Papers for
ai research teams
Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.
AI system breaks down and tracks research idea steps for novelty
Ground, Iterate, Trace: Decomposing AI Research Ideation
Abstract: Large language models have been widely used to generate plausible research ideas, yet it remains unclear what makes automated ideation effective and how generated novel proposals arise from prior scientific work. We introduce the Creative AI Researcher (CAR), an agentic framework combining literature grounding, iterative development and explicit conceptual operators. CAR constructs a semantic knowledge graph from retrieved literature and equips an ideation agent with graph-retrieval tools and four operators: bisociation, analogy, exploration and elaboration. Each operator call is checked against structural acceptance conditions and recorded, producing a trace of intermediate transformations and declared literature sources. We evaluate CAR with two language-model backbones across 40 LiveIdeaBench research topics and two execution benchmarks. CAR achieves significantly higher methodological novelty (MN) than single-shot prompting and The AI Scientist. Its methodological novelty does not differ significantly from the Research Ideation Agent, while its feasibility scores are higher on both backbones. With a shared experiment executor, CAR achieves the highest pairwise win rate among the evaluated AI methods on FML-bench-Lite and BioXArena. In a ten-condition ablation study, MN point estimates favor CAR over one-step proposal on both backbones, but no MN contrast survives correction for multiple comparisons. The results do not establish a consistent quality benefit from explicit operators beyond free-form iteration. CAR makes the sequence of operations and declared literature sources inspectable, separating judged idea quality from procedural traceability.
AI suggests new ways to fix intestinal cell barriers physically
Structured AI ideation yields physical strategies for epithelial barrier repair
Abstract: AI systems increasingly support literature analysis, candidate evaluation and hypothesis refinement, but enabling researchers to explore research directions beyond established disciplinary approaches remains challenging. We developed Abelio, an AI system designed to generate novel research directions by expanding investigators' avenues of inquiry. Abelio reformulates research objectives, uses these alternative formulations to identify relevant elements outside conventional disciplinary boundaries, and refines the resulting hypotheses. Each stage integrates divergent exploration with convergent selection through a dedicated subsystem. In quantitative evaluations, proposals generated by the Ideation Core (Abelio's hypothesis-generation core) outperformed those of four frontier large language models in scientific soundness and hypothesis diversity, and three of four in distance from existing literature. Subsystem ablation showed that each subsystem contributes to the property it targets, with the contribution of Element Selection measurable before Refinement. To investigate intestinal epithelial barrier restoration, Abelio proposed physical interventions involving membrane-tethered poly(ethylene glycol) (PEG) and cholesterol-anchored complementary DNA. Both proposals were archived before experimentation and subsequently tested independently in damaged Caco-2 monolayers. PEG preserved barrier function in a membrane-surface-dependent manner, whereas complementary DNA enabled sequence-specific restoration that was reversible on demand. These findings demonstrate how structured research ideation can identify experimentally testable approaches beyond established disciplinary conventions.
Scaling rules reveal how vision language models combine images and text
Scaling Laws for Vision-Language Fusion
Abstract: Vision-language models (VLMs) have emerged as central foundation models combining multimodal perception with text comprehension. While VLMs are traditionally built via late fusion by continued pre-training an off-the-shelf language model on multimodal (image-text) data, mixing modalities from the start of pre-training (early fusion) has recently gained traction. However, the fundamental trade-offs between these paradigms remain poorly understood, resulting in expensive and heuristic-driven training recipes. In this paper, we systematically analyze the learning dynamics of these multimodal pre-training processes, showing that validation losses during vision-language fusion can be modeled by separating the base learning trajectory, where different data types contribute at distinct rates, from a distribution shift driven by the competition between modalities. We derive a scaling law that unifiedly describes these dynamics, enabling accurate loss predictions across mixture ratios, training steps, and model sizes. Crucially, our law characterizes the trade-off between early and late fusion: under a fixed data budget, early fusion preserves text performance while late fusion favors vision performance, yielding a trade-off that increasingly tilts toward early fusion as model size scales. Extensive experiments confirm that our scaling law reliably captures multimodal learning dynamics and governs fusion decisions.
Auditors face difficulty detecting sabotage in ai safety research logs
Auditors Struggle to Catch Sabotage In End-to-End Transcripts of AI Safety Research
Abstract: As AI systems are increasingly used to automate AI R&D, misaligned systems could sabotage research intended to improve model safety and alignment. To stress-test defenses against such sabotage, we use a human expert as a proxy saboteur on AI safety research. We conduct a red-team/blue-team game where a human researcher runs AI safety experiments from scratch and auditors try to tell which were sabotaged and how. Given access to the full research transcript and the ability to run experiments, human auditors mostly fail to distinguish sabotaged from benign research: they achieve an AUROC of 0.59, not significantly better than chance, and fully identify the specific sabotage in only 17% of cases. Despite not being allowed to run experiments, the best frontier models outperform human auditors (Fable 5.1: AUROC 0.73, 32% full identification) within noise and weaker models perform comparably. When the saboteur chooses which experiments are attacked, rather than being required to sabotage every question given, sabotage identification drops further. Our results suggest that research sabotage by expert-level agents is difficult to detect with current auditing methods and provide a new setting for studying research sabotage.
AI discovery improves when asking the right questions early
Asking the right question is key to AI discovery in a chain of hypotheses
Abstract: Artificial intelligence (AI) scientists pursue discovery through a chain of hypotheses, fitting candidate models to existing observations and often choosing experiments where those models disagree most. Yet when early data miss a narrow feature, model selection favors simple curves that agree across the unmea- sured gap, so disagreement-guided search has no reason to look where the new phenomenon lies. When we tested two frontier AI scientists, Claude Sonnet 5.5 and GPT-6.1-Sol, across 32 closed-loop trials with a narrow peak omitted from the starting data, both sampled inside the peak region less often than random search. Surprisingly, the agent that followed model disagreement more closely missed the peak in all 32 trials, whereas discovery succeeded when experiments probed the unmeasured gap early and followed up across the anomaly. Asking the right question, rather than testing only where current models disagree, is therefore key to AI discovery in a chain of hypotheses.
Naive language models spark more creative scientific ideas
The Value of Not Knowing: How Fresh Eyes through Naive LLMs Fuel Creativity
Abstract: Many major scientific advances come from approaching a problem with a new or novel perspective. While large language models (LLMs) with internet-scale knowledge seem poised to take such perspectives, their consistent generation remains elusive. Instead, models are drawn to dominant views and preexisting, in-field bias. To explore the potential of fresh eyes to increase creativity, this paper proposes naive LLMs, models that are made intentionally ignorant of a specific topic. Through deprivation in training or intentional forgetting, naive models are unburdened by existing dogma, allowing them to be free to make new associations and conjectures. These naive models can then be paired with a stronger mentor model that can refine their unique perspectives into concrete proposals. For the 1B and 7B models tested in this paper, naive models consistently produce seed ideas that are more novel and useful than baseline counterparts, and mentor refinement improves the ideas even further. Thus, contrary to the prevailing tradition of scaling up training data, we demonstrate a case where less information is actually more powerful, in particular when the aim is to discover genuinely outside-the-box solutions.
Policy optimization improves reward handling in multi-reward reinforcement learning
Whitened Advantage Summation Policy Optimization: Correlation-Aware Aggregation for Multi-Reward Reinforcement Learning
Abstract: Reinforcement learning for language models often combines several rewards by normalizing each one and adding them with fixed weights. When the rewards are correlated, the variation they share is counted once for every reward that carries it. In terms of the ZCA-decorrelated components of the reward signal, the weighted sum applies weights $Σ^{1/2} w$ rather than the chosen weights $w$, where $Σ$ is the reward correlation matrix. We study WAS-PO, which whitens the normalized advantages with $Σ^{-1/2}$ before adding them, so that the chosen weights apply to those components. We derive the scalarized objective whose gradient it follows when normalization statistics are held fixed, show that its weights on the raw rewards can be negative, give an exact two-reward condition under which the lighter-weighted reward is reversed, and give a loose bound on the error from estimating $Σ$ during training. On tool calling with Qwen2.5-1.5B-Instruct, WAS-PO reached a higher mean training reward than GDPO in all seven seeds with three rewards (3.8% higher) and all three seeds with four (3.9% higher). WAS-PO also replaces GDPO's per-prompt normalization with batch-level normalization, and we did not separate the two changes with three or four rewards. The training-reward gains did not carry over to held-out accuracy, which was slightly lower than GDPO's with three rewards (not significant) and 0.03 lower in absolute terms with four (two seeds evaluated). With two rewards, WAS-PO trailed GDPO in training reward, and a three-seed control weakly suggests that the gap comes from batch-level normalization rather than whitening. In a math setting where the two rewards were nearly identical by construction, WAS-PO drove down the lighter-weighted reward, as the sign condition predicts.
Claimnov checks which parts of ai science ideas are truly new
Do AI Scientists Know What's New?
Abstract: A major challenge facing AI scientific discovery pipelines is how to determine when a generated idea is meaningfully novel. Ideally, novelty checks in these pipelines must be able to reject non-novel ideas while not rejecting too many novel ideas, and they should also be able to distinguish which parts of a proposal are non-novel, as opposed to opaquely discarding entire proposals. We design an analytical framework for evaluation of novelty checks with temporal ground truth at both the proposal- and claim-level, using four datasets of 1,060 proposals built from published papers, reviewer disputes, and generated ideas matching ones the field went on to publish. Auditing existing novelty checks with this framework, we find that checks with high rejection rates on non-novel work also tend to have high false alarms on ideas that are later realized by human researchers, and observe that checks built to evaluate proposals as a whole can point to claim-level prior art only by rejecting the whole proposal. We therefore introduce ClaimNov, a fine-grained, interpretable claim-level novelty module. ClaimNov catches 97% of disguised papers at a 1.5% false-alarm rate, and its claim-level novelty profile cuts the LLM judge's false alarms from 11.8% to 1.5%. Because its verdict rests on claim-level evidence, ClaimNov also keeps proposals that build on established contributions, rejecting only 3% of ICLR-accepted proposals with a reviewer-disputed contribution and keeping 91 of 100 novel proposals with one spliced prior contribution, against 76 for the strongest existing check. Code and data are available at https://github.com/emmatliu/claimnov.
Large language models use debate and graphs to generate better research ideas
Distill, Debate, Discover: Grounding Multi-Agent Research Ideation in Graph-Distilled Insights
Abstract: Large language models (LLMs) are increasingly used to support research idea generation and scientific discovery. Multi-agent debate offers a collaborative approach to this task by bringing together diverse perspectives to develop research ideas. However, a key challenge is how to equip agents with distinct and meaningful expertise to guide their contributions. Inspired by how researchers draw on accumulated research experience during collaboration, we introduce InsightAgents, a framework that constructs literature-grounded agent expertise by distilling insights from graphs of related papers. Guided by insights that trace evolving research approaches, reveal underlying assumptions, and identify unresolved challenges, within a line of research, agents jointly identify a research problem and propose, challenge, and refine potential approaches through debate to develop a research proposal. We evaluate InsightAgents against three research idea generation frameworks on 100 NLP research topics through pairwise comparisons by two LLM judges. In both online and offline knowledge settings, InsightAgents achieves the highest average score across the five evaluation dimensions under both judges.
Multi-agent coordination improves scientific discovery with language models
MELD: Recursive Self-Improvement via Hierarchical Coordination for Multi-Agent Scientific Discovery
Abstract: Multi-agent scientific discovery with large language models raises a coupled learning problem: research assignments determine both which experiments are pursued and the experiences from which agents adapt. We introduce MELD, a unified hierarchical test-time training framework that jointly adapts a team-level Coordinator policy and agent-level individual Scientist policies within a single discovery process. The Coordinator generates a joint portfolio of search branches and research directives over a shared derivation tree, while Scientists develop hypotheses and execution plans implemented by a shared frozen Executor. Both levels receive alternating GRPO-style parameter updates under a shared smooth best-of-budget objective, allowing team organization and individual research behavior to co-evolve in a recursive self-improvement loop. Our analysis establishes objective alignment for the hierarchy, characterizes conditions for coordination gains over independent parallel discovery and the first-order benefit of private adaptation, and derives a finite-budget regret bound for adapted models. On 22 MLGym and MLE-bench tasks, MELD is best on 20 of them and achieves the highest aggregate normalized scores on MLGym, using a 4B backbone for coordination and research. Most baselines use proprietary Scientists, and all systems share one frozen Executor.
Large models improve spotting research gaps but struggle to create them
Where Research Ideation Begins: Forward-Citation Supervision for Actionable Limitation Generation
Abstract: Automated research ideation has advanced rapidly, with multi-agent workflows that generate whole research ideas end to end. These frameworks leave largely unexamined the step that seeds a good idea: identifying an actionable limitation from literature. In this paper, we focus on this first step in the research ideation pipeline. Starting from a pilot study, we find that as large language models (LLMs) scale from 0.6B to 32B parameters, they become far better at recognizing which limitation later work addressed, but no better at generating one, suggesting that the bottleneck lies in the supervision signal rather than model capacity. We therefore construct ActLim-Bench from forward citations, yielding 18,279 evidence-linked supervision triples with human verification. On this signal, we propose a three-stage post-training pipeline: supervised fine-tuning, verifier-filtered rejection-sampling fine-tuning, and reinforcement learning with the verifier as an online dense reward. Experimental results show that our method achieves the highest recall score compared to existing methods among all frontier models. We further show that our generated limitations are preferred in 64.9% of non-tied human comparisons and improve downstream ideation systems. We will release our data and code upon publication.
AI agents make progress on open-ended scientific discoveries in Station
Can AI Agents Make Open-Ended Scientific Discovery? Evidence from Station
Abstract: Recent AI systems have made rapid progress in scientific discovery when given well-defined metrics, but whether they can autonomously undertake open-ended scientific discovery remains unclear. We investigate AI's ability to tackle open-ended tasks in Station, an open-world environment in which multiple agents simulate a scientific ecosystem. To tackle challenges specific to open-ended tasks, we propose augmenting Station with two mechanisms: a Supervisor mechanism and periodic Meta Reflection, which encourage persistent exploration even when intermediate metrics are lacking. We construct open-ended tasks from three recent oral papers presented at ICLR. We give agents the main research question studied in each paper while withholding the paper's results and disabling web access. We then measure how many of the original findings-partitioned into individual criteria-agents rediscover. We find that Station rediscovers 62.7% of the criteria on average, compared with 15.4% for Codex Multiagent-v2 and 14.4-20.6% for AI Scientist-v2. Ablation and behavioral analyses indicate that adding the two mechanisms together improves research coverage and continuity. We further evaluate Station on two open-ended tasks without oracle papers and find that some of the discoveries made by the agents closely match discoveries reported by researchers after the knowledge cutoff date. Together, these results indicate that a suitable environment can enable agents to autonomously make meaningful progress in open-ended scientific discovery.
Rised guides training of generalist ai agents across multiple environments
RISED: RubrIcs for agentic multi-environment Selection and sElf-Distillation
Abstract: Training a single LLM agent jointly across diverse interactive environments has attracted increasing attention as a route to generalist agents. Existing curriculum and data-selection strategies often allocate training at the environment level or prioritize local reward-based signals, without explicitly considering relationships between current rollouts across environments for prompt-group selection. Meanwhile, as environments are learned at different rates, all-failure and all-success rollout groups can coexist within a batch, leaving those data without group-relative reward signals. Both challenges highlight limitations of relying solely on scalar rewards in multi-environment RL: they provide limited information about cross-environment relationships and no within-group reward contrast when rewards are identical. This motivates richer textual feedback, such as rubrics describing rollout behaviours, to guide learning. Beyond rubrics' usage as reward, we repurpose rubrics to guide both online data selection and policy supervision. An LLM judge tags each rollout using a predefined rubric vocabulary shared across environments. The resulting profiles guide the selection of data that aligns with the overall behavioural composition of the mixed-environment batch while limiting overlap with already-selected data. Available positive rubrics (describing desired behaviours) provide privileged context for an on-policy self-distillation teacher, supplying additional token-level supervision, while negative rubrics (describing undesired behaviours) guide subsequent rollout generation away from recurring failure modes. Together, these components form RISED. Across model backbones, RISED achieves the highest mean pass rate across environments and ranks first or second in every individual environment. Rubric-based analysis of RISED can further characterize the behavioural changes accompanying these gains.
Scale equalized rollout allocation improves reinforcement learning gradient estimates
SERA: Scale-Equalized Rollout Allocation for Maximum Likelihood Reinforcement Learning
Abstract: Maximum Likelihood Reinforcement Learning (MaxRL) targets prompt-wise log-success and has shown strong performance on reasoning tasks. Under finite rollout budgets, however, the estimator used by MaxRL attenuates each prompt's likelihood gradient by a factor that depends on its success probability and rollout count. Under uniform rollout allocation, the common rollout count fails to compensate for success-dependent attenuation, leaving low-success prompts more strongly attenuated and distorting their relative contributions to the expected aggregate gradient. We introduce SERA (Scale-Equalized Rollout Allocation), which redistributes a fixed rollout budget to approximately equalize these finite-rollout scaling factors. Building on our theoretical analysis of how finite rollouts distort prompt-wise likelihood gradients, we formulate the allocation as a fixed-budget max--min problem, derive a waterline solution to its continuous relaxation, and introduce a multiplicity correction to remove the additional prompt weighting induced by heterogeneous rollout counts. Experiments show stronger alignment with exact likelihood gradients in a controlled ImageNet setting and improved multi-sample solution coverage over MaxRL on maze navigation and mathematical reasoning under matched training rollout budgets.
Deep research agents use rubric feedback to improve process and results
Dr.Credit: Rubric-Grounded Process Credit Assignment for Deep Research Agents
Abstract: Rubric-based tasks are increasingly addressed through reinforcement learning (RL), with rubric scores used as training rewards. However, these rewards typically supervise final answers without distinguishing the contributions of intermediate decisions. Many existing credit assignment methods rely on ground-truth answers to define process rewards, limiting their applicability to open-ended tasks without canonical solutions. To address this limitation, the proposed rubric-grounded credit uses task requirements as a shared reference for final answer evaluation and process supervision. The information returned by tools is assessed for the additional support it provides toward satisfying each rubric relative to that rubric's history of accepted support. By referencing these histories, credit distinguishes new support from evidence already present in the trajectory while recognizing partial support for each rubric. Dr.Credit uses rubric-grounded credit to supervise intermediate tool turns in an RL framework for deep research agents. The resulting process advantages are combined with GRPO outcome advantages to guide research decisions while retaining supervision of final-report quality. Evaluations on four in-domain and out-of-domain benchmarks show that Dr.Credit outperforms the evaluated open deep research baselines on every primary metric and submetric. Meanwhile, with an 8B-parameter backbone, the trained agent achieves average performance competitive with the evaluated frontier proprietary models. Further analyses suggest more efficient evidence acquisition and higher-quality reports under limited research-turn budgets, motivating the extension of rubric-grounded process supervision to a broader range of rubric-based tasks.
Language models learn to generate scientific ideas with greater impact
Learning to Ideate for Scientific Impact
Abstract: Scientific ideation is increasingly mediated by large language models, but current ideation systems are usually trained and evaluated on immediately judgeable proxies such as novelty, clarity, and feasibility. This leaves open whether delayed signals of scientific uptake can be used as feedback for steering models toward research directions with higher expected \emph{impact}. We study this question using citation-normalized impact as a noisy but scalable proxy for scholarly uptake. We construct a large-scale dataset from over 100K computer science papers by extracting goal-conditioned idea descriptions and assigning each paper an ordinal, year-normalized citation label. We then train a goal-conditioned reward model to predict citation-impact labels from research goal and idea pairs, and use this reward to align an idea generator through supervised fine-tuning followed by reinforcement learning. To reduce circularity, we evaluate generated ideas with a held-out, reference-grounded protocol that compares model outputs against historical ideas under the same research goal and weights judgments by the reference idea's citation-impact label. Experiments show that our RL-tuned model consistently produces ideas with higher estimated impact than both the base model and supervised fine-tuning baselines. Our findings position scientific impact as a practical, outcome-grounded feedback signal for aligning LLMs in open-ended scientific discovery.