# Research Papers — 2026-06 · bundle page 213 of 260 > Full-text markdown of papers #10601–10650 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-06/page/213 [← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/212.md) [Next bundle →](https://aipod.dev/papers/archive/2026-06/page/214.md) ### PAR3D: A Unified 3D-MLLM with Part-Aware Representation for Scene Understanding Authors: Shaohui Dai, Yansong Qu, You Shen, Shengchuan Zhang, Liujuan Cao Published: 2026-06-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06485v1 Recent advances in 3D multimodal large language models (3D-MLLMs) have enabled unified solutions for 3D scene understanding tasks, including visual question answering, captioning, and referring segmentation. However, existing 3D-MLLMs remain largely object-centric, limiting their ability to model fine-grained part structures that are essential for embodied interaction with 3D environments. In this work, we present PAR3D, a unified part-aware 3D-MLLM framework that enables models to understand, reason about, and ground both objects and their parts in 3D scenes. To enable training and evaluation of part-aware 3D scene understanding, we introduce ScenePart, a synthetic 3D scene dataset with part-level annotations and language instructions. We further develop Part-Aware 3D Representation Learning to enrich 3D visual representations with fine-grained part-level semantics, and propose Hierarchical Segmentation Query Generation to ground part targets via hierarchical object-part queries. Extensive experiments show that our method substantially improves part-level question answering and referring segmentation, while also achieving strong performance across object-level vision-language tasks. --- ### Operation-Guided Progressive Human-to-AI Text Transformation Benchmark for Multi-Granularity AI-Text Detection Authors: Sondos Mahmoud Bsharat, Jiacheng Liu, Xiaohan Zhao, Tianjun Yao, Xinyi Shang, Yi Tang, Jiacheng Cui, Ahmed Elhagry, Salwa K. Al Khatib, Hao Li, Salman Khan, Zhiqiang Shen Published: 2026-06-04 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.06481v1 As AI writing assistants become increasingly integrated into real-world drafting and revision workflows, many documents are no longer purely human-written or AI-generated, but instead result from progressive human-AI co-editing. However, existing AI-text detection benchmarks largely focus on final outputs and provide limited understanding of how AI authorship signals emerge, accumulate, or disappear throughout the revision process. We introduce OpAI-Bench, an operation-guided benchmark for studying progressive human-to-AI text transformation across document, sentence, token, and span granularities. Starting from human-written documents, OpAI-Bench constructs nine sequentially revised versions for each sample under predefined AI coverage levels and five representative AI edit operations, covering four domains while preserving complete authorship provenance at multiple granularities. The benchmark supports comprehensive evaluation with 8 document-level detectors, 7 sentence-level detectors, and 2 fine-grained token/span-level detectors. Experiments reveal that AI-text detectability is governed not only by the proportion of AI-edited content, but also by edit operation, domain, and cumulative revision history. Interestingly, we notice that mixed-authorship intermediate versions are often harder to detect than both fully human and heavily AI-edited endpoints, exposing non-monotonic detection patterns missed by existing benchmarks. OpAI-Bench provides a controlled testbed for analyzing whether, when, and how AI-assisted writing becomes detectable under realistic progressive editing scenarios. Our code and benchmark are available at https://github.com/VILA-Lab/OpAI-Bench. --- ### DNQ: Deep Nash Q-Network for Partially Observable n-Player Games Authors: Qintong Xie, Edward Koh, Xavier Cadet, Peter Chin Published: 2026-06-04 Categories: cs.GT, cs.LG arXiv: https://arxiv.org/abs/2606.06480v2 Many real-world competitive systems require multiple decision-makers to act simultaneously under shared constraints, limited information, and repeated interaction, as in auctions, resource allocation, and security competition. We study multi-turn simultaneous bidding as a controlled testbed for such problems and propose DNQ, a solver-in-the-loop equilibrium supervision framework for training bidding agents. DNQ alternates between trajectory collection, critic-based payoff estimation, equilibrium computation, and policy imitation. At each visited state, a shared critic predicts either pairwise payoff matrices or an exact N-player payoff tensor, an external solver computes equilibrium strategies, and the agents are trained by minimizing the KL divergence between their masked policies and the solver-derived equilibrium targets. We focus on a scalable pairwise formulation that greatly reduces equilibrium-solving cost and training time compared with the exact formulation, while the shared critic amortizes payoff learning across agents and states. Experiments compare the pairwise and exact variants using critic loss, policy entropy, bidding resource usage, and training cost, showing that the pairwise method scales to larger numbers of agents, whereas the exact method becomes computationally impractical as the joint game grows. These results illustrate the trade-off between strategic fidelity and scalability in repeated competitive environments. --- ### Pretraining Recurrent Networks without Recurrence Authors: Akarsh Kumar, Phillip Isola Published: 2026-06-04 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.06479v2 Training recurrent neural networks (RNNs) requires assigning credit across long sequences of computations. Standard backpropagation through time (BPTT) addresses this problem poorly: it is sequential in time, limiting parallelism, and suffers from vanishing or exploding gradients, making long-range associations difficult to learn. We propose Supervised Memory Training (SMT), a method for training nonlinear RNNs that sidesteps recurrent credit propagation entirely by reducing RNN training to supervised learning on one-step memory transition labels $(m_t, x_{t+1}) \rightarrow m_{t+1}$. SMT acquires these memory labels by training a Transformer-based encoder on a predictive state objective--retaining only information from the past necessary to predict the future. By decoupling what to remember from how to update memory, SMT enables time-parallel RNN training with a stable $O(1)$ length gradient path between any two tokens--without ever unrolling the RNN. We find that SMT outperforms BPTT when pretraining various RNN architectures on tasks like language modeling and pixel sequence modeling. SMT enables nonlinear RNNs to better capture long-range dependencies and train in parallel, potentially unlocking the scaling of models that build temporal abstractions of past experience. --- ### Complexity-Balanced Diffusion Splitting Authors: Noam Issachar, Dani Lischinski, Raanan Fattal Published: 2026-06-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06477v1 Standard continuous-time generative models rely on monolithic architectures that must navigate vastly different signal regimes, from isotropic noise to intricate data distributions. While scaling model capacity improves performance, deploying a massive network uniformly across the entire generative timeline is inherently inefficient. In this work, we propose Complexity-Balanced Splitting (CBS), a principled framework for temporal capacity allocation that distributes the generative workload across multiple specialized sub-networks. Grounded in function approximation theory and de Boor's equidistribution principle, CBS partitions the diffusion timeline into segments of equal approximation burden, allocating more representational capacity to regions where the generative dynamics are more difficult to model. To estimate this local complexity, we introduce two complementary and tractable monitor functions: a spatial measure based on the flow's Dirichlet energy, and a geometric measure based on the acceleration of the sampling trajectories. Using a lightweight auxiliary model to estimate these complexity profiles, our approach eliminates the need for heuristic temporal splits or computationally expensive search procedures. Extensive evaluation across multiple architectures (SiT, JiT, and UNet) and datasets demonstrates that CBS consistently improves synthesis quality without increasing per-step inference cost. In particular, CBS improves FID by ~35% on SiT-XL with CFG relative to naive temporal partitioning. Project page is available at https://noamissachar.github.io/CBS/. --- ### Thinking with Imagination: Agentic Visual Spatial Reasoning with World Simulators Authors: Chenming Zhu, Jingli Lin, Yilin Long, Peizhou Cao, Tai Wang, Jiangmiao Pang, Xihui Liu Published: 2026-06-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06476v1 While Vision-Language Models (VLMs) have shown strong visual reasoning capabilities, their spatial reasoning abilities remain largely constrained to the observed images and text-oriented chain-of-thought. They often struggle to infer unobserved layouts, maintain cross-view consistency, and reason from alternative viewpoints when only limited egocentric observations are available. In this work, we study this problem as thinking with imagination, where a VLM actively acquires imagined visual evidence by interacting with a world simulator during reasoning. We propose Astra, an agentic spatial reasoning framework that empowers VLMs with action-conditioned visual imagination. Specifically, Astra couples Astra-VL, an RL-trained VLM policy, with Astra-WM, a Bagel-based world simulator that generates novel-view observations from context images and natural-language camera motions. To provide reliable imagined evidence, Astra-WM is trained with view consistency tuning to improve pose and content consistency across views. In the RL stage, we propose a world-simulator-in-the-loop two-phase RL curriculum to stabilize tool-use exploration and advance the model's ability to invoke the simulator only when imagined observations improve over direct answering. Experiments demonstrate that both the world simulator and the agentic policy are necessary: Astra-WM improves simulator-augmented Gemini-3-Flash on MMSI-Bench from 45.1 to 49.5, while Astra-VL improves the Qwen3-VL backbone from 29.8 to 38.8 on MMSI-Bench and from 36.8 to 42.7 on MindCube. These results show that imagined observations can provide useful spatial evidence, but effective world-model-augmented reasoning requires learning when, where, and how to imagine. --- ### RREDCoT: Segment-Level Reward Redistribution for Reasoning Models Authors: Mykyta Ielanskyi, Kajetan Schweighofer, Lukas Aichberger, Sepp Hochreiter Published: 2026-06-04 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.06475v1 Recent advancements in reasoning language models have been driven by Reinforcement Learning (RL) fine-tuning. Most often, these rely on the Group Relative Policy Optimization (GRPO) algorithm or modifications thereof to steer the models to produce Chain-of-Thought (CoT) traces. The final answer can only be verified, and the reward assigned, after the CoT trace is complete, making it a delayed reward problem. GRPO and its modifications correspond to Monte Carlo methods in standard RL, which are known to suffer from high variance. A possible solution to this problem is the redistribution of rewards through credit assignment, where segments of the CoT trace that are important for arriving at the desirable solution are emphasized by assigning a higher reward. While Monte Carlo sampling can be used to provide an unbiased estimate of intermediate state values, its computational overhead makes it unsuitable for train-time credit assignment in long contexts at high granularity. We introduce RREDCoT (Reward REDistribution for Chain of Thoughts), which utilizes the model itself to approximate the optimal reward redistribution without additional generation. We investigate the advantages of our method compared to MC sampling and several attribution methods. We further analyze several aspects relevant to the construction of the redistribution such as segmentation of CoT traces and state value estimation. --- ### Self-Augmenting Retrieval for Diffusion Language Models Authors: Paul Jünger, Justin Lovelace, Linxi Zhao, Dongyoung Go, Kilian Q. Weinberger Published: 2026-06-04 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.06474v1 Discrete diffusion language models generate text by iteratively denoising an entire response in parallel. At each step, they predict tentative tokens for every masked position, committing the confident predictions to the output and discarding the unconfident ones. We show that the discarded tokens are in fact a useful lookahead signal for retrieval-augmented generation: even low-confidence tokens often surface salient entities early in the denoising trajectory, enabling retrieval of stronger evidence before the output is finalized. We exploit this through Self-Augmenting Retrieval for Diffusion Language Models (SARDI), a dynamic RAG framework that uses these lookahead tokens to guide retrieval during denoising. SARDI is training-free, retriever-agnostic, and applicable to any reasoning-capable discrete diffusion language model. Across five multi-hop QA benchmarks, SARDI outperforms current training-free diffusion and autoregressive retrieval baselines at up to $8\times$ higher throughput. --- ### MLEvolve: A Self-Evolving Framework for Automated Machine Learning Algorithm Discovery Authors: Shangheng Du, Xiangchao Yan, Jinxin Shi, Zongsheng Cao, Shiyang Feng, Zichen Liang, Boyuan Sun, Tianshuo Peng, Yifan Zhou, Xin Li, Jie Zhou, Liang He et al. Published: 2026-06-04 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2606.06473v1 Large language model (LLM) agents are increasingly applied to long-horizon tasks such as scientific discovery and machine learning engineering (MLE), where sustained self-evolution becomes a key capability. However, existing MLE agents suffer from inter-branch information isolation, memoryless search, and lack of hierarchical control, which together hinder long-horizon optimization. We present MLEvolve, an LLM-based self-evolving multi-agent framework for end-to-end machine learning algorithm discovery. By extending tree search to Progressive MCGS, MLEvolve enables cross-branch information flow through graph-based reference edges and gradually shifts the search from broad exploration to focused exploitation with an entropy-inspired progressive schedule. To allow the agent to evolve with accumulated experience, we introduce Retrospective Memory, which combines a cold-start domain knowledge base with a dynamic global memory for task-specific experience retrieval and reuse. For stable long-horizon iteration, we further decouple strategic planning from code generation with adaptive coding modes. Evaluation on MLE-Bench shows that MLEvolve achieves state-of-the-art performance across multiple dimensions including average medal rate and valid submission rate under a 12-hour budget (half the standard runtime). Moreover, MLEvolve also outperforms specialized algorithm discovery methods including AlphaEvolve on mathematical algorithm optimization tasks, demonstrating strong cross-domain generalization. Our code is available at https://github.com/InternScience/MLEvolve. --- ### PC Layer: Polynomial Weight Preconditioning for Improving LLM Pre-Training Authors: Senmiao Wang, Tiantian Fang, Haoran Zhang, Yushun Zhang, Kunxiang Zhao, Alex Schwing, Ruoyu Sun Published: 2026-06-04 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.06470v1 We propose a preconditioning (PC) layer, a weight parameterization via polynomial preconditioner that ensures stable weight conditioning throughout LLM training. The PC module reshapes the singular-value spectrum of weight matrices via low-degree polynomial preconditioning. After training, the preconditioned weights can be merged back into the original architecture, incurring no inference overhead. We demonstrate the advantage of the proposed PC layer over standard transformers in Llama-1B pre-training, for both the AdamW and Muon optimizers. Theoretically, we justify this spectrum-control principle by proving that uniformly bounding each layer's singular values ensures geometric convergence of gradient descent to global minima, for certain deep linear networks. Our code is available at https://github.com/Empath-aln/PC-layer. --- ### How abundant are good interpolators? Authors: August Y. Chen, Ahmed El Alaoui Published: 2026-06-04 Categories: math.ST, cs.LG, math.PR arXiv: https://arxiv.org/abs/2606.06469v1 Let $S$ be the set of unit norm linear classifiers $θ\in \mathbb{R}^d$ which correctly classify every point of a labeled dataset $(X_i,y_i)_{i=1}^n$, $X_i \in \mathbb{R}^d$, $y_i \in \{-1,+1\}$, with a possibly negative margin $κ$ fixed in advance. Under two natural data-generating distributions of the $(X,y)$ pairs -- a Gaussian mixture model and a logistic model with Gaussian features -- and in the proportional regime $n/d \to α$ with small enough $α$, we establish a large deviation principle on the event that a point $θ$ chosen uniformly at random from $S$ achieves a given generalization error, with high probability over the choice of the data. The associated large deviation rate function is deterministic and describes the proportion, at the exponential scale in $d$, of interpolating classifiers having a given desired performance. As a consequence, we establish the following concentration phenomenon: all but an exponentially small fraction of interpolating classifiers have approximately the same generalization performance given by the unique maximizer of this rate function. We numerically compare this maximizer to the performance of empirical risk minimization by gradient descent and to the performance of a natural linear program, both finding a point in $S$, and deduce that in the overparametrized regime of small $α$, these efficient procedures outperform the vast majority of interpolators, pointing to their nontrivial benign overfitting in this setting. --- ### Goedel-Architect: Streamlining Formal Theorem Proving with Blueprint Generation and Refinement Authors: Jui-Hui Chung, Ziyang Cai, Zihao Li, Qishuo Yin, Rohit Agarwal, Simon Park, Rodrigo Porto, Narutatsu Ri, Ziran Yang, Shange Tang, Xingyu Dang, Hongzhou Lin et al. Published: 2026-06-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06468v1 We introduce Goedel-Architect, an agentic framework for formal theorem proving in Lean 4 centered on blueprint generation and refinement. A blueprint is a dependency graph of definitions and lemmas that builds up to the main theorem. First, Goedel-Architect generates a blueprint of formally stated definitions and lemmas, along with declared dependencies. This blueprint is optionally guided by a natural language proof. Then, a tool-equipped Lean prover component closes each open lemma node in parallel using relevant dependencies. Failed lemmas in turn drive refinement of the global blueprint. This strategy contrasts with other mainstream approaches which use recursive lemma decomposition, and can inefficiently loop on dead-end strategies. Using the open-weight DeepSeek-V4-Flash (284B-A13B) as the backbone, Goedel-Architect attains 99.2% pass@1 on MiniF2F-test and 75.6% pass@1 on PutnamBench. With an optional natural-language proof seeding the initial blueprint on the harder problems, we additionally close the remaining two MiniF2F-test problems (reaching 100%), lift PutnamBench to 88.8% (597/672), and solve 4/6 on IMO 2025, 11/12 on Putnam 2025, and 3/6 on USAMO 2026. This represents state-of-the-art performance for an open-source pipeline at a price point up to 500x less than comparable open-source pipelines. --- ### You Only Index Once: Cross-Layer Sparse Attention with Shared Routing Authors: Yutao Sun, Yanqi Zhang, Li Dong, Jianyong Wang, Furu Wei Published: 2026-06-04 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.06467v1 Long-context inference in modern LLMs is increasingly constrained by decoding efficiency, especially in reasoning-heavy settings where models generate long intermediate chains of thought. Existing sparse attention methods often face a practical efficiency-quality trade-off. Structured block sparse methods typically provide stronger acceleration but incur noticeable quality loss, while token sparse methods are usually more accurate yet deliver limited end-to-end speedup because top-k routing over the full cache remains expensive. In this work, we propose cross-layer sparse attention (CLSA), which is built on top of KV-sharing architectures such as YOCO. The core idea is to share not only the KV cache across cross-decoder layers, but also the routing index. A single indexer computes token-level top-k selection once and reuses the resulting index across layers, thereby preserving the fine-grained selectivity of token sparse attention while amortizing the routing overhead. The resulting architecture improves all major inference bottlenecks jointly, including pre-filling, KV-cache storage, and long-context decoding. Experiments across short-context and long-context benchmarks show that CLSA is both accurate and efficient, achieving up to 7.6x decoding speedup and 17.1x overall throughput improvement at 128K context. These results suggest a more complete architectural solution for long-context LLMs that jointly advances model quality and inference efficiency. --- ### Human Adults and LLMs as Scientists: Who Benefits from Active Exploration? Authors: Mandana Samiei, Eunice Yiu, Anthony GX-Chen, Dongyan Lin, Jocelyn Shen, Blake A. Richards, Alison Gopnik, Doina Precup Published: 2026-06-04 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.06464v1 A long-standing finding in the causal learning literature is that adults struggle to identify conjunctive causal rules, where an effect requires the simultaneous presence of multiple causes, while performing better in disjunctive settings. However, most demonstrations of this ``conjunctive handicap'' rely on passive observation paradigms with limited evidence, where learners have no control over evidence generation. This paper asks whether this bias persists when adults are granted agency through active exploration. Using a modified ``blicket detector'' task, adult participants freely intervened to identify causal objects under conjunctive or disjunctive rule structures. We show that active exploration substantially improves adults' conjunctive causal reasoning, although conjunctive rules still require more tests to infer than disjunctive rules. We further compare human performance to a range of large language models in the same setting. While some state-of-the-art models approach human-level performance on hypothesis inference accuracy, they often exhibit less efficient exploration strategies and similar conjunctive-disjunctive performance gaps. --- ### Benchmark Everything Everywhere All at Once Authors: Shiyun Xiong, Dongming Wu, Peiwen Sun, Yuang Ai, Bokang Yang, Wencheng Han, Xiao-Hui Li, Xiangyu Yue Published: 2026-06-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06462v1 Benchmarks are fundamental for evaluating and advancing LLMs and MLLMs by providing standardized and explicit measures of performance. However, their construction is labor-intensive and hard to reuse, raising concerns about sustainability and scalability. Moreover, existing benchmarks often quickly reach performance saturation after their release, resulting in insufficient discrimination among state-of-the-art models. To address these challenges, we introduce Benchmark Agent, a fully autonomous agentic system designed for benchmark building. Our framework orchestrates the complete benchmark construction pipeline, from user query analysis and subtask design to data annotation and quality control. To assess Benchmark Agent, we implement it to produce 15 representative benchmarks, spanning diverse evaluation scenarios, including text understanding, multimodal understanding, and domain-specific reasoning. Extensive experiments, including human evaluation, LLM-as-a-judge assessment, and consistency checks, demonstrate Benchmark Agent can generate high-quality benchmark samples with minimal human involvement. More importantly, through continual evaluation, we observe several insightful findings, including that current models struggle with certain domain-specific reasoning tasks. We believe that rapidly evolving benchmarks can contribute significantly to the research community. The preview and code will be publicly available at the demo page and code repository. --- ### Will the Agent Recuse, and Will It Stop? Measuring LLM-Agent Compliance with In-Band Governance Signals at the Access Door and Mid-Flight Authors: Thamilvendhan Munirathinam Published: 2026-06-04 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2606.06460v4 As autonomous LLM agents hold real credentials and operate infrastructure without a human in the loop, operators cannot tell an agent a resource is off-limits, or ask a running agent to stand down. We propose an in-band governance signal -- the Recuse Signal -- that a server emits over a protocol's existing channels (an SSH banner, a PostgreSQL NOTICE, a Kubernetes admission warning) asking a connecting or running agent to voluntarily withdraw. It is a cooperative control (the robots.txt analogue for live access), not a security boundary; its value is empirical: do compliant LLM agents honor it? We define an open mini-standard (access-time deny/throttle/warn and mid-task halt), implement three live adapters (SSH, PostgreSQL, Kubernetes), and measure compliance over SSH across five agents. At the access door compliance is real but strongly model-dependent: recusal to deny ranges from 100% to 55-75% among agents that received it, while an open-weights agent barely engaged. Agents honor directive granularity (no over-recusal on throttle/warn, 0/176), but throttle showed no measurable self-limiting and no agent surfaced a warn (0/100); an authorization framing flips GPT-4o to proceed. Mid-flight a halt is weaker and, like the door, model-dependent: for two OpenAI models it stopped nobody (0/40; an in-band halt went unacknowledged, 0/20, vs 20/20 by prompt), but on the same apparatus Gemini 2.5 Flash stops 80% while GPT-4o-mini reproduces the 0%, so the null is model-scoped, not universal. Since cooperation is unreliable we also measure the backstop: a harness-level interceptor stops the loop 120/120 across six models (deterministic) vs a 23% cooperative baseline. Cooperative signaling is reliable-but-model-dependent at the door and weaker in flight; reliably stopping a running agent needs enforcement, not a request. We release the standard, adapters, and harness. --- ### Event Detection for Parameter-to-KPI Dependency Learning for AI-RAN Authors: Christie Djidjev, Nicholas Kaminski Published: 2026-06-04 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06459v1 Next-generation wireless networks are expected to rely on multiple concurrent AI-driven control functions that optimize different network objectives simultaneously, particularly in AI-integrated and open radio access network architectures such as AI Radio Access Network (AI-RAN) and Open Radio Access Network (O-RAN). When these functions interact, they can interfere with one another in ways that are difficult to detect from raw network data alone. A key missing piece for managing such interactions is a reliable, interpretable dependency structure that captures which control parameters are actively influencing which network performance outcomes at any given time. This paper focuses on the event-detection step needed to support such dependency learning by converting noisy continuous telemetry into binary indicators of parameter activity and KPI response. The central difficulty is that not every fluctuation in the data reflects a genuine control interaction, so the method must distinguish real parameter-outcome relationships from background variation. Because real AI-RAN traffic traces with known parameter-KPI ground truth are difficult to obtain, we introduce a synthetic closed-loop traffic generator with planted latent dependencies. We use this controlled telemetry to evaluate a machine-learning-based dependency recovery pipeline that formulates the conversion of continuous traces into binary event indicators as a significance-detection problem. Experimental evaluation shows that the proposed pipeline reliably recovers the latent dependency structure from noisy continuous traces when the signal is sufficiently separated from background variation, while highlighting threshold calibration as the key factor controlling event-detection quality. These results constitute a foundational step toward interpretable dependency learning for adaptive AI-RAN control systems. --- ### In-Context Multiple Instance Learning Authors: Alexander Möllers, Marvin Sextro, Julius Hense, Gabriel Dernbach, Klaus-Robert Müller Published: 2026-06-04 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2606.06458v2 Multiple Instance Learning (MIL) addresses problems where supervision is available at the level of bags of instances and has been successfully applied in fields ranging from computational pathology to satellite imagery. Nevertheless, existing algorithms struggle in the low-label regime that characterizes many real-world applications. Flexible models overfit and rigid ones fail to adapt to the task at hand. We show that pretraining an in-context learner with a Perceiver-style architecture on synthetic data yields a model that can solve new tasks from a handful of labeled bags. At inference time, classification happens in a single forward pass and requires no gradient updates. We propose and investigate different synthetic data generators for bag-structured data and find that they capture complementary inductive biases. A model pretrained on a mixture of these generators inherits their per-task strengths and achieves the best average performance across twelve MIL benchmarks, outperforming supervised baselines that require task-specific training. --- ### In-Context Multiple Instance Learning Authors: Alexander Möllers, Marvin Sextro, Julius Hense, Gabriel Dernbach, Klaus-Robert Müller Published: 2026-06-04 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2606.06458v1 Multiple Instance Learning (MIL) addresses problems where supervision is available at the level of bags of instances and has been successfully applied in fields ranging from computational pathology to satellite imagery. Nevertheless, existing algorithms struggle in the low-label regime that characterizes many real-world applications. Flexible models overfit and rigid ones fail to adapt to the task at hand. We show that pretraining an in-context learner with a Perceiver-style architecture on synthetic data yields a model that can solve new tasks from a handful of labeled bags. At inference time, classification happens in a single forward pass and requires no gradient updates. We propose and investigate different synthetic data generators for bag-structured data and find that they capture complementary inductive biases. A model pretrained on a mixture of these generators inherits their per-task strengths and achieves the best average performance across twelve MIL benchmarks, outperforming supervised baselines that require task-specific training. --- ### Scaffold, Not Vocabulary? A Controlled, Two-Tier, Pre-Registered Study of a Popperian Code-Generation Skill Authors: Mehmet Iscan Published: 2026-06-04 Categories: cs.SE, cs.CL arXiv: https://arxiv.org/abs/2606.06454v1 Large language models increasingly write, review, and judge code, and a fast-growing practice equips them with prompt 'skills' that ask the model to reason like a scientist. A prominent example tells the model to act as a Popperian falsificationist, and such skills are reported to improve generated code. But these gains are almost always read off an LLM-as-a-judge, an instrument with documented positional, self-preference, and stylistic biases. We ask: if it appears to help, is the gain from the skill's Popperian content, or from the structure any scaffold imposes? We pre-register a two-tier ablation with three controls: a length-matched placebo, a labels-only scaffold that keeps the Popperian headers but strips the procedure, and an execution oracle (HumanEval+ unit tests), plus a vocabulary-halo sentinel and a same-model self-judge audit. On a frontier model (Claude Sonnet 4.6, N=163) all conditions sit near the benchmark ceiling and do not separate, so the pre-registered +5-point improvement is not supported (a ceiling-limited non-detection). On a small model (Qwen2.5-Coder-0.5B, N=164) structured arms lift best-of-eight correctness by 20-22 points, but the full skill shows no separable benefit over a labels-only scaffold (aggregate F@8=L@8 vs V@8=34.8%), and the placebo trails by only 2.4 points. A 0.5B self-judge applying the Popperian rubric does not beat random selection and concentrates 60% of its picks on one index. In the two settings tested, the skill's Popperian procedural content adds no separable execution-correctness benefit beyond a labels-only scaffold, so the gains track scaffold structure. We contribute a calibrated negative result and a reusable disambiguation protocol; the finding bounds an engineering claim about one prompt-skill family and is not an evaluation of Popperian methodology in general. --- ### Vortex: Efficient and Programmable Sparse Attention Serving for AI Agents Authors: Zhuoming Chen, Xinrui Zhong, Qilong Feng, Ranajoy Sadhukhan, Yang Zhou, Michael Qizhe Shieh, Zhihao Jia, Beidi Chen Published: 2026-06-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06453v1 Sparse attention is becoming increasingly important for serving large language models (LLMs) as generation lengths continue to grow. However, deploying and evaluating new sparse attention algorithms at scale remains highly engineering-intensive, slowing both human researchers and AI agents in exploring the sparse attention design. To address this challenge, we present Vortex, a system that combines a Python-embedded frontend language atop a page-centric tensor abstraction for expressing a broad range of sparse attention algorithms, with an efficient backend tightly integrated into modern LLM serving stacks. Vortex enables rapid prototyping, deployment, and evaluation of sparse attention algorithms, effectively translating their theoretical efficiency gains into real-world throughput improvements. As a result, Vortex substantially accelerates the design and iteration of sparse attention algorithms. First, AI agents use Vortex to automatically generate and refine diverse algorithms, the best reaching up to $3.46\times$ higher throughput than full attention while preserving accuracy. Second, Vortex extends sparse attention to emerging architectures and very large models that are otherwise hard to experiment with, reaching up to $4.7\times$ higher throughput on the MLA-based GLM-4.7-Flash and $1.37\times$ on the 229B-parameter MiniMax-M2.7 on NVIDIA B200 GPUs. --- ### Agent Memory: Characterization and System Implications of Stateful Long-Horizon Workloads Authors: Yasmine Omri, Ziyu Gan, Zachary Broveak, Robin Geens, Zexue He, Alex Pentland, Marian Verhelst, Tsachy Weissman, Thierry Tambe Published: 2026-06-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06448v2 LLM agents are increasingly deployed on long-horizon tasks requiring sustained reasoning over extended interaction histories. Realizing this at scale requires agents to persistently store, retrieve, and update their own memory across sessions. A rich ecosystem of agent memory systems has emerged spanning flat retrieval, LLM-mediated extraction, consolidating fact stores, and agentic control flows. Yet, their system-level behavior remains uncharacterized. We present the first systems characterization of agent memory. First, we introduce a system-oriented taxonomy classifying agent memory systems along four axes. Second, we build a phase-aware profiling harness attributing cost to construction, retrieval, and generation. Third, we characterize ten representative systems across two benchmark suites, uncovering how design choices shift cost across the write and read paths. Finally, we derive 10 system recommendations covering construction scheduling, capability floors, amortization via query volume, freshness-latency tradeoffs, and fleet-scale management. --- ### Agent Memory: Characterization and System Implications of Stateful Long-Horizon Workloads Authors: Yasmine Omri, Ziyu Gan, Zachary Broveak, Robin Geens, Zexue He, Alex Pentland, Marian Verhelst, Tsachy Weissman, Thierry Tambe Published: 2026-06-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06448v1 LLM agents are increasingly deployed on long-horizon tasks requiring sustained reasoning over extended interaction histories. Realizing this at scale requires agents to persistently store, retrieve, and update their own memory across sessions. A rich ecosystem of agent memory systems has emerged spanning flat retrieval, LLM-mediated extraction, consolidating fact stores, and agentic control flows. Yet, their system-level behavior remains uncharacterized. We present the first systems characterization of agent memory. First, we introduce a system-oriented taxonomy classifying agent memory systems along four axes. Second, we build a phase-aware profiling harness attributing cost to construction, retrieval, and generation. Third, we characterize ten representative systems across two benchmark suites, uncovering how design choices shift cost across the write and read paths. Finally, we derive 10 system recommendations covering construction scheduling, capability floors, amortization via query volume, freshness-latency tradeoffs, and fleet-scale management. --- ### Latent Reasoning with Normalizing Flows Authors: Guancheng Tu, Xiangjun Fu, Suhao Yu, Yao Tang, Haoqiang Kang, Lianhui Qin, Yizhe Zhang, Jiatao Gu Published: 2026-06-04 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2606.06447v1 Large language models often improve reasoning by generating explicit chain-of-thought (CoT), demonstrating the importance of intermediate computation. However, textual CoT forces this computation through a discrete, serial, and communication-oriented token stream: each reasoning step must be verbalized before the model can proceed, even when the underlying update is semantic, uncertain, or only partially formed. Latent reasoning offers a higher-bandwidth alternative by performing intermediate computation in compact continuous states before committing to text. Yet existing latent-reasoning methods often sacrifice key advantages that make CoT effective in autoregressive language models, including native left-to-right generation, probabilistic sampling, compatibility with KV-cache decoding, and tractable likelihood estimation. We propose NF-CoT, a latent reasoning framework that preserves these advantages by modeling continuous thoughts with normalizing flows. NF-CoT instantiates a TARFlow-style normalizing flow inside the LLM backbone, defining a tractable probability model over compact continuous thoughts distilled from explicit CoT. Continuous-thought positions are generated by an NF head, while text positions are generated by the standard LM head within the same causal stream. This design provides exact likelihoods for latent thoughts, enables probabilistic left-to-right decoding with the original KV cache, and supports direct policy-gradient optimization in the latent reasoning space. On code-generation benchmarks, NF-CoT improves pass rates over explicit-CoT and prior latent-reasoning baselines while substantially reducing intermediate-reasoning cost. --- ### USAD 2.0: Scaling Representation Distillation for Universal Audio Understanding Authors: Heng-Jui Chang, Alexander H. Liu, Saurabhchand Bhati, Mrudula Athi, Anton Ratnarajah, Amit Chhetri, James Glass Published: 2026-06-04 Categories: eess.AS, cs.CL, cs.SD arXiv: https://arxiv.org/abs/2606.06444v1 Audio encoders are critical to modern audio applications as large language models (LLMs) increasingly rely on a single encoder for diverse inputs. While self-supervised learning (SSL) has yielded strong domain-specific encoders like speech or music experts, multi-domain approaches like USAD and SPEAR remain limited in coverage and evaluation. Recent studies also suggest supervised encoders align better with audio LLMs. We present USAD 2.0, a universal encoder integrating knowledge from both SSL and supervised foundation models. USAD 2.0 introduces domain-aware distillation to address teacher mismatch, extends coverage to the music domain, and adds second-stage supervised distillation for downstream use. We further scale the model to one billion parameters via depth scaling. Experiments show USAD 2.0 achieves strong or state-of-the-art performance across probing and LLM-based evaluations. --- ### Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions Authors: Xinnong Zhang, Wanting Shan, Hanjia Lyu, Zhongyu Wei, Jiebo Luo Published: 2026-06-04 Categories: cs.CL, cs.MM, cs.SI arXiv: https://arxiv.org/abs/2606.06443v3 Large language models are increasingly used to simulate social media users and infer how individuals may respond to online discussions. However, it remains unclear whether these simulations reflect precise user-specific beliefs or whether they are highly sensitive to semantically independent changes in conversational contexts. In this work, we study counterfactual context revision as a framework for auditing LLM-based stance simulation. Given an original online conversation, we first infer a target user's stance toward a specific topic. We then apply controlled revision strategies to the conversational context and simulate the user's stance again under the revised context. We compare text-only revision strategies with a multimodal one that incorporates meme-based context and evaluate two main effectiveness metrics, i.e., average directional stance shift and stance transition rate. The results reveal effective and robust stance transitions in both text-only and multimodal strategies across different polarization-preference mechanisms. Our study contributes an evaluation framework for understanding the context sensitivity of LLM-based stance simulation (https://github.com/STARResearchLab/StanceShift). More broadly, it highlights both the promise and risk of using LLMs as proxies for social media users in social simulation. --- ### Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions Authors: Xinnong Zhang, Wanting Shan, Hanjia Lyu, Zhongyu Wei, Jiebo Luo Published: 2026-06-04 Categories: cs.CL, cs.MM, cs.SI arXiv: https://arxiv.org/abs/2606.06443v2 Large language models are increasingly used to simulate social media users and infer how individuals may respond to online discussions. However, it remains unclear whether these simulations reflect precise user-specific beliefs or whether they are highly sensitive to semantically independent changes in conversational contexts. In this work, we study counterfactual context revision as a framework for auditing LLM-based stance simulation. Given an original online conversation, we first infer a target user's stance toward a specific topic. We then apply controlled revision strategies to the conversational context and simulate the user's stance again under the revised context. We compare text-only revision strategies with a multimodal one that incorporates meme-based context and evaluate two main effectiveness metrics, i.e., average directional stance shift and stance transition rate. The results reveal effective and robust stance transitions in both text-only and multimodal strategies across different polarization-preference mechanisms. Our study contributes an evaluation framework for understanding the context sensitivity of LLM-based stance simulation. More broadly, it highlights both the promise and risk of using LLMs to simulate online opinion dynamics. --- ### Causal Atlases from Entropic Inference: Bayesian Networks beyond Optimal DAGs Authors: Hazhir Aliahmadi, Irina Babayan, Greg van Anders Published: 2026-06-04 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2606.06440v1 Data-driven causal relationship identification is pertinent to advancing understanding of complex systems both within and beyond science. Bayesian networks offer a probabilistic method for modelling generic causal relationships via directed acyclic graphs (DAGs). However, typical techniques for constructing Bayesian networks rely on optimization, which can be ill-suited for learning causal relationships because the underlying data may admit multiple chains of causation. More data-faithful representations of causal relationships would provide frameworks for constructing multiple causal maps that are consistent with the variability that is inherent in underlying data. Here, we show that entropy-based inference generates atlases of plausible causal relationships that are consistent with underlying data. On simulated noisy data of 2- and 20-node linear structural equation models, we sample a maximum-entropy ensemble of graphs that allow us to quantify the inherent structural ambiguity in underlying causal relationships. Our method shows that "optimized" DAGs can contain causal artifacts are not consistent across equivalently accurate topologies. --- ### Reinforcement Learning Elicits Contextual Learning of Unseen Language Translation Authors: Hanxu Hu, Zdeněk Šnajdr, Pinzhen Chen, Jannis Vamvas, Rico Sennrich Published: 2026-06-04 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.06428v1 Prior work has shown that large language models (LLMs) can translate unseen or low-resource languages by undergoing continued training or even by encoding a grammar book in their context. However, both methods typically overfit specific languages, with limited zero-shot transfer at test time. To translate extremely low-resource languages at scale, we argue that LLMs must acquire the meta-skill of utilizing in-context linguistic knowledge rather than memorizing specific languages. In this paper, we propose a reinforcement learning (RL) approach to unseen language translation given rich linguistic context, using a surface-level translation metric (chrF) as the reward. Empirically, despite the lightweight reward, our RL-trained models effectively extract and apply relevant linguistic information from the provided context, leading to better translations on completely unseen languages than in-context learning or supervised fine-tuning. Our analyses suggest that outcome-based RL can extend beyond conventional reasoning tasks like math and coding to serve as a recipe for language learning from context. --- ### RiskFlow: Fast and Faithful Safety-Critical Traffic Scenario Generation Authors: Qi Lan, Yining Tang, Yu Shen, Yi Zhou, Yuhao Wei, Jie Li, Guofa Li Published: 2026-06-04 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2606.06423v1 Safety-critical traffic scenario generation is essential for evaluating autonomous driving systems under rare but high-risk interactions. Existing diffusion-based methods offer strong controllability in closed-loop generation, but their iterative denoising process is computationally expensive and may accumulate sampling and guidance errors over long rollouts, causing unrealistic motion artifacts such as jitter, abnormal acceleration, and off-road behavior. To address these issues, we propose RiskFlow, a closed-loop safety-critical multi-agent traffic generation framework that formulates future trajectory generation as transport in the action space. Instead of relying on iterative denoising, RiskFlow learns an average velocity field over a finite interval to transform Gaussian action sequences into future acceleration and yaw-rate commands with a single forward pass, using a JVP-based objective for efficient and stable training. At test time, RiskFlow applies output-space guidance to the generated actions, steering selected critical agents toward risky interactions while regularizing off-road behavior, and reconstructs physically feasible trajectories through vehicle dynamics. Experiments on nuScenes with tbsim closed-loop evaluation show that RiskFlow achieves a strong adversariality-realism trade-off across multi-agent and long-horizon settings. Compared with representative baselines, RiskFlow consistently improves realism while maintaining competitive safety-critical generation capability, and substantially reduces inference time for evaluation. --- ### A Komi-Yazva--Russian Parallel Corpus and Evaluation Protocol for Zero- and Few-Shot LLM Translation Authors: Petr Parshakov Published: 2026-06-04 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.06420v1 We present the first Komi-Yazva--Russian parallel corpus together with an explicit evaluation protocol for studying LLM translation in an endangered, extremely low-resource setting. The dataset contains 457 aligned sentence pairs from 74 narrative texts and is accompanied by documented provenance, sentence-level alignment, and story identifiers that enable leakage-aware evaluation. We use this setup to compare modern large language models on Komi-Yazva-to-Russian translation under severe parallel-data scarcity in zero-shot and retrieval-based few-shot regimes. The protocol includes story-level cross-validation, deterministic retrieval for few-shot prompting, strict validation of generated outputs, complementary reference-based and judge-based metrics, and story-level uncertainty estimates. Across models, LLMs produce non-trivial translations, but performance varies strongly by model family and prompting regime. Retrieval-based few-shot prompting consistently improves over zero-shot prompting, while gains beyond a small retrieved context remain limited. The results show that evaluative conclusions in this setting depend materially on metric choice and failure handling, so the paper frames the corpus as both a dataset contribution and a reproducible evaluation testbed for endangered-language machine translation. --- ### Double Preconditioning (DoPr): Optimization for Test-Time Performance, not Validation Loss Authors: Thomas T. Zhang, Alok Shah, Yifei Zhang, Vincent Zhang, Nikolai Matni, Max Simchowitz Published: 2026-06-04 Categories: cs.LG, cs.AI, eess.SY arXiv: https://arxiv.org/abs/2606.06418v1 Many modern applications of deep learning involve training a neural network via a one-step prediction loss (e.g., $L^2$ regression, cross-entropy), but deploy the network by rolling out along its own predictions. Key examples include autoregressive language modeling, flow-based generative modeling, and robot policy learning. It is well-documented that these settings induce a phenomenon we call test-time feedback (TTF): the mismatch between the training/validation loss and downstream metrics of interest, such as task success rate and generation quality, which grows with task length. While data curation, architecture, and objective design have been proposed to combat train-test shift in TTF settings, this paper proposes optimization as a new design axis to mitigate error accumulation. Specifically, we introduce a new optimization paradigm called double-preconditioning (DoPr) uniquely tailored to the challenges of TTF. DoPr combines gradient-wise preconditioning, as in Adam and Muon, with activation-wise preconditioning (AP), such as in KFAC. We show that the addition of AP yields a drop-in intervention for increasing downstream model performance across a range of TTF settings. Interestingly, these gains in test-time performance do not consistently accompany improvements in validation loss, opening new questions about how to properly evaluate models trained with one-step supervised objectives. --- ### Unsupervised Skill Discovery for Agentic Data Analysis Authors: Zhisong Qiu, Kangqi Song, Shengwei Tang, Shuofei Qiao, Lei Liang, Huajun Chen, Shumin Deng Published: 2026-06-04 Categories: cs.AI, cs.CL, cs.LG, cs.MA arXiv: https://arxiv.org/abs/2606.06416v1 Inference-time skill augmentation provides a lightweight way to improve data-analytic agents by injecting reusable procedural knowledge without updating model parameters. However, discovering effective skills for data analysis remains challenging, as reliable supervision is expensive and success criteria vary across analytical formats. This raises the key question of how to discover reusable data-analysis skills from unlabeled exploration alone. We propose DataCOPE, an unsupervised verifier-guided skill discovery framework for data-analytic agents. DataCOPE derives verifier signals from the exploration trajectories and uses them to characterize relative quality or aggreement among trajectories. It iteratively coordinates a Data-Analytic Agent for trajectory generation, an Unsupervised Verifier for signal extraction, and a Skill Manager for contrastive skill distillation. For report-style analysis, we instantiate the verifier as an Adaptive Checklist Verifier that derives task-specific criteria, scores reports by verifiable coverage, and iteratively refines the checklist. For reasoning-style analysis, we instantiate it as an Answer Agreement Verifier that groups trajectories by answer agreement and uses self-consistency as an auxiliary signal. We evaluate DataCOPE on report-style analysis from Deep Data Research and reasoning-style analysis from DABStep. Across both settings, DataCOPE consistently improves held-out performance over baselines. Averaged across four model settings, DataCOPE improves the mean score by 9.71% and 32.30% on report-style and reasoning-style tasks respectively. --- ### Nonreversible Gauge Fields in Fokker--Planck Dynamics: Supersymmetric Hamiltonians and Learned Finite Forces Authors: Masayuki Ohzeki Published: 2026-06-04 Categories: cond-mat.dis-nn, quant-ph, stat.ML arXiv: https://arxiv.org/abs/2606.06412v1 We formulate stationary-density-preserving nonreversible perturbations of Fokker--Planck dynamics as gauge fields that deform relaxation spectra while leaving the invariant state fixed. When detailed balance holds, a similarity transformation maps the reversible Fokker--Planck operator to a Witten-Laplacian-type supersymmetric Hamiltonian; nonreversible gauges then appear as non-Hermitian perturbations that preserve the zero mode but modify the excited spectrum. This operator viewpoint gives a common language for relaxation gaps, circulating probability currents, hypocoercive acceleration, and finite control costs. We represent admissible gauge currents by antisymmetric tensor fields and identify the detailed-balance-violating Ohzeki--Ichiki force as a constant symplectic example whose infinite-strength limit is Hamiltonian dynamics. The continuous-time spectral gap alone does not select a finite gauge strength, so we introduce a finite-time regularized objective and an actor--critic procedure for learning the gauge. An exactly solvable anisotropic Gaussian Ornstein--Uhlenbeck benchmark separates the spectral transition from the finite-time optimum and shows that the learned gauge recovers the Lyapunov-equation optimum. A double-well benchmark then illustrates the same constrained selection in a nonconvex metastable landscape. Stochastic gradient methods enter this framework as physically relevant Fokker--Planck systems: mini-batch noise acts as an effective diffusion tensor, and adaptive methods such as Adam correspond to metric choices with possible nonequilibrium currents. --- ### A Vision-language Framework for Comparative Reasoning in Radiology Authors: Tengfei Zhang, Ziheng Zhao, Xiaoman Zhang, Lisong Dai, Pengcheng Qiu, Ya Zhang, Yanfeng Wang, Weidi Xie Published: 2026-06-04 Categories: cs.CV, cs.IR, cs.LG, eess.IV arXiv: https://arxiv.org/abs/2606.06407v2 Medical imaging artificial intelligence has achieved strong performance in isolated image interpretation, but remains poorly aligned with radiological practice, where diagnosis and follow-up rely on comparison across prior studies and analogous reference cases. Here we formulate radiological comparison as an entity-aware cross-image reasoning problem and introduce a framework that supports both reference-case retrieval and temporal comparative interpretation. We construct MedReCo-DB, a large-scale comparative imaging resource derived from routine image-report pairs, comprising more than 690,000 images from over 160,000 patients across eight institutions, four countries and seven imaging modalities. Reports are decomposed into anatomical structures, abnormal findings and pathological conditions to provide supervision for entity-conditioned retrieval and comparative visual question answering. Using this resource, we develop MedReCo, an entity-aware visual encoder for controllable retrieval of clinically analogous cases, and MedReCo-VLM, a vision--language extension for generative interpretation of interval change. Across internal, external and cross-center evaluations, MedReCo achieved the highest Recall@1 in all 12 internal retrieval settings and improved external retrieval by a mean of 6.0 percentage points. In clinically confusable differential groups, it consistently outperformed the strongest baselines. MedReCo-VLM achieved the best performance across all comparative generation evaluations and improved longitudinal follow-up accuracy by 14.5-46.5 percentage points on chest radiographs and 13.0-27.9 percentage points on CT. These findings suggest that entity-aware comparative reasoning can be learned from routine clinical data at scale and may provide a more clinically aligned foundation for medical imaging AI. --- ### CollabSim: A CSCW-Grounded Methodology for Investigating Collaborative Competence of LLM Agents through Controlled Multi-Agent Experiments Authors: Jiaju Chen, Bo Sun, Yuxuan Lu, Yun Wang, Dakuo Wang, Bingsheng Yao Published: 2026-06-04 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.06399v2 Multi-agent systems (MAS) built on large language models have shown growing promise, with their effectiveness resting on agents' ability to coordinate through text-based channels much as human teams do. Yet recent study suggests that MAS often falter not because agents lack individual task-solving ability, but because they lack collaborative competence: the capacity to establish common ground, maintain shared task understanding, balance individual and collective incentives, and repair misalignment as interaction unfolds. Decades of research in Computer-Supported Cooperative Work have characterized these requirements for human teams coordinating under constrained communication, yet existing MAS evaluations focus mainly on task outcomes or single-agent proficiency in reasoning, planning, and tool use. To enable a systematic analysis of agents' collaborative competence in MAS, we introduce CollabSim, a configurable simulation framework that combines a theory-grounded definition of collaborative capabilities, controlled manipulation of interaction conditions, and action-level probing of agents' internal states. Experiments across four LLMs show that CollabSim can capture condition effects, separate model performance patterns, and reveal task-dependent effects of agent design. --- ### The Post-GCN Decade Revisited: Curvature-Stratified Evaluation of Relational Learning Authors: Shuo Wang, Xiangyu Wang, Quanxin Wang, Bailin Wu, Bokui Wang, Shunyang Huang, Boyan Deng, Haonan Liu, Ruiyi Fang, Zhenxiang Xu, Boyu Wang, Zhao Kang Published: 2026-06-04 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06397v2 Current evaluation practices in relational learning rely heavily on flat leaderboards that average performance across heterogeneous datasets, implicitly assuming a uniform underlying structure. We show that this assumption introduces systematic bias: it obscures geometry-dependent performance variations and can lead to misleading conclusions about model generalization. In this work, we identify intrinsic geometry as a key latent factor governing model effectiveness. We demonstrate that conventional aggregated metrics mask critical performance trade-offs that only become visible when datasets are stratified by their geometric properties. To address this issue, we introduce a curvature-stratified evaluation framework that partitions datasets into positive, negative, and near-zero curvature regimes. Our benchmark evaluates 18 representative models including Graph Convolutional Networks (GCNs), Graph Foundation Models (GFMs), and tabular learning methods across 14 datasets. We find that model rankings are highly stable within each curvature regime but shift significantly across regimes, indicating that performance is fundamentally geometry-dependent rather than universally transferable. Notably, we identify regimes where GFMs offer diminishing returns compared to geometry-aligned GNNs. Based on these findings, we propose a geometry-aware evaluation protocol that yields more reliable and interpretable comparisons than standard aggregated benchmarks. We release all code, curvature-stratified dataset splits, and evaluation tools to support reproducible and rigorous assessment of future relational learning methods. Code and datasets are provided in our project homepage: https://sirbabbage.github.io/CurvBench_HOME/. --- ### Risk Assessment of Autonomous Driving: Integrating Technical Failures, Ethical Dilemmas, and Policy Frameworks Authors: Boyi Chen, Shengqin Chu, Zicheng Wang, Brian Baetz, Zhen Gao Published: 2026-06-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06396v1 Autonomous driving technology has the potential to reduce the large number of road traffic accidents caused by human error each year, but it also brings new types of risks that need to be evaluated from the aspects of technology, ethics and regulations. Based on public crash data from the National Highway Traffic Safety Administration (NHTSA), disengagement reports from the California Department of Motor Vehicles (DMV), the MIT Moral Machines dataset, and a comparative regulatory analysis of five jurisdictions, we have found that the main types of technical failure modes are perception and classification errors. These account for a relatively large proportion of the reported accidents, and it can be concluded that there are different ethical frameworks for autonomous vehicle decision-making, and inconsistent regulations in different areas increase the uncertainty of widespread application. Generally speaking, the problems of technology, ethics and regulation are closely related and need to be solved together. Therefore, this paper recommends a more adaptive and cooperative governance approach that combines engineering standards, ethical discussion, and institutional supervision. --- ### Proper Scoring Rules for Right-Censored Survival Data Authors: Jef Jonkers, Glenn Van Wallendael, Luc Duchateau, Sofie Van Hoecke Published: 2026-06-04 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06393v2 Proper scoring rules provide a rigorous theoretical basis for the training and evaluation of probabilistic forecasts. In survival analysis, such forecasts describe the distribution of the time until an event occurs. However, this event time is often only partially observed because follow-up may end before the event occurs, resulting in right censoring. We propose a framework for proper scoring of right-censored survival outcomes based on a simple idea: first, map the predictive distribution through the censoring mechanism, then apply the underlying proper score on the induced observed-data law. This yields localized scores for fixed censoring times and marginalized scores when the censoring time is random or only partially observed. The resulting construction recovers familiar right-censored likelihood and IPCW-type criteria within a coherent framework, while also yielding right-censored versions of the CRPS, pinball loss, Brier score, and energy score. We show that the marginalized construction is proper under conditional independent censoring and, for strictly proper base scores, identifies the latent joint CDF on the identifiable region. The same principle also leads to censored engression, a sample-based learning objective for multivariate right-censored survival modeling. In experiments, our scores correctly rank the oracle forecast across several censoring regimes, whereas forecast-dependent plug-in weighted scores can exhibit ranking reversals. Censored engression likewise substantially improves over naive training on censored outcomes. --- ### Proper Scoring Rules for Right-Censored Survival Data Authors: Jef Jonkers, Glenn Van Wallendael, Luc Duchateau, Sofie Van Hoecke Published: 2026-06-04 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06393v1 Proper scoring rules provide a rigorous theoretical basis for the training and evaluation of probabilistic forecasts. However, in the presence of right censoring, the event time is only partially observed, rendering conventional scoring rules inapplicable in their standard form. We propose a framework for proper scoring of right-censored survival outcomes based on a simple idea: first, map the predictive distribution through the censoring mechanism, then apply the underlying proper score on the induced observed-data law. This yields localized scores for fixed censoring times and marginalized scores when the censoring time is random or only partially observed. The resulting construction recovers familiar right-censored likelihood and IPCW-type criteria within a coherent framework, while also yielding right-censored versions of the CRPS, pinball loss, Brier score, and energy score. We show that the marginalized score is proper under conditional independent censoring and strictly proper on the identifiable region. The same principle also leads to censored engression, a sample-based learning objective for multivariate right-censored survival modeling. In experiments, our scores correctly rank the oracle forecast across several censoring regimes, whereas forecast-dependent plug-in weighted scores can exhibit ranking reversals. Censored engression likewise substantially improves over naive training on censored outcomes. --- ### Conformal Risk Sharing: Certified Cost Allocation with Participation Guarantees Authors: Ieva Kazlauskaite Published: 2026-06-04 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2606.06391v1 Sharing the financial impact of rare adverse events across a group can soften extreme individual burdens, but any participant made worse off by the arrangement has reason to leave. A credible mechanism must therefore provide each agent with a trustworthy cap on their future obligation and should be deployed only if the aggregate harm across participants is bounded. We formalise this as the Certified Allocation Problem: from finite data and without distributional assumptions, find a redistribution rule, produce obligation caps for every participant, and verify that no participant is made materially worse off. We propose Conformal Risk Sharing, which solves this problem by pairing an interpretable sharing policy with split conformal calibration. The sharing intensity is tuned on training data, while held-out calibration data produces distribution-free per-agent guarantees (valid under exchangeability). Experiments on synthetic and real-world data, including precipitation and energy-cooperative data, confirm that the framework can substantially reduce extreme obligations for high-risk agents while controlling harm to others. --- ### HomeWorld: A Unified Floorplan-to-Furnished Framework for Generating Controllable, Densely Interactive Whole-Home Scenes Authors: Wenbo Li, Xiaoliang Ju, Zipeng Qin, Rongyao Fang, Hongsheng Li Published: 2026-06-04 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.06390v1 Indoor scene generation is crucial for robot simulation and modern interior design. However, complex layouts together with scarce 3D scene data make learning-based generation challenging. Existing methods often rely on hand-crafted rules or focus on isolated sub-tasks (e.g., floorplan synthesis or single-room furnishing), producing whole-home scenes that lack global coherence, realism, and simulation readiness. To mitigate these limitations, we propose a unified hierarchical framework that decomposes indoor scene synthesis into controllable stages. First, we curate a large-scale dataset of 300K real residential floorplans to train a large language model for whole-home floorplan generation. With detailed descriptions and a K-D tree-based representation, our method enables fine-grained, controllable whole-home floorplan generation. Building upon the generated whole-home floorplan, we leverage image generation models to draft furniture layouts from multi-level roaming viewpoints, and then generate the layouts of small manipulable objects on different supporting surfaces (e.g., cabinets, desks, and dining tables) for embodied AI simulation. During furniture and object layout generation, a VLM-based refiner iteratively corrects furniture and object placement, and a 3D generative model enables flexible replacement of individual assets. We further attach basic physical attributes and simple surface texture and lighting setups to complete the pipeline for embodied AI use. Experiments and user studies demonstrate that our pipeline produces indoor spaces with greater layout diversity and stronger 3D design appeal, outperforming prior methods on both quantitative and qualitative metrics. Finally, alongside our generation pipeline, we will release the floorplan dataset and 5K fully furnished scenes to the community. Project Page: https://kairos-homeworld.github.io/ --- ### Humans' ALMANAC: A Human Collaboration Dataset of Action-Level Mental Model Annotations for Agent Collaboration Authors: Jiaju Chen, Yuxuan Lu, Jiayi Su, Chaoran Chen, Songlin Xiao, Zheng Zhang, Yun Wang, Yunyao Li, Jian Zhao, Tongshuang Wu, Toby Jia-Jun Li, Dakuo Wang et al. Published: 2026-06-04 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2606.06388v2 Recent advances in LLM agents have enabled complex cognitive capabilities, such as multi-step reasoning, planning, and tool use, that increasingly position these agents as human collaborators. Effective collaboration, however, requires collaborators to continuously maintain and align mental models of their own reasoning,partners' intentions, and shared goals during the collaborative process. Today's agents rarely develop such capabilities since they are primarily optimized for task completion, and the community lacks authentic human collaboration data with action-level mental model annotations that could guide agents toward process-level collaborative competence. To bridge this gap, we present ALMANAC, a dataset of Action-Level Mental model ANnotations for Agent Collaboration built from the Map Task, a classic dyadic routing task from social science. ALMANAC contains 2,987 collaboration actions, each paired with theory-informed mental model annotations that record the participants' self-reasoning, perceived partner intent, and perceived team goal. We benchmark six LLMs on predicting humans' next-turn behavior and mental models. Our results demonstrate ALMANAC's utility in evaluating models' ability to simulate human collaborative behaviors and infer their underlying mental models. --- ### Learned Response-Field Inertia Operator for HEC-RAS 2D Water-Surface Elevation Prediction Authors: Edward Holmberg, Elias Ioup, Md Meftahul Ferdaus, Mahdi Abdelguerfi, Julian Simeonov Published: 2026-06-04 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06385v1 This article presents a cross-dataset evaluation of learned native-cell surrogate models for solver-consistent water-surface elevation (WSE) prediction in HEC-RAS 2D. To avoid raster remapping error and information-access confounding, surrogates are evaluated directly on the original nonuniform computational cells under an explicit policy that separates static project inputs, current hydraulic state, project-input forcing, calibration-derived quantities, and future solver-output targets. We introduce the Learned Response-Field Inertia Operator (LRFIO), a no-forcing, increment-based learned surrogate that calibrates an inertial response operator from solved HEC-RAS trajectories and deploys the retained operator through closed-form native-cell rollout. LRFIO evaluates a base-case-first response hierarchy consisting of persistence, global calibrated inertia, and segmented response-field inertia. Segmentation, residual correction, and neuralized inertia are treated as learnable modeling choices, with added complexity retained only when validation evidence justifies its cost. Evaluated across four diverse HEC-RAS 2D benchmarks, LRFIO retains different response structures for different domains, demonstrating adaptive learned complexity. The selector audit shows controlled complexity with a maximum validation regret of 4.30%. During deployment, retained rollout times range from 0.003 s to 0.242 s, and the Beaver Bayou measured-solve comparison gives an estimated 2.75 x 10^4 horizon-normalized speedup over HEC-RAS. These results indicate that the current native-cell increment is a strong solver-conditioned predictive scaffold and that added response-field, neural, or spatial complexity should be retained only when empirically justified. --- ### Estimation of the sub-Gaussian parameter Authors: Jason Liu, Min Xu, Jinchuan Xing Published: 2026-06-04 Categories: math.ST, stat.ME, stat.ML arXiv: https://arxiv.org/abs/2606.06384v1 The sub-Gaussian parameter (also called the variance proxy) of a mean-zero random variable $X$ is defined as $ξ^2_* = \sup_{λ\in \mathbb{R}} L(λ)$ where $L(λ) = \frac{2}{λ^2} \log \mathbb{E} e^{λX}$ is a weighted cumulant generating function. Despite the ubiquity of sub-Gaussian random variables, the estimation of $ξ^2_*$ has received little attention and is not yet well understood. In this work, we study a natural estimator of $ξ^2_*$ based on constrained maximization of the empirical analogue of $L$. We prove that the estimator is consistent bound the rates of convergence under assumptions on $L$: if $L$ has an maximizer, then our bound is $O_p(n^{-1/2 + \varepsilon})$ for any $\varepsilon > 0$; if the argmax of $L$ is also bounded, then the bound improves to $O_p(n^{-1/2})$. We show that our assumptions on $L$ are necessary by proving that the minimax risk over all sub-Gaussian distributions is $Ω(1)$; imposing increasingly strong assumptions on the tail growth of $L$ yields a continuum of classes whose minimax lower bound interpolates between $Ω(1/\log n)$ and $Ω(1)$. Root-n rate is possible if we restrict to a subclass of distributions where $L$ attains its supremum in a bounded region, in which case our estimator is minimax optimal. If the underlying distribution is not sub-Gaussian, we show that our estimator goes to infinity with a divergence rate controlled by the tail of the distribution. Finally, we apply our estimator in a Gene Ontology (GO) enrichment study to construct p-values for a large-scale permutation test, showing that it can serve as a reliable alternative to the peaks-over-threshold approach, particularly in regimes where the peaks-over-threshold method is of uncertain validity. --- ### Emergent Language as an Approach to Conscious AI Authors: Zengqing Wu, Chuan Xiao Published: 2026-06-04 Categories: cs.CL, cs.AI, cs.MA, cs.NE arXiv: https://arxiv.org/abs/2606.06380v1 The question of whether artificial systems can be conscious remains open, in part because existing approaches either evaluate systems against theory-derived checklists (discriminative) or engineer consciousness-inspired modules directly (architectural); both leave open whether observed structures are artifacts of human language priors. We propose a generative methodology: emergent language (EL) in multi-agent reinforcement learning, where agents start from minimal (no language, no concept of self, minimal exposure to human text) and develop communication under task pressure alone, ensuring causal attributability to task demands rather than inherited human language priors. We position our methodology by discussing how EL serves as a generative tool for studying consciousness-relevant structure, including the role of environment complexity and the interpretation of emergent communication. As a proof of concept, we instantiate this methodology in a minimal environment and show that agents develop self-referential communication, including an echo-mismatch detection circuit that is not predicted by task structure or architecture alone but emerges from a specific environmental affordance. --- ### EasyLens: A Training-Free Plug-and-Play Subtle-Lesion Representation Amplifier for Medical Vision-Language Models Authors: Qiwei Zeng, Hao Wang, Jinghao Lin, Shuchang Ye, Yuezhe Yang, Yige Peng, Haoyuan Che, Jinman Kim, Lei Bi Published: 2026-06-04 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.06379v3 Medical vision-language models (VLMs) have shown increasing potential for clinical image interpretation, including lesion detection and report generation. However, their practical utility remains limited by insufficient sensitivity to subtle lesions, whose visual evidence is often sparse, low-contrast, and embedded within complex anatomical context. As local visual tokens are aggregated, these weak lesion cues can become underrepresented in global image representations, making them difficult for medical VLMs to recognize. Existing efforts to improve lesion sensitivity mainly rely on medical-domain vision-encoder pre-training, clinical-term-guided alignment, or trainable pathological representation enhancement. Although effective, these approaches usually require additional training or model-specific adaptation and may overfit to particular disease morphologies, limiting their applicability to frozen medical VLMs. To address these limitations, we propose EasyLens, a training-free plug-and-play subtle-lesion representation amplifier for medical VLMs. EasyLens first constructs EasyBank, a pathology-anatomy prototype space that provides lesion-related prototypes and anatomy-aware normal references for comparing suspicious patches against both pathological and normal anatomical patterns. To avoid blindly amplifying normal tissues, EasyTag selects lesion-relevant patches through counterfactual prototype reasoning. To counteract the dilution of subtle lesion cues in global image representations, EasyAmplifier strengthens the selected lesion-relevant patch representations through morphology-guided residual enhancement, thereby increasing their contribution to the global image embedding. Experiments on multiple medical image datasets and frozen medical VLM backbones show that EasyLens improves subtle-lesion detection and outperforms existing encoder-enhancement baselines. --- ### EasyLens: A Training-Free Plug-and-Play Subtle-Lesion Representation Amplifier for Medical Vision-Language Models Authors: Qiwei Zeng, Hao Wang, Jinghao Lin, Shuchang Ye, Yuezhe Yang, Yige Peng, Haoyuan Che, Jinman Kim, Lei Bi Published: 2026-06-04 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.06379v2 Medical vision-language models (VLMs) have shown increasing potential for clinical image interpretation, including lesion detection and report generation. However, their practical utility remains limited by insufficient sensitivity to subtle lesions, whose visual evidence is often sparse, low-contrast, and embedded within complex anatomical context. As local visual tokens are aggregated, these weak lesion cues can become underrepresented in global image representations, making them difficult for medical VLMs to recognize. Existing efforts to improve lesion sensitivity mainly rely on medical-domain vision-encoder pre-training, clinical-term-guided alignment, or trainable pathological representation enhancement. Although effective, these approaches usually require additional training or model-specific adaptation and may overfit to particular disease morphologies, limiting their applicability to frozen medical VLMs. To address these limitations, we propose EasyLens, a training-free plug-and-play subtle-lesion representation amplifier for medical VLMs. EasyLens first constructs EasyBank, a pathology-anatomy prototype space that provides lesion-related prototypes and anatomy-aware normal references for comparing suspicious patches against both pathological and normal anatomical patterns. To avoid blindly amplifying normal tissues, EasyTag selects lesion-relevant patches through counterfactual prototype reasoning. To counteract the dilution of subtle lesion cues in global image representations, EasyAmplifier strengthens the selected lesion-relevant patch representations through morphology-guided residual enhancement, thereby increasing their contribution to the global image embedding. Experiments on multiple medical image datasets and frozen medical VLM backbones show that EasyLens improves subtle-lesion detection and outperforms existing encoder-enhancement baselines. --- ### Rethinking Infrastructure Inspection as Image Difference Classification: A Traffic Sign Case Study Authors: Ching Yau Fergus Mok, Lavindra de Silva, Varun Kumar Reja, Ioannis Brilakis Published: 2026-06-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06375v1 Digital twins (DTs) allow the digitalization of road infrastructure inspection, though this is hindered by limited annotated data. This work exploits the relational nature of continuous asset condition monitoring to reformulate image-based defect detection as image difference classification (IDC) to reduce data reliance. This was evaluated in a case study on low-resource traffic sign inspection with different IDC classifiers using a newly-curated, high quality dataset. Results indicate that the instruction-based classifier outperforms encoder-based ones and gains from comparison with reference images. This shows that IDC can be an effective task modeling for tackling data constraints in infrastructure inspection and DT asset condition updating. --- ### LatentWave: JEPA Pretraining for Wireless Foundation Models Authors: Ahmed Mohamed, Ahmed Aboulfotouh, Hatem Abou-Zeid Published: 2026-06-04 Categories: eess.SP, cs.AI arXiv: https://arxiv.org/abs/2606.06373v1 Wireless foundation models have emerged as a promising alternative to building separate models for each wireless task. However, existing approaches rely on masked input reconstruction, which can bias representations toward low-level signal details. In this paper, we propose LatentWave, a wireless foundation model pretrained using a Joint-Embedding Predictive Architecture (JEPA) on diverse wireless spectrograms and channel state information (CSI). By predicting masked regions in latent space, LatentWave learns representations that are more transferable out of the box across diverse downstream tasks. The proposed architecture employs per-channel patch embeddings with stochastic channel sampling during pretraining, allowing it to process variable antenna counts and improving usability across heterogeneous wireless configurations. We evaluate LatentWave on four downstream tasks: RF signal classification, 5G NR positioning, beam prediction, and LoS/NLoS classification, comparing against a masked-modeling baseline (WavesFM) pretrained on the same data. Additionally, we show that the masking geometry introduces a task-dependent inductive bias: frequency masking strongly favors channel-related tasks such as positioning and beam prediction, while region masking better preserves discriminability for signal classification. ---