark, a dataset with expert annotations for error identification and repair. The framework implements a two-stage evaluation protocol: the \textit{Utility} stage evaluates error detection capability and repair instruction generation, while the \textit{Upgradability} stage assesses whether corrections faithfully restore scientific validity without compromising existing accurate information. Experiments reveal that current multimodal systems exhibit significant limitations in both scientific error assessment and faithful correction, exposing a fundamental gap between visual perception and scientific usability.

Multimodal LLMs are increasingly used to assist scientific peer review, where a core requirement is verifying whether claims in a paper are supported by its evidence. Prior work has shown that models perform substantially better at this task when the evidence is a table than when it is a chart of the same underlying data. This raises the question of whether models fail to extract information from charts, or do they extract it but fail to use it when forming their prediction? We study this question through layer-wise linear probing and attention analysis on three open-weight VLMs over table and chart evidence, representing the same underlying data. We find consistent evidence for the latter. Chart information is encoded in the models' intermediate representations but does not reach the prediction position, a gap that is absent for tables and holds across all conditions tested. Attention analysis further reveals that this disconnect takes two architecturally distinct forms across model families. These findings reframe the table-chart gap as a failure of how encoded visual information is routed at prediction time, rather than a failure of encoding itself.

Small-to-medium scientific datasets place machine learning pipelines under two compounding pressures. Single-run feature selection produces feature sets that change substantially under small perturbations of the training data, and any procedure that uses the same data for selection, tuning, and evaluation produces optimistically biased performance estimates. The two failure modes are routinely treated as separable, but in the regimes where scientific data live, they interact: an unstable selection inflates the variance of an already-optimistic score, and standard remedies for one rarely address the other. RobustModelMaker is a Python framework that couples bootstrap stability selection with strict nested cross-validation, performs all preprocessing and selection inside each fold, and produces a stability-tested feature subset together with a leakage-safe performance estimate. The framework supports nine algorithms across binary classification, multiclass classification, and regression. Behaviour is verified by a deterministic test suite spanning unit, performance, and reproducibility checks on three real scientific datasets comparing to three alternative selectors (ANOVA F-test, recursive feature elimination with cross-validation, and Boruta) on both predictive score and a Jaccard measure of selection stability. RobustModelMaker is competitive in score with the best alternative selector on each dataset, and occupies a position on the joint score-stability frontier that none of the alternatives match across all three task types. Two example applications, ovarian cancer biomarker discovery from the PLCO Trial and critical-temperature regression on the UCI Superconductivity Data, illustrate how the framework is used in practice and what trade-offs become visible when stability is treated as a first-class deliverable rather than an emergent property.

Scientific discovery is not only answer generation but revision of the representational regime in which evidence, artifacts, operations, and verifiers are typed. We develop a category-theoretic account of agentic discovery for materials science. In a fixed regime b with schema category S_b, the system state is a copresheaf I_t: S_b -> Set, and provenance is the category of elements \int_{S_b} I_t. Fixed-regime operation is an update on such states, endofunctorial only when provenance-preserving refinements are specified and preserved. Discovery is instead a verified regime transition u: S_b -> S_b': old artifacts are preserved, transported by the left Kan extension Lan_u I_t, and compared with the post-transition state to identify residual content beyond functorial transport. This separates retrieval, search, and discovery without subjective novelty. We instantiate the framework in two systems. In Builder/Breaker, a protein-mechanics world model is revised under a Minimum Description Length gate; the accepted law expresses within-chain flexibility as all-mode elastic compliance conditioned by slow collective-mode participation, or mode-conditioned compliance. In CategoryScienceClaw, typed skills, artifacts, open needs, workflow mutation, gates, stress tests, and public discourse become a proof-carrying knowledge-computation graph. A fiber-network example records candidate models, rejected alternatives, an AIC gate, perturbation tests, and an accepted orientation-tensor anisotropic stiffness surrogate over an isotropic fiber-count descriptor. Together, the cases show how category theory can be both a mathematical language for discovery and an engineering specification for self-revising AI discovery systems.

Scientific discovery demands intelligence, perseverance, and serendipity across vast search spaces. Today, top scientific capabilities remain siloed--one AI system for biological analysis, another for clinical reasoning, mathematical derivation, or materials simulation--and no pre-designed team can anticipate every skill a question will need. Science Earth is a planet-scale scientific runtime in which any capability--a simulation cluster, a wet-lab robot, a proof engine, a single-cell pipeline--can connect to any other, with collaboration structure emerging from the question itself. Its underlying EACN protocol lets capabilities discover one another, negotiate task ownership, and adjudicate across incompatible evidentiary standards without prior knowledge of who will meet whom. This shifts the organizing challenge from workflow design to open-ended connectivity. Two runs validate this under structurally distinct conditions. In a trans-Pacific higher-order Kuramoto synchronization study, agents identified and corrected a closure-ratio assumption in Ott-Antonsen analytic theory that fails outside the Lorentzian limit, within thirty minutes. In an eight-agent single-cell run on the 4.88M-cell Kang 2024 pan-cancer atlas, heterogeneous capabilities coupled over a 64.9-hour window with one structural external instruction, producing three new result layers and anchoring findings against an independent wet-lab study on an adjacent CCR8- TIGIT+ Treg subset. These cases are a first empirical reading, not a benchmark sweep. They show that when AI capabilities are truly connectable and coordination emerges from the problem, scientific reasoning becomes a distributed, self-correcting process--a step towards scaling AI-native discovery to the planet.

We argue that the AI community is now ready to move beyond benchmarking and consolidate scattered efforts in model analysis into a systematic discipline, a direction we term Model Science. Complex AI models now serve billions of users, yet our understanding of how they work lags far behind our ability to deploy them. Decades of benchmark-driven research have delivered remarkable progress: extensive leaderboards, a wide range of performance metrics, tracking capability gains across diverse tasks; yet this success has also revealed the limits of benchmarks as they tell us whether models perform but not why they succeed or fail, they miss critical failure modes, such as hallucinations or shortcuts. Precedents from established sciences point the way forward: cognitive science shows that understanding complex systems requires complementary levels of analysis; neuroscience demonstrates that deep study of single cases reveals what population studies miss; medicine teaches that specialised training must develop alongside research practice; and agriculture models how shared infrastructure and principles enable cumulative progress. These lessons inform three foundations for Model Science. First, we propose to consolidate research around four functional perspectives: Verify, Explore, Steer, and Refine that address complementary questions about model behaviour. Second, we discuss the required infrastructure for cumulative knowledge: catalogues of datasets, models and findings. Third, we highlight the need for deep analysis of individual model instances, not just model families, because single cases can reveal what population studies miss.

While Reasoning Language Models (RLMs) are rapidly emerging as powerful tools for scientific research, their impact is primarily concentrated in "hard science" fields. The slow -- or lack of -- adoption of RLMs in other branches of science is causing a widening gap in research productivity. In this survey, we provide the first comprehensive analysis of RLM adoption across 28 scientific disciplines following the classification used by the European Research Council (ERC), spanning the Social Sciences and Humanities, Physical Sciences and Engineering, and Life Sciences. We examine how RLMs are developed, evaluated, and applied across disciplines. Furthermore, we introduce a maturity-oriented assessment framework based on available domain-specific development and evaluation resources, revealing substantial disparities in RLM maturity that become even more pronounced when only publicly available resources are considered. Finally, we highlight current implementation paradigms that are gaining popularity across disciplines, current challenges, and future directions in enabling RLM adoption across science.

Modern Machine Learning (ML) and Artificial Intelligence (AI) models, especially large language models (LLMs), are increasingly used to generate scientific hypotheses and mechanistic explanations from observational data. This position paper argues that in the high-dimensional proxy regimes where modern ML excels, mechanistic learning is generically underdetermined: many incompatible mechanisms induce essentially the same observational relationships on the support of the data, so predictive success and coherent explanations are insufficient evidence of mechanism discovery. This underdetermination becomes uniquely hazardous with large language models (LLMs), which tend to collapse large equivalence classes of explanations into a single fluent narrative. This paper proposes concrete standards for ``mechanistic ML,'' and argues these norms are necessary if LLM-centered workflows are to support science rather than merely simulate it.

Large language models have rapidly evolved in multilingual competence and reasoning capacity, enabling their integration into Social Sciences and Humanities research workflows. Yet existing evaluation paradigms remain anchored in task-based NLP benchmarks and fail to address interpretive validity, cultural situatedness, and epistemic mediation. This paper reconceptualizes multilingual reasoning LLMs as hermeneutic instruments that actively structure meaning production across linguistic and cultural contexts. Drawing on hermeneutics, philosophy of technology, science and technology studies, multilingual NLP research, and computational social science methodology, we develop a theoretically grounded framework for evaluating multilingual reasoning in Social Sciences and Humanities (SSH) research. We articulate a rigorous experimental protocol with operationalized metrics for cultural alignment, cross-lingual stability, and reasoning faithfulness, along with transparency requirements tailored to interpretive research tasks. We illustrate the framework through a concrete application scenario involving multilingual political discourse analysis. The paper contributes a conceptual and methodological foundation for responsible integration of multilingual reasoning LLMs into computational social science infrastructures.

Science news is an important medium to communicate discoveries between the research communities and the public. Yet, most metrics for generated or summarized text evaluate semantic similarity and factual consistency, but do not measure how much knowledge readers learn from the news. We introduce KnowledgeGain, a metric that evaluates the quality of science news by measuring how much knowledge readers gained after reading it. To evaluate the metric, we first performed a controlled human study and showed that the metric successfully captures the differential knowledge gained by human readers reading different types of science media. The data allowed us to calibrate a prompt-only LLM reader simulator. We use it to rank and filter candidate articles before human evaluation. A second human study shows that articles selected with this simulator improve post-reading accuracy and normalized KnowledgeGain over a strong generation baseline. Our work is a step toward generating science news that better meets the knowledge and comprehension goals of Bloom's Taxonomy.

Knowledge graphs over corpora of inter-referencing documents - scholarly papers, legal opinions, policy briefs - encode the topology of reference but not its stance. The standard representation collapses a rich evaluative relation into an untyped edge, losing the very content that supports community-level queries about how one document is received by another. We propose the claim network: a representational pattern in which each cross-document reference is reified as a typed claim, carrying source, target, claim text, and a four-class stance label grounded in the citation-intent literature. We give a construction pipeline applicable to any corpus of scholarly inter-referencing documents and instantiate it on a corpus of 127 papers in 3D point cloud semantic segmentation, producing a network of 8,260 typed claims. Three downstream task families demonstrate what the network enables: retrieval signal augmentation, aggregated-stance summarisation, and topological analytics. Head-to-head evaluation against standard Retrieval-Augmented Generation (RAG) baselines shows that the gain over flat retrieval is the gain from the right intermediate representation rather than the wrong one.

Scientific figures are among the most effective means of communicating complex research ideas, yet producing publication-quality illustrations remains one of the most labor-intensive parts of paper preparation. Existing automated systems each target a single figure type under text-only input, leaving the diversity of types and conditions researchers actually use unaddressed; their raster outputs further cannot be locally revised. Because scientific figures are structured compositions of discrete semantic components, the localized errors generators produce on such layouts demand not a stronger backbone but a harness. We instantiate this harness in two complementary systems: Crafter, a multi-agent harness for figure generation that generalizes across figure types and input conditions without architectural changes, and CraftEditor, which applies the same pattern to convert raster outputs into editable SVGs. Moreover, we introduce CraftBench, a benchmark spanning three figure types and four input conditions with human quality annotation. Experiments show that Crafter substantially outperforms both standalone generators and the agentic baseline on PaperBanana-Bench and CraftBench, with ablations confirming each component's independent contribution; CraftEditor faithfully converts outputs into editable SVGs that surpass all baselines. Our code and benchmark are available at https://github.com/HaozheZhao/Crafter.

A plausible future mathematical claim must satisfy two constraints: it should follow the direction of prior work and respect the formal dependencies that constrain what can validly follow. Existing approaches typically model only one of these sources, producing claims that are either weakly grounded or insufficiently motivated. We introduce grounded future mathematical generation, where the goal is to generate a plausible future theorem-like claim for an anchor paper using two complementary sources of context: its scientific citation graph and aligned formal theorem dependency graph. To address this setting, we propose COMPOSE, a dual-graph framework that conditions a language model on both scientific citation context and formal theorem structure. To support this setting, we construct a dataset of 108K paired scientific-formal graph examples from arXiv and Mathlib, together with a benchmark of 47K future papers from 2024--2025. Experiments show that COMPOSE outperforms strong baselines on retrieval to real future papers and achieves the best overall performance under LLM-judge evaluation, producing more grounded and mathematically richer outputs. These results show that future mathematical generation benefits from combining scientific context with formal structure. Project page is available at https://david-busbib.github.io/COMPOSE-page/.

Self-improvement at scale has been a longstanding goal for reasoning models, and there are two natural places to do it: at test time, through verification-refinement (V-R) loops; and at training time, through self-training methods. Both are gated by the same bottleneck: the verifier. V-R loops stall when verifier scores inflate while accuracy stagnates, and when feedback is too generic to act on; self-training fails similarly when bad self-generated data are added to training. Better verification would unlock both, but the capability we want to train, i.e., catching self-generated errors, lacks training signal. To address this challenge, we propose self-trained verification (STV). Our key observation is that, while a model cannot catch these errors alone, it can when shown the reference solution. We turn this asymmetry into a supervision target and train the verifier to imitate a more informed version of itself. At test time, STV substantially improves V-R loops on hard problems, while alternatives (e.g., SFT, RL on verifier scores, and even meta-verifiers) do not. STV roughly doubles accuracy on hard math and lifts it 14x on scientific reasoning tasks (1.5% to 21%). At training time, we additionally train the generator using RL with STV verifier's feedback inside the V-R loop - a procedure we call verifier-in-the-loop training (ViL). Starting from an RL-converged generator, ViL yields a further 33% gain in pass@1. More notably, the generator's standalone pass@1, with no verifier at test time, climbs 30% relative past where standard RL had converged. Hence, the next frontier in reasoning on hard problems may lie in how we train for and with verification. Website: https://ar-forum.github.io/stv-webpage

Scientific discovery is an inherently creative and uncertain process, requiring reasoning beyond the recall of known knowledge. While many benchmarks have been proposed to evaluate large language model (LLM) performance on deep research tasks via multi-hop retrieval, their innovative reasoning abilities essential for true scientific discovery remain largely untested. We introduce a benchmark framework for evaluating model performance in scientific discovery and reasoning, building up from a raw problem to the classical null hypothesis test. In our framework, models initially receive only the topic and research question from a recent paper, with technical details progressively revealed. At each stage of information disclosure, the model is tasked with generating hypotheses that address the research question, which is compared with the conclusions from the original paper and evaluated via automated semantic similarity of constituent atomic claims. This progressive evaluation of semantic divergence from ground-truth conclusions enables assessment of a model's innovativeness (under minimal information) to grounded reasoning capabilities (under full experimental details), both critical for using LLMs for scientific discovery purposes. Our framework provides a foundation for systematically evaluating scientific reasoning and discovery capabilities in LLMs, crucial for advancing the development of next-generation AI scientist/co-scientist systems. Specifically, here we evaluate GPT-5, GPT-5.4, Gemini 2.5 pro, and Gemini 3.1 pro preview across 45 papers spanning bioactive materials, mechanical materials, and nanomaterials. We find that GPT-5.4 and Gemini 3.1 pro outperform their previous generation counterparts as expected, and GPT-5.4 in particular maintains 0.7 F1 score alignment with ground truth conclusions even under minimal context.

Marine lead (Pb) and its isotopes are critical tracers for ocean circulation and anthropogenic pollution, yet in-situ observations remain costly and sparse. While vast historical records exist, they lie buried within the unstructured content of academic papers, creating "data silos" inaccessible to comprehensive analysis. Manual extraction is unscalable, while general-purpose Large Language Models (LLMs) lack the necessary domain-specific knowledge, leading to hallucinations and scientifically invalid outputs. To address this, we introduce an expert-guided adaptation approach that enables LLMs to perform rigorous scientific data extraction without fine-tuning. We operationalize this approach through Compass, an LLM agent framework enhanced by a Knowledge Tree co-designed with marine scientists, which decomposes complex tasks into verifiable steps, guiding the agent's reasoning to ensure scientific validity. Deploying Compass across a corpus of over 230,000 relevant open-access papers, we successfully extract 3,751 previously unincorporated Pb records. This effort establishes the largest integrated marine Pb database to date. Beyond standard metrics, Compass demonstrates superior reliability through multi-layered validation, achieving 92% accuracy as confirmed through expert manual verification. The newly integrated data expand coverage in previously under-sampled regions such as the East China Sea and the Southern Ocean, providing an enriched data foundation for future scientific discoveries. We release an interactive visualization platform to facilitate open scientific access. Our work demonstrates that expert-guided agents can effectively bridge the gap between general-purpose LLMs and high-stakes scientific domains, enabling scalable data discovery in geosciences.

As multimodal language models play an increasingly important role in scientific research, materials science offers a critical testbed due to its interdisciplinary, multimodal, and application-driven nature. However, existing materials benchmarks mainly focus on property prediction, knowledge QA, or characterization understanding, leaving the broader reasoning process from materials knowledge to application underexplored. To fill this gap, we present OmniMatBench, a human-calibrated multimodal reasoning benchmark for materials science. OmniMatBench contains 3,171 expert-curated QA and calculation problems across 19 materials-science subfields, spanning fundamental materials knowledge, structural and engineering materials, materials processing and manufacturing, and functional and applied materials. We evaluate 13 open-source and closed-source MLLMs and find that the best model achieves only a 0.372 overall score, revealing a substantial gap in current materials-science reasoning. Further analysis shows strong variation across subfields, fixed reasoning heuristics, uneven materials knowledge, and limited high-level knowledge application under formula-, retrieval-, and code-assisted settings. OmniMatBench provides crucial insights into the capabilities and limitations of current MLLMs and establishes a foundation for reliable AI assistants in materials-science research.