# Research Papers — 2026-05 · bundle page 85 of 326 > Full-text markdown of papers #4201–4250 of 16266 published in 2026-05, newest first. AI-readable mirrors of the full 16266-paper 2026-05 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-05/page/85 [← Previous bundle](https://aipod.dev/papers/archive/2026-05/page/84.md) [Next bundle →](https://aipod.dev/papers/archive/2026-05/page/86.md) ### Label-NTK Alignments and A Tighter Convergence Bound in the NTK Regime Authors: Ruchirinkil Marreddy, Chaoyue Liu Published: 2026-05-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.25275v1 The Neural Tangent Kernel (NTK) framework explains optimization in over-parameterized neural networks via approximately linearized dynamics, yielding exponential convergence guarantees. However, existing results are often overly pessimistic and do not match the fast training in practice, as they depend on the smallest NTK eigenvalue, which is typically extremely small in practice. In this work, we develop sharper convergence guarantees by characterizing the interaction between data labels and the NTK eigen-spectrum. We identify two key phenomena, Label-NTK alignment and Residual-NTK alignment, showing that projections of labels and residuals onto NTK eigenvectors scale with the corresponding eigenvalues. We provide empirical evidence and theoretical justification under mild data assumptions. Exploiting these alignment properties, we derive a refined convergence bound that depends on the full spectrum and closely matches practical training dynamics, significantly improving over classical worst-case results. We further obtain improved generalization bounds. Experiments on MLPs and CNNs across multiple datasets validate our theory. --- ### AI Cartography: Mapping the Latent Landscape of AI Benchmark Ecosystems Authors: Michael Hardy, Anka Reuel, Lijin Zhang, Jodi M. Casabianca, Sang Truong, Yash Dave, Hansol Lee, Benjamin Domingue, Sanmi Koyejo Published: 2026-05-24 Categories: cs.AI, cs.CY, stat.AP arXiv: https://arxiv.org/abs/2605.25272v1 While aggregate leaderboard scores drive AI development, they contain substantial measurement noise whose sources and magnitudes remain unquantified, making it unclear when rankings reflect genuine capability differences versus evaluation artifacts. We introduce a framework for measuring the latent landscape in AI benchmark ecosystems. Applying Confirmatory Factor Analysis (CFA) and Generalizability Theory to 4,000+ models from the Open LLM Leaderboard, we decompose sources of ranking variance and establish: (1) structures assumed in current reporting practice underestimate the strength of relationships between benchmarks; (2) evidence of local dependence among leaderboard items, undermining uses of benchmarks as measurement instruments under current scoring systems; (3) contributor metadata explains more rank-relevant variance ($\approx9\%$) than architecture or deployment categories in this context; (4) a manifest-score "scaling law" slope has low reliability ($R_β=0.53$); by contrast, the latent general-factor size slope is highly stable across ecosystem controls ($R_g=0.97$). We are able to provide unique insights into benchmark dynamics, such as which benchmarks are a function of LLM size and which can be oppositely impacted by post-training practices. We provide actionable diagnostics to determine how benchmark rankings can be trusted and how benchmark design can be improved. --- ### Positivity in classical enumerative geometry: a case study in synchronized AI-assisted mathematics Authors: Gergely Bérczi, László M. Fehér Published: 2026-05-24 Categories: math.AG, cs.AI, cs.NE arXiv: https://arxiv.org/abs/2605.25271v1 We study the symmetric polynomial $\prod_{α\in A_{n,d}}\bigl(1+α_1 x_1+\cdots+α_n x_n\bigr)$ where $A_{n,d}:=\{α\in\mathbb{Z}_{\ge 0}^n:|α|=d\}$, which is the total Chern class of $\mathrm{Sym}^d(\mathbb{C}^n)$, viewed as a torus representation whose Chern roots are the weights $α_1 x_1+\cdots+α_n x_n$ for $α\in A_{n,d}$. Its homogeneous degree-$k$ part $c_k(n,d)$ is the $k$-th Chern class of $\mathrm{Sym}^d(\mathbb{C}^n)$. These Chern classes, together with their coefficients in various symmetric function bases, play a central role in enumerative geometry. Despite their simple definition, general closed formulas for their coefficients are subtle, and many structural properties of these classes have remained poorly understood. In this paper we prove several conjectures concerning their structure, establish explicit formulas, and study log-concavity properties for both the Chern classes and their $K$-theoretic analogue. In rank two, passing to the Schur basis and expanding the Schur coefficients in the binomial basis of $d$, we uncover a new binomial log-concavity phenomenon and prove refined positivity results. The paper demonstrates a novel methodology: we combine several AI systems with human mathematical insight in a coordinated workflow, deploying each tool according to its strengths in experimental discovery, conjecture formation, symbolic proof construction, and verification. To our knowledge, this is one of the first detailed case studies of orchestrating multiple AI tools to make substantial progress on a coherent mathematical research project. --- ### Latent Q-Barrier Shielding for Safe In-Context Reinforcement Learning Authors: Minjae Kwon, Amir Moeini, Shangtong Zhang, Lu Feng Published: 2026-05-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.25267v1 Safe in-context reinforcement learning (ICRL) adapts online from interaction history without test-time parameter updates while controlling episode cost under a safety budget. Under out-of-distribution (OOD) deployment shifts, pretraining-only safe ICRL can give poor reward-safety tradeoffs because the remaining budget affects behavior only through frozen policy conditioning, not an explicit action-level check against predicted future cost. We propose a latent Q-Barrier shield that learns a context representation, latent dynamics, and an ensemble cost critic before deployment. Without parameter updates, the shield infers context from history and filters or softly reweights candidate actions using the remaining budget and predicted future cost. We prove a conditional, error-decomposed barrier-margin result: a Q-Barrier-satisfying action leaves the next latent-budget state with an approximately budget-safe continuation under the learned critic, up to Bellman and latent-prediction errors. Across five safe ICRL benchmarks, the shield improves deployment-time reward-safety tradeoffs over a strong safe-ICRL baseline: after a short context window, it achieves higher return in four of five benchmarks while matching or lowering average episode cost in all five. --- ### DeltaCam: Differential Intrinsic Camera Modeling for Video Generation Authors: Debabrata Mandal, Zhihan Peng, Yujie Wang, Praneeth Chakravarthula Published: 2026-05-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.25266v1 Incorporating camera intrinsics into video generation models offers a principled way to control not only scene dynamics but also the imaging process that governs visual appearance. Prior work has primarily focused on extrinsic control, such as camera pose and motion, while treating intrinsic camera parameters as implicit or fixed. A key bottleneck is the lack of large-scale video datasets with accurate and diverse temporally varying camera metadata, which makes learning absolute camera parameterizations difficult. As a result, current models struggle to incorporate photographic camera behavior, including depth-of-field transitions, exposure variations, lens distortions, and color processing, in a controllable and temporally consistent manner. We introduce DeltaCam, a video diffusion framework that models camera behavior through $Δ$-parameterized neural camera adaptors, operating on relative changes in camera motion and intrinsics instead of absolute states. By learning this differential formulation from synthetic video data, we mitigate reliance on precise real-world camera labels and enable smooth, consistent control over imaging factors such as focal length, aperture, ISO, color temperature, and lens distortion. We extend this framework to real-world footage through two mechanisms: finetuning the controls on real image-metadata pairs for precise shot matching, and extracting disentangled embeddings for implicit video-to-video style transfer without requiring explicit camera parameters. By effectively separating scene content from intrinsic imaging behavior, DeltaCam enables camera-consistent video generation and editing operations that are difficult to achieve with existing models. Ultimately, our results establish a practical and scalable approach for bridging synthetic control and real-world photographic emulation. --- ### Mimir: Large-scale Multilingual Concept Modeling Authors: Elio Musacchio, Lucia Siciliani, Pierpaolo Basile Published: 2026-05-24 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2605.25263v2 Current language modeling approaches are built around tokens. Text corpora are split into tokens, and models are trained by performing computations on these tokens, such as predicting the next token given the preceding ones as context. This paradigm has become the standard in modern language modeling, especially given the outstanding performance obtained by token-based architectures. However, recent works have not only begun to question how language models process and understand meaning from tokens, but also to question whether using higher levels of granularity could advance the research field. This led to the idea of Concept Modeling, that is, to directly train models for next-concept prediction rather than next-token prediction. In this work, we introduce Mimir, a 1.6B Large Concept Model trained for multilingual concept understanding and generation. We leverage a large-scale multilingual pre-training corpus (38,883,987,240 sentences) spanning 46 languages and a large-scale multi-turn, and multilingual instruction-tuning dataset (66,816,428 sentences) covering a total of 35 languages. We extensively evaluate model performance against a language model with a comparable number of parameters. --- ### Mimir: Large-scale Multilingual Concept Modeling Authors: Elio Musacchio, Lucia Siciliani, Pierpaolo Basile Published: 2026-05-24 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2605.25263v1 Current language modeling approaches are built around tokens. Text corpora are split into tokens, and models are trained by performing computations on these tokens, such as predicting the next token given the preceding ones as context. This paradigm has become the standard in modern language modeling, especially given the outstanding performance obtained by token-based architectures. However, recent works have not only begun to question how language models process and understand meaning from tokens, but also to question whether using higher levels of granularity could advance the research field. This led to the idea of Concept Modeling, that is, to directly train models for next-concept prediction rather than next-token prediction. The goal is to change the input from tokens to concepts, forcing the underlying language model to shift its granularity from fine-grained tokens to broad concepts. In this work, we introduce Mimir, a 1.6B Large Concept Model trained for multilingual concept understanding and generation. We leverage a large-scale multilingual pre-training corpus (38,883,987,240 sentences) spanning 46 languages and a large-scale multi-turn and multilingual instruction-tuning dataset (66,816,428 sentences) covering a total of 35 languages. We extensively evaluate model performance against a language model with a comparable number of parameters. --- ### Semantics-Guided Multimodal Masked Autoencoder Pretraining for 3D BEV Object Detection Authors: Prabuddhi Wariyapperuma, Rajitha de Silva, Marc Hanheide, Thomas Bohné, Leonardo Guevara Published: 2026-05-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.25262v1 Accurate 3D bird's-eye view (BEV) object detection is essential for autonomous driving, and depends strongly on effective multimodal representations from complementary sensors such as cameras and LiDAR. Multimodal masked autoencoders have shown strong potential for learning such representations for downstream 3D BEV object detection. However, existing methods typically apply uniform random masking to camera and LiDAR inputs, treating all regions equally, and learn representations only through masked reconstruction. We propose a semantics-guided multimodal masked autoencoder framework that introduces semantic information during pretraining through two separate components: (i) semantics-guided LiDAR voxel masking, which preserves semantically important LiDAR regions more strongly, and (ii) an auxiliary point-wise LiDAR semantic decoder branch that injects semantic guidance in addition to reconstruction. On BEVFusion 3D object detection, our semantics-guided pretraining strategy improves performance on the nuScenes mini validation set compared to the standard UniM2AE baseline: semantics-guided LiDAR voxel masking yields +1.49% mean Average Precision (mAP) and +1.66% nuScenes Detection Score (NDS), while decoder-side point semantic supervision yields +1.39% mAP and +3.22% NDS over the baseline. --- ### First, do no harm: Breaking suicidogenic echo chambers in media recommendation Authors: Alberto Díaz-Álvarez, Raúl Lara-Cabrera, Fernando Ortega-Requena, Víctor Ramos-Osuna Published: 2026-05-24 Categories: cs.IR, cs.AI, cs.CY, cs.LG arXiv: https://arxiv.org/abs/2605.25258v1 Recommender systems generally optimises user engagement, but this approach is dangerous in mental health contexts. When vulnerable users show signs of suicidal ideation, standard algorithms often trap them in echo chambers of harmful content, worsening their psychological state. In response, we introduce RankAid, a re-ranking method that prioritises clinical safety alongside predictive relevance. It works as an add-on layer to existing models: it penalises risky items and boosts therapeutic content depending on the user's current level of vulnerability. We evaluated this approach using the MovieLens 1M dataset, where items were semantically annotated for clinical risk and therapeutic value using large language models. Our simulations show that our algorithm successfully blocks the recommendation of harmful content during crisis peaks, actively reshaping the feed to support emotional de-escalation. Furthermore, this safety intervention only causes a controlled, acceptable drop in standard accuracy metrics like NDCG. By using asymmetric hyperparameters, RankAid also gives system administrators the flexibility to tune the severity of the intervention based on specific clinical guidelines. --- ### Whose Alignment? Comparing LLM Process Alignment Across Diverse Organizational Decision Contexts Authors: Niklas Weller, Emilio Barkett Published: 2026-05-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.25256v2 Steerable pluralism requires a model to faithfully represent one specified perspective. Organizations are a natural setting for this demand, since they deploy LLMs to make decisions that must reflect their own policy. Yet, most existing work fixes that perspective at the level of individuals or demographic groups. We rely on a decision-policy capturing method to measure process alignment in organizational settings, assessing whether an LLM faithfully reproduces the organization's decision policy rather than merely reaching the same conclusions. We find heterogeneity along two axes. Across models, baseline alignment varies strongly and tracks neither pricing nor general benchmark performance. Across organizations, the structure of alignment changes. In ECHR Article 6 decisions, process alignment predicts output accuracy ($r = 0.85$, $p < .001$), and making the organization's past decision policy explicit improves poorly aligned models. In consumer credit decisions, process alignment is low overall but varies more than output accuracy, and the models resist adopting the organization's weighting of protected attributes. Because historical credit decisions encode potentially discriminatory patterns, higher alignment there is not always desirable. Process-level measurement is therefore necessary, and depending on whether the target policy is normatively desirable, the same procedure can calibrate or audit a model. Deciding which policy to align to, and whether higher alignment is feasible or desirable, makes organizational alignment a pluralistic problem in its own right. --- ### Boosted Stochastic Frank-Wolfe for Constrained Nonconvex Optimization Authors: Navil Nandhan, Abbas Khademi, Antonio Silveti-Falls Published: 2026-05-24 Categories: math.OC, stat.ML arXiv: https://arxiv.org/abs/2605.25255v1 The boosted Frank-Wolfe algorithm accelerates the classical Frank-Wolfe algorithm by better aligning the update direction with the negative gradient. Its analysis, however, has been limited to deterministic convex problems, with step sizes that require either line search or knowledge of the Lipschitz constant of the gradient. We develop a novel step size strategy that does not depend on the Lipschitz constant of the gradient, which allows us to extend the boosted Frank-Wolfe algorithm to the stochastic setting. We prove that boosting with this step size strategy can be combined with many modern gradient estimators, including SAGA, L-SVRG, SAG, Heavy Ball momentum, and zeroth-order estimators, among others, while retaining the worst-case convergence rates of ordinary stochastic Frank-Wolfe. Our analysis also yields the first convergence rates for boosted Frank-Wolfe on nonconvex and quasar-convex objectives, results which are new even for deterministic problems. Experiments on sparse logistic regression and quantum process tomography show that stochastic boosted Frank-Wolfe achieves faster convergence per gradient oracle call (and on wall-clock) compared to the non-boosted baseline. --- ### Guess the Unified Model: How Much Can We Recover from Generated Images? Authors: Jasin Cekinmez, Ryo Mitsuhashi, Addison J. Wu, Yida Yin Published: 2026-05-24 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2605.25254v1 With unified model-generated images now widespread online, attributing their model of origin offers a path toward transparency and deeper insight into the characteristic behaviors of individual models. Prior work has explored provenance in LLM-generated text, diffusion model images, and datasets, but the separability of unified model-generated images remains an underexplored area. We address this gap by examining separability across corruption, domains, and prompt languages using images generated by seven unified models. We show that model attribution is highly feasible as our model achieves near-perfect accuracy with around 20K images per model. Corruptions and structural perturbations have only a modest effect on attribution performance, and cross-domain generalization reveals that semantic content contributes to separability but is not the dominant signal. Finally, we observe that for most models, prompt language attribution is around chance levels, suggesting minimal language-specific visual signatures. These findings highlight consistent model-specific visual characteristics in unified models outputs and open new directions for tracing and auditing generative image pipelines. --- ### Quantifying Empirical Compute-Supervision Tradeoffs in RLVR Authors: Ryo Mitsuhashi, Patrick Chen, Isabelle Tseng, Jasin Cekinmez, Addison J. Wu Published: 2026-05-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.25252v2 Reinforcement learning with verifiable rewards (RLVR) has become a standard paradigm for post-training language models, but in practice, verifiers are rarely perfect. Recent theoretical work predicts that verifier noise affects the rate of learning but not its final outcome, implying that sufficient compute should close any gap induced by imperfect supervision. We test this prediction empirically by post-training Qwen2.5 (0.5B, 1.5B) with GRPO on GSM8K while injecting controlled false-positive and false-negative noise into the binary correctness signal, and varying rollouts per prompt as a compute axis. In practice, the gap in validation accuracy persists under substantial compute scaling, with returns to compute that are sharply diminishing. We further find a structural asymmetry where false negatives monotonically degrade performance more quickly than false positives. These findings suggest verifier quality and training compute are not interchangeable, and that reducing false negatives is a more effective lever than scaling compute alone. --- ### LipoAgent: Coordinating Fine-Tuned LLM Agents for Safer Lipid Design Authors: Leshu Li, An Lu, Haiyu Wang, Zhibin Feng, Conghui Duan, Qing Bao, Zongmin Zhao, Sai Qian Zhang Published: 2026-05-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.25250v1 Lipid nanoparticles (LNPs) are among the most clinically mature platforms for nucleic acid delivery, yet designing lipids that are both effective and biologically safe remains a major bottleneck. In practical screening, toxicity is a decision-level constraint: if a lipid is toxic, its efficiency prediction is clinically irrelevant. We propose LipoAgent, a safety-aware multi-agent LLM framework for lipid discovery. LipoAgent combines domain-specific finetuning with a conditional prediction objective that enforces toxicity as a prerequisite for efficiency prediction, and further improves reliability via multi-agent verification with lightweight human oversight when disagreement persists. Across multiple foundation models, LipoAgent achieves an average 32% relative improvement in mRNA transfection efficiency prediction compared with other reported models for lipid design. Wet-lab validation confirms that virtual screening rankings reliably translate to biological transfection outcomes. The code is publicly available at https://github.com/SAI-Lab-NYU/LipoAgent.git. --- ### FrontierOR: Benchmarking LLMs' Capacity for Efficient Algorithm Design in Large-Scale Optimization Authors: Minwei Kong, Chonghe Jiang, Ao Qu, Wenbin Ouyang, Zhaoming Zeng, Xiaotong Guo, Zhekai Li, Junyi Li, Yi Fan, Xinshou Zheng, Xi Jing, Yikai Zhang et al. Published: 2026-05-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.25246v3 Large language models (LLMs) are increasingly used for optimization modeling and solver-code generation, yet practical operations research and optimization problems often require a harder capability: designing scalable algorithms that exploit problem structure and outperform direct formulation-and-solve baselines. Existing benchmarks are limited to small or simplified examples far below real-world scale and complexity. We introduce FrontierOR, among the first benchmarks to systematically evaluate LLM-based efficient algorithm design for realistic large-scale optimization problems. FrontierOR includes 180 tasks derived from methodologically diverse papers published in top-tier operations research venues, each with standardized instances and a hidden, expert-verified evaluation suite. We evaluate seven LLMs spanning frontier, cost-effective, and open-source models both in one-shot and test-time evolution settings. The results reveal that frontier models still struggle to move from executable formulations to efficient optimization algorithms: the strongest one-shot model outperforms Gurobi in only 31% of cases in both solution quality and computational efficiency, and even strong coding agents with test-time evolution achieve only 50% on selected hard tasks. FrontierOR establishes a practical evaluation platform for LLM-based optimization algorithm design, which enables future LLMs and agents to be systematically tested on whether they can move beyond correct formulation toward a feasible, high-quality, and efficient algorithm. Code and data are publicly released at https://github.com/Minw913/FrontierOR. --- ### Inference Time Optimization with Confidence Dynamics Authors: Yu Wang, Minghao Liu, Jiayun Wang, Jinrui Huang, Ankit Shah, Wei Wei Published: 2026-05-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2605.25244v1 Inference time optimization techniques, such as repeated sampling, have significantly advanced the reasoning capabilities of Large Language Models (LLMs). However, the critical role of model uncertainty remains largely underexplored in these optimization strategies. In this paper, we investigate the dynamics of confidence along reasoning trajectories and for first time reveal a surprising and unique pattern: correct answer traces tend to exhibit confidence improvement over time (positive confidence gain), while incorrect traces show attenuated or declining confidence as reasoning proceeds. Based on this observation, we propose Confidence Dynamic Gain (CDG) based voting, which incorporates how the confidence trajectory of the response evolves along the reasoning chain. Experiments across four open-source architectures (DeepSeek-R1, gpt-oss, Gemma-3, Qwen-QwQ) on the AIME24/25, HMMT25, and BRUMO25 benchmarks demonstrate that CDG yields a significant performance boost over baselines. These results demonstrate that our method provides a robust discriminative signal for improving answer selection in LLM reasoning. We also provide theoretical insights for this phenomenon. Code will be released at https://github.com/Accenture/CDG.git. --- ### JudgmentBench: Comparing Rubric and Preference Evaluation for Quality Assessment Authors: Russell Yang, Ruishi Chen, Pierce Kelaita, Riya Ranjan, Sibo Ma, Charles Dickens, Matthew Guillod, Megan Ma, Julian Nyarko Published: 2026-05-24 Categories: cs.CL, cs.AI, cs.CY arXiv: https://arxiv.org/abs/2605.25240v3 Two methodologies dominate current practices of benchmarking: rubric-based scoring evaluates items against predefined criteria, whereas comparative judgment elicits pairwise preferences between outputs. Although both methodologies are widely used, the choice between them is rarely justified. We release JudgmentBench, a benchmark of 30 real-world legal tasks, paired with 1,539 rubric scores and 1,530 pairwise preference judgments collected from practicing attorneys--including at major U.S. law firms--with substantial experience. The annotations constitute the first publicly available dataset in a high-expertise domain in which both supervision signals are elicited from the same experts on the same items. Using LLM-generated outputs at three constructed quality levels, we provide an initial empirical comparison: comparative judgments recover the intended quality ordering substantially better than rubrics under both a per-task rank-correlation metric (mean Spearman's rank correlation of 0.908 vs. 0.150, estimated difference = 0.758 (0.494, 1.021)) and a per-judgment pairwise win-rate metric (0.669 vs. 0.542, estimated difference = 0.127 (0.067, 0.186)), while requiring less than half the annotation time. The patterns hold for human annotators and LLM autograders. Beyond this initial comparison, the paired structure of the dataset supports a broader research agenda on how expert judgment should be elicited, aggregated, and used as supervision in domains without verifiable ground truth. --- ### Constraint-Anchored Attribution: Feasibility-Certified Counterfactuals and Bonferroni-PAC Sufficient Subsets for Neural CO Policies Authors: Sohaib Lafifi Published: 2026-05-24 Categories: cs.LG, cs.AI, math.OC arXiv: https://arxiv.org/abs/2605.25235v1 We give an attribution method for neural combinatorial-optimisation (CO) policies that (i) decomposes a decision by constraint families via LP-relaxation duals, (ii) certifies counterfactuals through a combinatorial feasibility model (implemented as a CSP feasibility-decision model), and (iii) bounds the size of a PAC-sufficient explanation with a Bonferroni-corrected Hoeffding sufficient-subset test along a greedy ordering. Across three CO problems and three seeds, our LP-anchored $Λ$-attribution matches the CF-derived signal at 96.5% on CVRPTW (n_cert=344) and 77.2% on the Orienteering Problem (n_cert=281) vs 75.0% and 35.2% for proxy gradient (paired diffs +0.215 and +0.420; McNemar exact $p \le 10^{-14}$). In the rank-aligned regime of the Flexible Job-Shop Scheduling Problem, both backends agree on every CSP-certified flip (n_cert=59), confirming the no-gain prediction. Bonferroni-PAC subsets average 5.0 nodes per step ($M=70$, $\varepsilon=δ=0.2$, $k_{\max}=25$). Reference implementation: https://github.com/sohaibafifi/neuro-co-cax --- ### On the Epistemic Uncertainty of Overparametrized Neural Networks Authors: David Rügamer Published: 2026-05-24 Categories: cs.LG, cs.AI, stat.CO, stat.ML arXiv: https://arxiv.org/abs/2605.25234v1 Epistemic uncertainty is often viewed as a reducible uncertainty that vanishes with increasing data. This perspective implicitly assumes parameter identifiability and equates epistemic uncertainty with predictive variability. In overparametrized neural networks, however, model parameters are typically non-identifiable due to symmetries and redundant representations. As a consequence, substantial parameter uncertainty can persist even when the underlying function is fully identified. In this work, we analyze epistemic uncertainty through the lens of non-identifiability and characterize both discrete and continuous sources of residual uncertainty. Focusing on one-hidden-layer ReLU networks, we thoroughly analyze the resulting posterior structure and validate our theoretical insights through empirical studies. --- ### Meta-Agent: From Task Descriptions to Verified Multi-Agent Systems Authors: Andy Xu, Yu-Wing Tai Published: 2026-05-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.25233v1 AI agents are increasingly used to solve complex, multi-step tasks, but existing multi-agent frameworks remain brittle as workflows grow in scale and depth. Small errors at intermediate stages can propagate through agent interactions, while insufficient grounding and weak verification mechanisms further limit reliability. We present Meta-Agent, a two-phase framework that automatically constructs and executes specialized multi-agent systems from natural-language task descriptions. In the construction phase, a task planner decomposes a problem into a directed acyclic graph of agent specifications with explicit input/output contracts and verification criteria. A web search module grounds each specification with external evidence, and a code generation module produces system prompts and tool configurations. A construction-time verification stage then validates generated artifacts and triggers targeted regeneration when failures are detected. In the execution phase, a coordinator dispatches subtasks across the agent graph while execution-time verification gates intermediate outputs. We further introduce a three-level error attribution mechanism that distinguishes local, upstream, and structural failures, enabling targeted recovery strategies ranging from localized retries to partial re-execution and re-decomposition. We evaluate Meta-Agent across coding, contextual learning, and open-ended reasoning tasks. Experiments against strong multi-agent baselines and ablation studies demonstrate consistent improvements in task success rate, error recovery, and workflow stability. The results highlight the importance of tightly integrating planning, grounding, and verification for building reliable multi-agent systems. --- ### Specification-Based Code-Text-Code Reengineering for LLM-Mediated Software Evolution Authors: Oleg Grynets, Vasyl Lyashkevych, Arsen Dolichnyi, Roman Piznak, Taras Zelenyy, Volodymyr Morozov Published: 2026-05-24 Categories: cs.SE, cs.AI, cs.LO arXiv: https://arxiv.org/abs/2605.25232v1 Direct Code2Code transformation remains challenging to control because it can preserve surface-level syntax while introducing semantic drift, hidden behavioral changes, loss of traceability, non-idiomatic target implementations, or incomplete reconstruction of domain logic. This paper proposes a specification-based Code2Text2Code reengineering framework for LLM-mediated software evolution. The central idea is to transform source code into a neutral textual specification that captures program behavior, identifiers, computational flow, conditions, side effects, data dependencies, and domain-specific intent without directly transferring the source language syntax. The proposed framework combines factual context extraction, Code2Text generation, iterative verification between source code and text specification, Text2Code generation, target code verification, retrieval-augmented grounding, and semantic-aware chunking, and transformation loss estimation. The knowledge representation layer integrates metadata derived from AST, graph-based dependency structures, neutral natural language specifications, technical documentation, business documentation, and architecture-level representations. The conducted experiments include a Code2Text2Code dataset built from multiple programming languages and SQL dialects, comparison of intermediate representations, retrieval evaluation, documentation transformation evaluation, and prompt tuning using DSPy. A graph formalization using structural preservation, reverse compatibility, interface stability, and total graph similarity is implemented to estimate transformation losses. The results support the interpretation of the Code2Text2Code approach not as a simple code transformation, but as a controlled specification-based reengineering process for LLM-mediated software evolution. --- ### Boosting Inference with Guided Reasoning: Stochastic Exploration for Recursive Models Authors: Andrew Corbett, Archit Sood, Anna Tzatzopoulou, Sai-Aakash Ramesh, Tim Dodwell Published: 2026-05-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.25230v2 Recent work on recursive architectures has shown that tiny neural networks can be surprisingly powerful on structured reasoning tasks. The trick is to model reasoning trajectories with a latent dynamical system. We argue that the inference-time behaviour of these architectures is best understood as approximate inference over latent reasoning trajectories, with deterministic recursion as the one-particle, zero-noise limit. We make this view operational through guided stochastic exploration: stochastic perturbations of the reasoning dynamics propose neighbouring trajectories, and the model's existing early-stopping head reweights them online. The framework yields three label-free diagnostics: local stability, guide alignment, and cloud-token entropy. These predict, from inference traces alone, whether the procedure will help and which of its outputs to trust. On Sudoku-Extreme it lifts exact-solve accuracy from $85.9\%$ to $98.0\%$ without retraining; on Maze-Hard the diagnostics flag a misaligned guide, as validation performance later confirms. The same machinery thus characterises both when recursive reasoning has room to improve at the trajectory level and when the model's internal guide can recover it. --- ### A Blended Likelihood Approach for Achieving Fairness Using Naive Bayes Authors: John Arthur Junior, Abdul Lateef Yussif, Maame G. Asante-Mensah, Charles R. Haruna, Sandro Amofa, Elliot Attipoe Published: 2026-05-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.25228v1 Concerns about algorithmic bias and fairness have increased as artificial intelligence has been incorporated into high-stakes decision-making. Traditional Naive Bayes classifiers, while efficient and interpretable, lack fairness-awareness mechanisms and perpetuate historical biases in sensitive domains such as hiring, credit scoring, and criminal justice. This study develops a fairness-aware extension of the Naive Bayes classifier that mitigates bias while maintaining computational efficiency. We propose the Bias Mitigating Naive Bayes (BMNB) classifier, integrating in-processing and post-processing interventions. The in-processing stage employs a blended likelihood approach combining group-specific and pooled likelihood estimates through a tunable blending parameter alpha to balance fairness and accuracy. The post-processing stage applies output calibration with adaptive thresholding to fine-tune group-specific decision boundaries. Experimental results indicate that BMNB attains Disparate Impact (DI) values of 1.000, 1.171, and 0.997 and Equal Opportunity Difference (EOD) values of -0.217, -0.226, and -0.053 on the Adult, ProPublica, and Framingham datasets, respectively, while maintaining computational efficiency. Ablation studies confirm that the combination of blended likelihood and adaptive thresholding yields superior performance compared to either technique in isolation. --- ### From Automation to Collaboration: Human-in-the-Loop Methods for Safe and Trustworthy NLP Authors: Most. Sharmin Sultana Samu, MD. Tanvir Ahmed Seum, Md. Rakibul Islam Published: 2026-05-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2605.25226v1 Large language models are widely deployed in high-stakes NLP tasks, yet risks such as bias, hallucination, adversarial vulnerability and unreliable generalization remain. Probe-based auditing reveals inconsistencies in model behavior. Adversarial text generation uncovers robustness gaps, especially in lower-resourced languages with limited benchmarks. Enterprise text-to-SQL settings expose the difficulty of validating outputs over private and large-scale databases. Human supervision is essential for probe validation, adversarial verification and domain-specific annotation, but it is costly and hard to scale. This survey examines recent human-in-the-loop methods that shift NLP from automation toward collaboration for safety and trustworthiness. We review how human expertise supports auditing, robustness evaluation, data construction and model steering. Our findings highlight gaps in scalable probing, sustainable robustness benchmarks, low-resource settings and governance of private systems. We outline practical research directions for adaptive auditing, collaborative evaluation and accountable deployment. --- ### Transformer Field Theory: A Response-Theoretic Approach to Mechanistic Interpretability Authors: David N. Olivieri, Antonio F. Pérez Rodríguez Published: 2026-05-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.25225v2 Mechanistic interpretability often studies Transformer behavior by intervening on internal activations through activation patching, causal tracing, path patching, and steering directions. This paper develops Transformer Field Theory: a response-theoretic framework in which the residual stream of a fixed forward pass is treated as a Transformer field over layer depth and token position. In this formulation, patching becomes a localized source insertion into the Transformer field, first-order sensitivity fields predict patch effects, Green functions describe downstream propagation, and patch selection is posed as an adjoint inverse problem. Empirically, we test the theory's forward response objects in GPT-2-style autoregressive Transformers. Localized Transformer-field interventions exhibit a bounded local linear regime; first-order sensitivities predict patch effects across layer-token sites; localized sources generate structured anisotropic Transformer-field propagation; high-sensitivity sites and sliced Green operators provide reduced response descriptions; and prompt-induced Transformer-field displacements partially transfer answer behavior. These results establish sensitivities, Transformer-field responses, and sliced Green operators as practical objects for organizing patching experiments, while providing the forward mathematical basis for patch-site inference and cross-scale response transfer. --- ### Data-Specific Hyper-Parameter Design: A Paradigm Shift in Reservoir Computing Authors: G Manjunath, Juan-Pablo Ortega, Alma van der Merwe Published: 2026-05-24 Categories: math.DS, cs.LG arXiv: https://arxiv.org/abs/2605.25221v1 Reservoir computing typically relies on large, randomly generated reservoirs, enabling simple, often linear readouts. Over the past two decades, most constructions have exploited the freedom to select the reservoir, constrained primarily by stability conditions based on state contraction or memory capacity. However, these designs are largely independent of the input data and learning objective, resulting in a trial-and-error methodology driven by randomness. In high dimensions, the reservoir acts as a random embedding of the input history, implicitly relying on Johnson--Lindenstrauss--type concentration phenomena to preserve information. In contrast, we develop reservoir design principles from a geometric perspective for inputs generated by deterministic dynamical systems. Rather than relying on random embeddings, we require reservoir state increments to align within a cone around an input-determined vector subspace, and prove that such a cone concentration reduces ridge-regression training error. When the cone angle is small, the variance of reservoir states concentrates in the input-determined subspace, improving conditioning of the empirical second-moment matrix and strengthening alignment between dominant covariance directions and the state-target cross-covariance. For echo state networks, we provide a constructive approach to reservoir design. The reservoir matrix is chosen so that associated Krylov-chain directions remain nearly closed within an input-determined subspace while permitting controlled mixing in its orthogonal complement. We also provide a spectral diagnostic for ridge regression training that identifies when reservoir geometry concentrates predictive information into a few dominant covariance modes and when ``spectral pollution'' inhibits forecasting. Numerical experiments demonstrate consistent performance gains over arbitrary reservoir constructions. --- ### Multi-view Consistent 3D Gaussian Head Avatars 'without' Multi-view Generation Authors: Aviral Chharia, Fernando De la Torre Published: 2026-05-24 Categories: cs.CV, cs.GR, cs.RO arXiv: https://arxiv.org/abs/2605.25220v1 High-fidelity 3D Gaussian head avatar generation is critical for applications such as AR/VR, telepresence, and digital humans. Existing methods depend on multi-view datasets, 3D captures, or intermediate 2D view synthesis. In contrast, we learn both conditional and unconditional 3D head models from randomly sampled 2D images alone, without using multi-view data, 3D supervision, or intermediate view generation. We introduce MVCHead, a single-shot state space model that enforces multi-view consistency (MVC) directly in the 3D representation while regressing 3D Gaussians under these constraints. At its core, we propose a Hierarchical State Space (HiSS) block that progressively refines Gaussians from coarse to fine, while capturing long-range dependencies. Within each HiSS block, we modify Mamba's standard unidirectional scan with the proposed Hierarchical Bi-directional State Scan (HiBiSS) that aligns recurrence with the axes along which multi-view inconsistencies are strongest. Finally, we design an SE(3) Multi-view Critic that judges whether a set of self-renders arises from a single underlying 3D configuration, rewarding cross-view pixel alignment without observing real multi-view pairs. MVCHead achieves state-of-the-art perceptual quality, surpasses prior methods in both texture and geometric consistency, and maintains comparable shape consistency. To demonstrate scalability, we release FaceGS-10K, the first large-scale dataset of ready-to-use 3D Gaussian head assets for training and evaluation of 3D head models. Project Page and code: https://humansensinglab.github.io/MVCHead/ --- ### Personalized Federated Learning by Energy-Efficient UAV Communications Authors: Shiqian Guo, Jianqing Liu, Beatriz Lorenzo Published: 2026-05-24 Categories: cs.LG, eess.SY arXiv: https://arxiv.org/abs/2605.25212v1 Federated learning (FL) is an effective paradigm for enhancing the learning capability of edge devices while preserving data privacy. In geographically dispersed FL systems, such as sensor networks in remote areas, unmanned aerial vehicles (UAVs) can flexibly establish high-quality communication links to support parameter exchange. However, device heterogeneity and the limited battery capacity of UAVs pose significant challenges. Specifically, data heterogeneity slows convergence, while scheduling all devices for global collaboration incurs excessive communication and energy costs. To overcome these challenges, we adopt a strict separation between a globally shared backbone and permanently local personalization heads, thereby mitigating the impact of data heterogeneity. Furthermore, we propose a gradient-based scheduling strategy that jointly considers energy efficiency and learning performance. In each communication round, the backbone is updated only by the top-$α$ devices ranked by gradient $\ell_{2}$-norm, ensuring that optimization focuses on the most informative updates. Simulation results demonstrate that the proposed scheme achieves higher learning accuracy than state-of-the-art approaches while significantly reducing UAV energy consumption. --- ### Evolving Causal Regulatory Networks (ECR-Net) Authors: Govind Vallabhasseri Binish, Abdhul Ahadh, Rano Roy Kavanal, Arya Ukunde Published: 2026-05-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.25211v1 Modern machine learning models excel at pattern recognition but remain brittle, often failing to generalize out of distribution (OOD) because they capture spurious correlations rather than the underlying causal data-generating process. Current causal discovery methods, while powerful, typically assume a static graph structure, rendering them unable to model systems that adapt or undergo structural changes across different environments. We introduce ECR-Net, Evolving Causal Regulatory Networks, a novel, bio-inspired framework for adaptive causal mechanism discovery. Our approach models the data-generating process not as a static graph, but as a dynamic system analogous to a Gene Regulatory Network (GRN), composed of localized, recursive functions where variables can activate and inhibit one another. To discover the latent structure of this network, we employ an evolutionary search algorithm that evolves a population of candidate regulatory graphs, optimizing for a fitness function that measures how well the simulated system dynamics reconstruct the observed data. The key innovation of ECR-Net is its ability to model structural adaptation, it explicitly ingests shifts in the data's statistical properties as signals of an environmental shock. In response, the evolutionary search identifies parsimonious modifications to the causal graph topology, such as link inhibitions or activations that explain the new data regime. We posit that ECR-Net represents a new class of adaptive Structural Causal Models capable of discovering how and why a system's fundamental rules change, offering a path toward robust generalization in complex, non-stationary systems. --- ### Multi-Objective Learning for Diffusion Models: A Statistical Theory under Semi-Supervised Learning Authors: Ziheng Cheng, Yixiao Huang, Hanlin Zhu, Haoran Geng, Somayeh Sojoudi, Jitendra Malik, Pieter Abbeel, Xin Guo Published: 2026-05-24 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2605.25210v1 Diffusion models are increasingly used as powerful conditional generators, yet real deployments often involve multiple target distributions arising from different tasks, e.g., diverse prompt domains in text-to-image generation, or multiple environments in robotics with diffusion policies. This naturally leads to a multi-objective learning (MOL) problem. A key challenge is that achieving good Pareto trade-offs can require a generalist model class with substantially larger capacity than what suffices for solving any individual task, thereby increasing statistical cost since sample complexity typically scales with the model complexity. To reconcile this, we develop a principled MOL framework for diffusion models with limited data: a semi-supervised regime where paired (labeled) samples are scarce, but (unlabeled) condition data are abundant. We propose a two-stage training procedure that first fits lightweight specialist models from limited paired data, and then distills them into a generalist model by generating pseudo-samples. We establish generalization bounds showing that the required number of paired samples only depends on the complexity of the specialist model classes. We further extend the theory to diffusion policies for sequential decision making to account for distribution shift in on-policy rollouts. Extensive experiments on robotic control and image restoration tasks are conducted to verify our theoretical results. --- ### They Are Not the Same: Direct Causes Are Not Grounded Emotion Explanations Authors: Zhuangzhuang Pan, Yan Xia, Chee Seng Chan Published: 2026-05-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2605.25208v1 Emotion-Cause Pair Extraction (ECPE) was introduced to explain why an emotion occurs, but this goal is now often reduced to binary pair/non-pair prediction. This proxy is useful for direct-cause extraction, yet easy to over-read as evidence grounded emotion explanation. We show that this interpretation is only partially valid. In IEMO-MECP, 90.9% of original positives remain emo-cause and 95.0% of original negatives remain non-pair, confirming that the binary ECPE task is largely preserved. The problem is that direct triggers alone do not constitute a grounded explanation. Emo-context, an utterance that helps interpret a target emotion without directly causing it, appears on both sides of the original boundary and is enriched near binary uncertainty, showing that the binary boundary has no stable place for such discourse evidence. Across evaluated ECPE models, direct triggers are recovered more reliably than contextual support. Under shortcut pressure, this imbalance becomes consequential. Binary-trained models assign higher pair scores to nearby lexically similar non-pair candidates than to evidence supported but structurally harder emo-cause and emo-context pairs. Thus, pair scores can reward convenient attributions over grounded explanations. High binary ECPE performance indicates that a model can identify direct triggers; it does not indicate that the model has explained the emotion. Code is publicly available at https://github.com/panzhzh/ECPExsame. --- ### Clarification Is Not Enough: Post-Clarification Answering Remains the Bottleneck in Multi-Turn QA Authors: Jinyan Su, Jennifer Healey Published: 2026-05-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2605.25204v1 Pluralistic alignment requires systems to adapt to diverse user values, communication styles, and contextual assumptions. We believe that a foundational prerequisite for such alignment enabling accurate preference elicitation from people when their intent is under-specified or ambiguous. We study the problem of preference elicitation in multi-turn question answering by decomposing the problem into two components: a \textbf{clarification policy}, which decides whether to ask a clarifying question or answer directly, and \textbf{post-clarification answering}, which produces the correct final answer once the missing information is provided. We show, using the PACIFIC benchmark, that supervised fine-tuning rapidly improves the clarification policy, however, final answer accuracy remains substantially lower even when the model takes the correct action. This gap indicates that understanding and correctly interpreting the user's response is the critical gap in multi-turn question-answering systems. --- ### Influence-Inspired Spectral Rotations for Extreme Low-Bit LLM Quantization Authors: Gorgi Pavlov Published: 2026-05-24 Categories: cs.LG, cs.AI, cs.LO arXiv: https://arxiv.org/abs/2605.25203v1 We apply the influence-adaptive Walsh geometry of a companion theory paper (arXiv:2605.01637) to extreme low-bit weight-only LLM quantization. The recipe is one math-invariant transformation: WHT-rotate each linear layer's weight matrix and rescale its columns by per-coordinate Walsh-basis activation energy before handing off to a reconstruction-error quantizer (Intel auto-round). This biases per-group integer rounding toward high-spectral-energy channels. On four pretrained decoder-only models from 135M to 1.5B parameters, BBT-spectral reduces wikitext-2 perplexity by 15-58% relative to vanilla auto-round at W2A16; we also report a TinyLlama-1.1B auxiliary data point. Three extensions transfer the recipe to families it failed on: a per-head PCA matrix-Gamma replacement of q_norm/k_norm for Qwen3 attention (PPL 136.76 -> 88.99 on Qwen3-0.6B); an SO(2) per-pair rotation that commutes with RoPE (PPL 36.93 -> 21.84 on Qwen2.5-1.5B); and an MoE-aware input-side absorption fix identified by architectural fuzzing of Laguna-style fused-expert layouts. A W2-vs-W4 ablation gives a deliberate negative control: the redistribution payoff falls within the +/-0.5 PPL noise floor at W4, consistent with the Schur-convexity intuition that the cost of unconcentrated influence vanishes as the noise budget shrinks. All quantized weights export to OpenVINO IR and run on Intel NPU + Arc dGPU + CPU with PPL invariant to device within +/-0.1. We do not claim a formal Boolean-to-real-valued transfer of the theory paper's majorization argument: the WHT activation energy used here is not the Boolean influence of the theory paper, the link is intuitive, and the contribution is engineering value rather than a transferred theorem. Head-to-head benchmarks against SpinQuant, QuaRot, QuIP-sharp, AQLM, OmniQuant, and ButterflyQuant at matched calibration are the main future-work item. --- ### GroupTravelBench: Benchmarking LLM Agents on Multi-Person Travel Planning Authors: Xiang Cheng, Yulan Hu, Lulu Zheng, Xiangwen Zhang, Zheng Pan, Xin Li, Yong Liu Published: 2026-05-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2605.25200v3 Travel planning in the real world is overwhelmingly a \textit{group} activity, yet existing LLM travel-planning benchmarks reduce it to a single user, where the field is approaching saturation. This single-user assumption sidesteps what makes group planning hard for an agent: discovering private preferences across multiple users, surfacing conflicts, and balancing utility against fairness. To bring the task back to its multi-user reality, we introduce \textbf{\textit{GroupTravelBench}}, the first benchmark for \textbf{multi-user, multi-turn} travel planning. Built from real user profiles, POI data, and ticket prices, it comprises 650 tasks across three difficulty levels, each running in a synchronous group-chat sandbox with cached tool data for reproducible offline evaluation. Beyond the multi-step reasoning and tool use that single-user benchmarks already test, GroupTravelBench probes three group-specific capabilities: \textit{(i) elicitation} of private preferences through multi-turn dialogue; \textit{(ii) coordination} of inter-user conflicts via compromise or subgrouping; and \textit{(iii) planning} that balances group utility against fairness. We pair this with a complementary evaluation framework combining rule-based outcome metrics and LLM-judge process metrics. Across a wide range of frontier models, even the strongest agents fall short on all four rule-based outcome metrics, with plan validity below 12\%, suggesting that group-level outcome quality is a key open challenge for LLM travel-planning agents. --- ### GroupTravelBench: Benchmarking LLM Agents on Multi-Person Travel Planning Authors: Xiang Cheng, Yulan Hu, Lulu Zheng, Zheng Pan, Xin Li, Yong Liu Published: 2026-05-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2605.25200v2 Travel planning in the real world is overwhelmingly a \textit{group} activity, yet existing LLM travel-planning benchmarks reduce it to a single user, where the field is approaching saturation. This single-user assumption sidesteps what makes group planning hard for an agent: discovering private preferences across multiple users, surfacing conflicts, and balancing utility against fairness. To bring the task back to its multi-user reality, we introduce \textbf{\textit{GroupTravelBench}}, the first benchmark for \textbf{multi-user, multi-turn} travel planning. Built from real user profiles, POI data, and ticket prices, it comprises 650 tasks across three difficulty levels, each running in a synchronous group-chat sandbox with cached tool data for reproducible offline evaluation. Beyond the multi-step reasoning and tool use that single-user benchmarks already test, GroupTravelBench probes three group-specific capabilities: \textit{(i) elicitation} of private preferences through multi-turn dialogue; \textit{(ii) coordination} of inter-user conflicts via compromise or subgrouping; and \textit{(iii) planning} that balances group utility against fairness. We pair this with a complementary evaluation framework combining rule-based outcome metrics and LLM-judge process metrics. Across a wide range of frontier models, even the strongest agents fall short on all four rule-based outcome metrics, with plan validity below 12\%, suggesting that group-level outcome quality is a key open challenge for LLM travel-planning agents. --- ### Hide to Guide: Learning via Semantic Masking Authors: Ruitao Liu, Qinghao Hu, Alex Hu, Yecheng Wu, Shang Yang, Luke J. Huang, Zhuoyang Zhang, Han Cai, Song Han Published: 2026-05-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.25198v1 Reinforcement learning with verifiable rewards (RLVR) has become a powerful paradigm for improving language models on reasoning-intensive tasks, but its effectiveness is often limited by exploration. For example, models often fail on hard problems, leaving little useful reward signal. External expert traces offer a natural source of guidance, yet they may also expose reward-relevant content along the critical path to the verifier target, such as final answers, intermediate values, executable implementations, or answer-related entities. This content can create an unintended reward hacking channel, allowing the policy to obtain reward by copying the trace rather than learning the underlying reasoning or agentic behavior. Existing guided-RL methods reduce this risk by using partial trajectories, but they mainly control how much expert information is shown heuristically rather than which parts should be hidden. To this end, we propose Semantic Masked Expert Policy Optimization (SMEPO), a fine-grained semantic masking strategy for expert-guided RLVR. Instead of truncating traces coarsely or revealing them unchanged, SMEPO masks reward-relevant semantic spans along the critical path while preserving the expert's decomposition, plan, and procedural structure. This turns hard problems from reasoning from scratch into a fill-in-the-blank process: the policy can follow the expert's problem-solving route, but must still reconstruct the missing values, code, or entities by itself. SMEPO is simple to apply and requires no changes to the reward function or RL objective. Across diverse domains, including math, code, and agentic search, SMEPO improves accuracy by up to 3.2 points over GRPO and reduces training time by up to 4.2x. The code is available at https://github.com/mit-han-lab/SMEPO. --- ### Beyond Killer Robots: General AI Attitudes and Public Support for Military AI in Nine Countries Authors: Andreas Jungherr, Antonia Schlude, Adrian Rauchfleisch Published: 2026-05-24 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2605.25196v1 AI-enabled military systems are a fixture of modern military conflict. Applications vary from autonomous drones for surveillance and attack to AI-supported target selection. The importance of AI for modern conflict shows also in public disputes between governments and technology companies over the conditions for military access to frontier AI. Both military uses and government attempts at enabling and steering them happen before a backdrop of public opinion, yet we still know little about how people think about military AI. Drawing on a preregistered survey of 9,000 respondents in nine countries, including China, Germany, and the United States, we examine whether support for military AI is shaped primarily by general attitudes toward AI, principled opposition to lethal autonomy, or foreign-policy and geopolitical orientations. Across six military AI scenarios that vary in lethality and human control, respondents who view AI as beneficial are substantially more supportive of military AI. Hawkish respondents are also more supportive. By contrast, principled opposition to lethal autonomy is not broadly associated with the full index but is related to the application of fully autonomous lethal force. Contrary to our expectation, perceived AI risks are positively associated with support. Cross-national differences are moderate and broadly consistent with geopolitical context. Overall, public opinion toward military AI appears conditionally permissive. Publics are not categorically opposed to various military uses of AI. Instead, unease is concentrated around fully autonomous lethal force. --- ### Baton: Explicit Semantic Blueprints for Joint Video-Audio Generation Authors: Shuyuan Tu, Qi Tian, Zihan Yang, Yue Wu, Xintong Han, Weijie Kong, Jiangfeng Xiong, Jian-Wei Zhang, Zhao Zhong, Liefeng Bo, Zuxuan Wu, Yu-Gang Jiang Published: 2026-05-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.25195v2 Current open-source diffusion models struggle to generate stable and synchronized audio-visual content, particularly in scenarios demanding complex semantic reasoning. The root cause is that existing methods rely on coarse text embeddings from off-the-shelf encoders to guide audio-video denoising, which discards fine-grained semantics and, critically, lacks a shared long-horizon plan, leading to uncoordinated denoising trajectories and fragile cross-modal alignment. We propose Baton, the first framework that introduces explicit semantic planning into joint video-audio generation. Our key insight is that complementing coarse text guidance with semantically rich, modality-aware planned tokens, jointly reasoned and mutually aligned before denoising, can simultaneously restore fine-grained semantic detail and establish a shared blueprint that coordinates both audio and video denoising trajectories. Concretely, Baton first introduces the VA-Planner, a multimodal language model equipped with dual semantic alignment towers, where learnable queries cross-attend to both video and audio features to produce a pair of semantically aligned video and audio planned tokens as keyframe-level blueprints. These planned tokens are injected into the diffusion backbone via cross-attention layers, providing temporally grounded guidance complementary to coarse text embeddings. Since planned tokens do not share one-to-one spatial-temporal correspondence with diffusion latents, we further propose Relative Semantic RoPE, a relative positional encoding that maps planned tokens and latents into a shared spatial-temporal coordinate frame, enabling each latent to accurately attend to its positionally corresponding semantic cues. Experiments on benchmarks show the effectiveness of Baton both qualitatively and quantitatively. --- ### Localize and Neutralize: Gradient-guided Token Suppression against Visual Prompt Injection Attack Authors: Dongpeng Zhang, Ke Ma, Yangbangyan Jiang, Gaozheng Pei, Longtao Huang, Qianqian Xu, Qingming Huang Published: 2026-05-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.25194v3 Adversarial images pose a severe security threat to multimodal large language models through prompt injection. Existing defenses largely lack a principled understanding of the underlying mechanisms and struggle to balance efficiency and defense utility. In this work, we show that successful adversarial attacks do not rely on the entire image uniformly but instead depend on a small subset of critical image tokens. Based on this insight, we propose Gradient Token Masking (GTM), which localizes these tokens via gradient analysis and neutralizes them through masking. We find that attribution based on the first generated token's output probability fails when attacks preserve the predicted token. To overcome this, GTM utilizes the Hidden-State Gradient Norm score for generation-influence attribution under adversarial inputs. We prove that its ranking is consistent with that of the full adversarial loss gradient, providing a theoretical guarantee for accurate localization. Our method requires only a single forward-backward pass to identify and zero out a small number of high-scoring tokens, effectively disrupting the adversarial attack path. Extensive experiments on prompt injection and multimodal jailbreak attacks demonstrate that our approach reduces attack success rates (ASR) to near zero while preserving model utility with negligible computational overhead. --- ### Localization then Neutralization: Gradient-guided Token Suppression against Visual Prompt Injection Attack Authors: Dongpeng Zhang, Ke Ma, Yangbangyan Jiang, Gaozheng Pei, Longtao Huang, Qianqian Xu, Qingming Huang Published: 2026-05-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.25194v1 Adversarial images pose a severe security threat to multimodal large language models through prompt injection. Existing defenses largely lack a principled understanding of the underlying mechanisms and struggle to balance efficiency and defense utility. In this work, we show that successful adversarial attacks do not rely on the entire image uniformly but instead depend on a small subset of critical image tokens. Based on this insight, we propose Gradient Token Masking (GTM), which localizes these tokens via gradient analysis and neutralizes them through masking. We find that attribution based on the first generated token's output probability fails when attacks preserve the predicted token. To overcome this, GTM utilizes the Hidden-State Gradient Norm score for generation-influence attribution under adversarial inputs. We prove that its ranking is consistent with that of the full adversarial loss gradient, providing a theoretical guarantee for accurate localization. Our method requires only a single forward-backward pass to identify and zero out a small number of high-scoring tokens, effectively disrupting the adversarial attack path. Extensive experiments on prompt injection and multimodal jailbreak attacks demonstrate that our approach reduces attack success rates (ASR) to near zero while preserving model utility with negligible computational overhead. --- ### SpongeBob: Sync-Aware Harmonious Audio-Visual Generative Editing Authors: Sen Liang, Cong Wang, Fengbin Guan, Zhentao Yu, Yiting Lu, Yuanzhi Wang, Yuan Zhou, Xin Li, Zhibo Chen Published: 2026-05-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.25193v2 Visual and acoustic events in the physical world are inherently coupled, yet existing video editing methods typically adopt decoupled pipelines, lacking bidirectional modality interaction. This results in two key limitations: (i) audio-visual desynchronization and (ii) contextual conflicts between generated audio and preserved content. To address these, we propose SpongeBob, the first end-to-end audio-visual joint editing framework featuring bidirectional cross-modal interaction. For synchronization, a Sync-Aware Mechanism aligns visual edits with sound events via bidirectional attention, temporal alignment, and spatial constraints. For contextual consistency, a Context-Aware Module leverages acoustic and visual context attention to prevent semantic clashes. Additionally, we introduce Sync-Preserving Training and Guidance (SPTG) to enhance alignment without degrading quality. Due to the scarcity of paired data, we construct a scalable data pipeline and a large-scale subject-level dataset. We also propose SpongeBob-Bench for systematic evaluation. Experiments show SpongeBob significantly outperforms existing baselines, improving Sync-C by 30% and Ctx-F1 by 12.5%. Our project page is available at: https://hy-spongebob.github.io/. --- ### Injecting Image Guidance into Text-Conditioned Diffusion Models at Inference Authors: Agata Żywot, Iason Skylitsis, Thijmen Nijdam, Zoe Tzifa-Kratira, Derck Prinzhorn, Konrad Szewczyk, Aritra Bhowmik Published: 2026-05-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.25191v1 Text-to-image diffusion models like Stable Diffusion generate high-quality images from text, but lack a way to inject visual guidance (e.g. sketches, styles) at inference without retraining. Existing methods either require computationally expensive fine-tuning or rely on style transfer techniques that risk semantic misalignment with textual prompts. We introduce Visual Concept Fusion (VCF), the first method offering dual conditioning on both an image and text prompt at inference time without any concept-specific training. VCF enables visual concept injection into Stable Diffusion by aligning CLIP image features with the text embedding space. VCF consists of three components: (1) a lightweight aligner that maps image tokens to the text embedding manifold using InfoNCE and cross-attention reconstruction losses, (2) a fusion strategy that preserves both textual and visual semantics, and (3) an optional Prompt-Noise Optimization (PNO) module for test-time refinement. Our experiments demonstrate that VCF successfully transfers visual attributes including style, composition, and color palette from reference images while maintaining prompt adherence. Quantitative results show a trade-off between text alignment (CLIP score) and visual correspondence (LPIPS), with VCF outperforming baselines in reference fidelity. --- ### Directional Alignment Mitigates Reward Hacking in Reinforcement Learning for Language Models Authors: Wenlong Deng, Jiaji Huang, Kaan Ozkara, Yushu Li, Christos Thrampoulidis, Xiaoxiao Li, Youngsuk Park Published: 2026-05-24 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2605.25189v1 Reward hacking arises when a model improves a proxy reward by exploiting shortcuts rather than solving the intended task. We study this failure mode through the geometry of reinforcement learning updates in language models and argue that hacking emerges when optimization drifts away from a stable low-dimensional learning trajectory. We analyze this drift through dominant singular directions of parameter updates and show that reward-hacking runs exhibit substantially larger directional change than clean runs. Motivated by this observation, we introduce trusted-direction projection, which constrains gradients to remain within a clean reference subspace. Across reward-hacking experiments on mathematical reasoning, the proposed approach delays shortcut exploitation and better preserves task performance. --- ### DarkForest: Less Talk, Higher Accuracy for Multi-Agent LLMs Authors: Yi Li, Songtao Wei, Dongming Jiang, Zhichun Guo, Qiannan Li, Bingzhe Li Published: 2026-05-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.25188v1 Multi-agent LLM systems improve reasoning by combining outputs from multiple agents, but interaction-heavy methods can introduce error propagation and high communication overhead. When agents exchange raw responses or reasoning traces, incorrect intermediate reasoning may be adopted and amplified, leading to confident but wrong consensus; multi-round communication also increases token consumption, latency, and inference cost. In this paper, we propose a controlled-communication coordination framework named DarkForest. DarkForest first keeps agents independent, so each agent produces an answer without seeing the others' outputs. It then parses the raw responses into structured candidate records, groups semantically equivalent candidates into clusters, and estimates a calibrated belief distribution over these clusters using agent reliability, confidence, parse quality, support-pattern reliability, and independence corrections. A coordinator receives only policy-permitted evidence from this belief state with controlled communication. Experiments on six reasoning benchmarks show that DarkForest achieves leading overall quality, improves the strongest baseline by up to 30.7\% on benchmark metrics, and reduces token consumption by up to $6.5\times$ compared with communication-heavy baselines. --- ### By Their Fruits You Will Know Them: Comparing Formalizations of Law by the Decisions They Encode Authors: Julius Vernie, Matthias Grabmair Published: 2026-05-24 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2605.25186v2 Formalizing legal provisions promises machine-accessible law and automated legal reasoning, and recent LLMs make it tempting to generate such formalizations directly from statutory text. However, any formalization makes implicit interpretive choices whose consequences are hard to anticipate, especially if an LLM is the author. We present a method for systematically comparing different formalizations of the same legal provision by their inferences on individual cases. Given multiple formalizations of a provision, we match them at the node level, derive a shared interface for each pair from the matching, and use a SAT solver to enumerate the edge cases on which any two formalizations disagree. Selected edge cases are then verbalized into concrete factual scenarios that a legal expert can examine and act on. We apply our method to formalizations of ten EU provisions generated by nine frontier LLMs. We find that behavioral divergence between formalizations is essentially uncorrelated with their structural agreement and that the verbalized cases reveal qualitatively distinct types of disagreement, including divergences that mirror genuine controversies in the legal commentary. --- ### By Their Fruits You Will Know Them: Comparing Formalizations of Law by the Decisions They Encode Authors: Julius Vernie, Matthias Grabmair Published: 2026-05-24 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2605.25186v1 Formalizing legal provisions promises machine-accessible law and automated legal reasoning, and recent LLMs make it tempting to generate such formalizations directly from statutory text. However, any formalization makes implicit interpretive choices whose consequences are hard to anticipate, especially if an LLM is the author. We present a method for systematically comparing different formalizations of the same legal provision by their inferences on individual cases. Given multiple formalizations of a provision, we match them at the node level, derive a shared interface for each pair from the matching, and use a SAT solver to enumerate the edge cases on which any two formalizations disagree. Selected edge cases are then verbalized into concrete factual scenarios that a legal expert can examine and act on. We apply our method to formalizations of ten EU provisions generated by nine frontier LLMs. We find that behavioral divergence between formalizations is essentially uncorrelated with their structural agreement and that the verbalized cases reveal qualitatively distinct types of disagreement, including divergences that mirror genuine controversies in the legal commentary. --- ### Knowledge Graph-Driven Expert-Level Reasoning for Neuroscience Authors: Jake Stephen, Niraj K. Jha Published: 2026-05-24 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2605.25183v2 Knowledge graph (KG) is an abstraction that can be extracted from text corpora and used for in-depth reasoning. Prior work has leveraged KGs to fine-tune language models (LMs), enabling domain-specific superintelligence. In this work, we explore whether KG-driven in-depth reasoning capabilities can emerge in neuroscience using only information contained within a single authoritative textbook. The central hypothesis is that structured knowledge, when distilled into a high-quality KG and converted into KG-grounded question-answer (QA) supervision, is sufficient to produce expert-level reasoning through a fine-tuned LM that surpasses large language models (LLMs) in accuracy, while employing orders of magnitude fewer parameters. We construct a textbook-derived KG via a dual-LLM validation pipeline, expand it with a masked LM trained on the KG topology, generate multi-hop QA items, which include QA pairs and reasoning traces, to fine-tune an LM exclusively on KG-derived supervision, and apply reinforcement learning using path-derived KG signals as implicit reward models. Our results demonstrate that deep, mechanistic neuroscience understanding can be induced in the model without reliance on large, heterogeneous web-scale corpora. The KG-based synthetic neuroscience curriculum that readers can quiz themselves on, and the fine-tuned LM, are available at the following GitHub location: https://kg-bottom-up-superintelligence.github.io/neuro-bench. --- ### SpecAlign: A Semantic Alignment Framework for SystemVerilog Assertion Generation Authors: Jaime Rafael Imperial, Hao Zheng Published: 2026-05-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.25181v1 Existing Large Language Model (LLM) approaches to SystemVerilog Assertion (SVA) generation primarily focus on syntactic validity and formal verification outcomes, while semantic alignment between generated assertions and natural language specifications remains difficult to quantify. As a result, hallucinated or misaligned SVAs can reduce confidence and increase debugging efforts in the absence of golden RTL. This paper presents SpecAlign, a framework for semantic evaluation and refinement of LLM-generated SVAs. SpecAlign introduces two iterative alignment loops that assess both natural language properties and SVAs against the design specification using entailment-based classification. We improve alignment decisions by generating multiple reasoning paths using chain-of-thought prompting and aggregating them via a self-consistency voting mechanism. Misaligned assertions are analyzed to generate actionable feedback for refinement. We further define a quantitative alignment score to measure semantic consistency across iterations. Experimental results demonstrate that SpecAlign effectively detects semantic inconsistencies and improves assertion alignment without relying on golden RTL, providing a scalable complement to traditional formal verification evaluation metrics. --- ### Locality Matters for Training-Free Audio Token Compression in Audio-Language Models Authors: Jiale Luo, Xiaoyu Liang, Haoji Hu Published: 2026-05-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2605.25179v1 Audio-language models (ALMs) are increasingly used for audio captioning, question answering, and open-ended audio understanding, but their inference cost remains high when audio inputs are represented as long prefix-token sequences. These audio prefixes consume context budget, increase memory usage, and make deployment harder in resource-constrained or latency-sensitive settings. Existing training-free audio-token reduction methods mainly rely on fixed pooling or score-based pruning. Fixed pooling is content-agnostic, while score-based pruning can preserve isolated salient tokens but discard nearby acoustic context. We propose Local Temporal Bipartite Merging (LTBM), a training-free encoder-space compression method that merges similar nearby audio tokens under an explicit temporal window constraint. Beyond introducing LTBM, we use a controlled Global Merge variant to isolate whether temporal locality itself is a useful inductive bias for audio-token compression. Experiments on AudioCaps, Clotho, and MMAU with Qwen2-Audio show evidence of a task-dependent locality effect: locality-aware merging is more favorable for captioning at several compression settings, especially under stronger compression, while global matching is more competitive for multiple-choice audio understanding. A cross-backbone validation on Audio Flamingo 3 further supports the captioning-side advantage of locality-aware merging under moderate and aggressive compression. --- ### Discrepancy Minimization Improves Cross-Hospital Robustness in Digital Pathology Authors: Ben Vardi, Dana Schonberger, Yuval Friedmann, Zohar Yakhini, Iris Barshack, Alexander Loebel, Ariel Shamir Published: 2026-05-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.25175v1 Pathology foundation models (PFMs) have advanced rapidly in recent years and support training classifiers for a range of histopathology tasks. However, their robustness across hospitals remains limited: performance often degrades when training a classifier on data from one hospital and evaluating it on another target hospital. We address this challenge by fine-tuning PFMs with a local maximum mean discrepancy (LMMD) objective that applies to two settings: domain adaptation, where unlabeled target-hospital data is available, and domain generalization, where target-hospital data is unavailable at all. Experiments at both the patch- and slide-level show consistent improvements across multiple PFMs and tasks. ---