# Research Papers — 2026-05 · bundle page 268 of 326 > Full-text markdown of papers #13351–13400 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/268 [← Previous bundle](https://aipod.dev/papers/archive/2026-05/page/267.md) [Next bundle →](https://aipod.dev/papers/archive/2026-05/page/269.md) ### EULER-ADAS: Energy-Efficient & SIMD-Unified Logarithmic-Posit Engine for Precision-Reconfigurable Approximate ADAS Acceleration Authors: Mukul Lokhande, Ratko Pilipovic, Omkar Kokane, Adam Teman, Santosh Kumar Vishvakarma Published: 2026-05-07 Categories: cs.AR, cs.AI, cs.CV, eess.IV, math.NA arXiv: https://arxiv.org/abs/2605.06875v1 Advanced driver-assistance systems (ADAS) require neural compute engines that deliver low-latency inference under strict power and area constraints. Posit arithmetic is attractive for such accelerators because it provides high numerical fidelity at low precision, but its variable-length regime encoding increases encode/decode cost and exposes the datapath to large regime-field fault effects. This paper presents EULER-ADAS, a SIMD-enabled logarithmic bounded-Posit neural compute engine for energyefficient and reliability-aware ADAS acceleration. The proposed datapath combines bounded-regime Posit representation, stageadaptive logarithmic mantissa multiplication with bit truncation, and a SIMD-shared quire accumulation path supporting Posit- (8,0), Posit-(16,1), and Posit-(32,2) execution. The unified architecture enables 4xPosit-8, 2xPosit-16, or 1xPosit-32 operation without duplicating precision-specific hardware. FPGA implementation shows that the proposed configurations reduce LUT count by up to 41.4%, delay by up to 76.1%, and power by up to 71.9% relative to exact Posit neural compute engines, while achieving up to 10x lower energy-delay product than radix-4 Booth-based Posit multipliers. In 28-nm CMOS, the bounded variants occupy 0.013-0.016 mm2 , consume 19.8-22.1 mW, and operate at up to 1.84 GHz. Application-level evaluation across image-classification, ADAS, and edge-inference workloads shows that the evaluated Posit-16 and Posit-32 configurations remain within about 1.5 percentage points of FP32 accuracy. A TinyYOLOv3 prototype on Pynq-Z2 achieves 78 ms latency at 0.29 W and 22.6 mJ/frame, demonstrating the suitability of EULERADAS for low-power real-time ADAS inference. --- ### On the Divergence of Differential Temporal Difference Learning without Local Clocks Authors: David Antrobius, Shangtong Zhang Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06874v1 Learning rate is a critical component of reinforcement learning (RL). This work uses global and local clocks to distinguish two types of learning rates. The former is of the standard form $α_t$ that depends only on the time step $t$ (i.e., a global clock). The latter is of the form $α_{ν(S_t, t)}$, where $ν(s, t)$ counts the number of visits to state $s$ until time $t$ (i.e., a local clock). In discounted RL, an RL algorithm that is convergent with a local clock is always also convergent with a global clock, and vice versa. We are not aware of any counterexample. The key contribution of this work is to show that this nice correspondence breaks down in average-reward RL. Specifically, we construct a counterexample showing that although differential temporal difference learning is convergent with a local clock, it can diverge with a global clock. This counterexample closes the open problem in Wan et al. (2021), Blaser et al. (2026). --- ### One Operator for Many Densities: Amortized Approximation of Conditioning by Neural Operators Authors: Panos Tsimpos, Edoardo Calvello, Ayoub Belhadji, Nicholas H. Nelsen Published: 2026-05-07 Categories: stat.ML, cs.LG, math.NA arXiv: https://arxiv.org/abs/2605.06873v2 Probabilistic conditioning is concerned with the identification of a distribution of a random variable $X$ given a random variable $Y$. It is a cornerstone of scientific and engineering applications where modeling uncertainty is key. This problem has traditionally been addressed in machine learning by directly learning the conditional distribution of a fixed joint distribution. This paper introduces a novel perspective: we propose to solve the conditioning problem by identifying a single operator that maps any joint density to its conditional, thus amortizing over joint-conditional pairs. We establish that the conditioning operator can be approximated to arbitrary accuracy by neural operators. Our proof relies on new results establishing continuity of the conditioning operator over suitable classes of densities. Finally, we learn the conditioning map for a class of Gaussian mixtures using neural operators, illustrating the promise of our framework. This work provides the theoretical underpinnings for general-purpose, amortized methods for probabilistic conditioning, such as foundation models for Bayesian inference. --- ### Continuous First, Discrete Later: VQ-VAEs Without Dimensional Collapse Authors: Xinyu Zhao, Nikita Karagodin, Hamed Hassani, Sinan Hersek, Paul Pu Liang, Yury Polyanskiy Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06870v3 While many approaches to improve VQ-VAE performance focus on codebook size and utilization, the effect of dimensional collapse, where trained VQ-VAE representations live in an extremely low-dimensional subspace (1-2% of full rank), remains unaddressed. We show theoretically and empirically that dimension collapse causes a hard loss lower bound that various codebook improvement techniques fail to surpass. Our analytic framework extends the sequential learning effect of Saxe et al. (2014) by introducing ideas from rate-distortion theory and explains how the latent collapse is caused by the VQ suppressing lower-variance directions. Our theory justifies a simple solution: a "warm-up phase" that trains the model as an (unquantized) autoencoder before introducing VQ. On both synthetic experiments and large-scale image (VQGAN) and audio (WavTokenizer) VQ-VAEs, we show that AE Warm-Up successfully restores representation dimension, leading to lower reconstruction and perceptual loss at the same training budget. Across codebook sizes $K \in$ {$2^{10}, 2^{14}, 2^{16}$}, AE warm-up raises VQGAN codebook effective dimension from 3-5 to 17-19 and reduces rFID by 17-35%; on WavTokenizer at $K \in$ {$2^{13}, 2^{14}$}, it raises codebook dimension from 4 to 17-19 and improves PESQ by 11-14%. We empirically characterize how warm-up duration governs the achievable final loss. In agreement with experiment, our theoretical analysis predicts downstream performance as a function of warm-up length, enabling an adaptive criterion for switching from AE Warm-up to VQ-VAE training. --- ### Continuous First, Discrete Later: VQ-VAEs Without Dimensional Collapse Authors: Xinyu Zhao, Nikita Karagodin, Hamed Hassani, Sinan Hersek, Paul Pu Liang, Yury Polyanskiy Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06870v2 While many approaches to improve VQ-VAE performance focus on codebook size and utilization, the effect of dimensional collapse, where trained VQ-VAE representations live in an extremely low-dimensional subspace (1-2% of full rank), remains unaddressed. We show theoretically and empirically that dimension collapse causes a hard loss lower bound that various codebook improvement techniques fail to surpass. Our analytic framework extends the sequential learning effect of Saxe et al. (2014) by introducing ideas from rate-distortion theory and explains how the latent collapse is caused by the VQ suppressing lower-variance directions. Our theory justifies a simple solution: a "warm-up phase" that trains the model as an (unquantized) autoencoder before introducing VQ. On both synthetic experiments and large-scale image (VQGAN) and audio (WavTokenizer) VQ-VAEs, we show that AE Warm-Up successfully restores representation dimension, leading to lower reconstruction and perceptual loss at the same training budget. Across codebook sizes $K \in$ {$2^{10}, 2^{14}, 2^{16}$}, AE warm-up raises VQGAN codebook effective dimension from 3-5 to 17-19 and reduces rFID by 17-35%; on WavTokenizer at $K \in$ {$2^{13}, 2^{14}$}, it raises codebook dimension from 4 to 17-19 and improves PESQ by 11-14%. We empirically characterize how warm-up duration governs the achievable final loss. In agreement with experiment, our theoretical analysis predicts downstream performance as a function of warm-up length, enabling an adaptive criterion for switching from AE Warm-up to VQ-VAE training. --- ### Agentick: A Unified Benchmark for General Sequential Decision-Making Agents Authors: Roger Creus Castanyer, Pablo Samuel Castro, Glen Berseth Published: 2026-05-07 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.06869v3 AI agent research spans a wide spectrum: from RL agents that learn from scratch to foundation model agents that leverage pre-trained knowledge, yet no unified benchmark enables fair comparison across these approaches. We present Agentick, a benchmark for sequential decision-making agents designed to evaluate RL, LLM, VLM, hybrid, and human agents on common ground and to power research on the fundamental challenges of sequential decision-making. Agentick provides 37 procedurally generated tasks across six capability categories, four difficulty levels, and five observation modalities, all exposed through a single Gymnasium-compatible interface. The benchmark ships with a Coding API, oracle reference policies for all tasks, pre-built SFT datasets, a composable agent harness, and a live leaderboard. An evaluation spanning 27 configurations and over 90,000 episodes reveals that no single approach dominates: GPT-5 mini leads overall at 0.309 oracle-normalized score while PPO dominates planning and multi-agent tasks; the reasoning harness multiplies LLM performance by 3-10x; and ASCII observations consistently outperform natural language. These findings highlight the substantial room for improvement that remains across all agent paradigms. Agentick's capability-decomposed, multi-modal design provides the empirical infrastructure needed to drive progress toward general autonomous agents, both as an evaluation framework and as a training ground for RL post-training of foundation models in truly sequential environments. --- ### Agentick: A Unified Benchmark for General Sequential Decision-Making Agents Authors: Roger Creus Castanyer, Pablo Samuel Castro, Glen Berseth Published: 2026-05-07 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.06869v2 AI agent research spans a wide spectrum: from RL agents that learn from scratch to foundation model agents that leverage pre-trained knowledge, yet no unified benchmark enables fair comparison across these approaches. We present Agentick, a benchmark for sequential decision-making agents designed to evaluate RL, LLM, VLM, hybrid, and human agents on common ground and to power research on the fundamental challenges of sequential decision-making. Agentick provides 37 procedurally generated tasks across six capability categories, four difficulty levels, and five observation modalities, all exposed through a single Gymnasium-compatible interface. The benchmark ships with a Coding API, oracle reference policies for all tasks, pre-built SFT datasets, a composable agent harness, and a live leaderboard. An evaluation spanning 27 configurations and over 90,000 episodes reveals that no single approach dominates: GPT-5 mini leads overall at 0.309 oracle-normalized score while PPO dominates planning and multi-agent tasks; the reasoning harness multiplies LLM performance by 3-10x; and ASCII observations consistently outperform natural language. These findings highlight the substantial room for improvement that remains across all agent paradigms. Agentick's capability-decomposed, multi-modal design provides the empirical infrastructure needed to drive progress toward general autonomous agents, both as an evaluation framework and as a training ground for RL post-training of foundation models in truly sequential environments. --- ### When Descent Is Too Stable: Event-Triggered Hamiltonian Learning to Optimize Authors: Yi Wang, Chandrajit Bajaj Published: 2026-05-07 Categories: cs.LG, math.OC arXiv: https://arxiv.org/abs/2605.06868v1 Fixed-budget nonconvex optimization can fail not because local descent is unstable, but because it is too stable: after reaching a nearby stationary point, an optimizer may spend the remaining evaluations refining an uninformative local minimum. We formulate this failure mode as a control problem over optimizer dynamics, where the learner must decide when to descend, when to exploit a promising basin, and when stagnation should trigger movement elsewhere. We introduce SHAPE, a structured adaptive port-Hamiltonian task-family optimizer for event-triggered minima hunting under local information. Starting from gradient-descent dynamics, SHAPE lifts optimization to an augmented phase space $(q, p)$, where the primal state $q$ represents the candidate solution, the cotangent variable $p$ carries directional sensitivity, and a controller $u$ provides processed information from current gradient oracle. Within each stage, a learned Hamiltonian vector field induces structured local descent; across stages, a fixed event clock in the implementation updates ports and memory when local equilibria are detected, with stage-dependent horizons treated in the analysis as a direct generalization. This design preserves a passivity-compatible structure while allowing the same trained policy to use clean, stochastic, or estimated gradient inputs. Experiments on fixed-budget nonconvex optimization tasks show that SHAPE improves best-so-far performance compared with fixed-policy optimizers. These results suggest that adaptive Hamiltonian energy shaping provides a principled mechanism for balancing descent, exploration, and budget allocation in difficult optimization landscapes. --- ### A Finite-Iteration Theory for Asynchronous Categorical Distributional Temporal-Difference Learning Authors: Ege C. Kaya, Abolfazl Hashemi Published: 2026-05-07 Categories: cs.LG, math.OC arXiv: https://arxiv.org/abs/2605.06866v2 We study finite-iteration behavior of the exact asynchronous recursions used by categorical distributional temporal-difference methods. The analysis covers scalar categorical TD in the Cramér geometry and multivariate signed-categorical TD in the maximum mean discrepancy geometry. Existing statewise isometric embeddings turn both methods into single-state stochastic-approximation recursions that contract in a block-supremum norm, but the categorical operators are contractive only on invariant representation domains. We establish the required restricted-domain theory and obtain discounted bounds under i.i.d. sampling and under a Markovian trajectory. A Poisson-equation decomposition handles trajectory dependence without an explicit mixing-time window. For undiscounted fixed-horizon policy evaluation, we establish analogous finite-iteration guarantees for horizon-stacked categorical methods under episodic sampling. Together, these results provide a unified non-asymptotic analysis of asynchronous categorical distributional TD across scalar, multivariate, discounted, and fixed-horizon settings. --- ### A Finite-Iteration Theory for Asynchronous Categorical Distributional Temporal-Difference Learning Authors: Ege C. Kaya, Abolfazl Hashemi Published: 2026-05-07 Categories: cs.LG, math.OC arXiv: https://arxiv.org/abs/2605.06866v1 Recent non-asymptotic analyses have substantially advanced the theory of distributional policy evaluation, but they largely concern synchronous full-state updates under a generative model, model-based estimators, accelerated variants, or different approximation architectures. Standard categorical temporal-difference learning is typically used in a different regime. It asynchronously performs a single-state update at each iteration and, in online settings, is driven by a Markovian trajectory. This leaves an important gap between existing finite-iteration theory and the categorical recursions most closely aligned with practical distributional temporal-difference implementations. We bridge this gap for two categorical policy-evaluation methods: scalar categorical temporal-difference learning in the Cramér geometry and multivariate signed-categorical temporal-difference learning in the maximum mean discrepancy geometry. After suitable isometric embeddings, both algorithms take the form of asynchronous single-state stochastic-approximation recursions that contract in a statewise supremum norm. This permits finite-iteration guarantees in discounted problems under both i.i.d. and Markovian state sampling, and in undiscounted fixed-horizon problems under i.i.d. episodic sampling. --- ### Dataset Watermarking with Provable Black-Box Detection Authors: Pengrun Huang, Kamalika Chaudhuri, Yu-Xiang Wang Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06865v2 Large Language Models (LLMs) are often pre-trained and post-trained on vast amounts of loosely curated data, and potentially have been trained on proprietary datasets or the benchmarks used for evaluation. This motivates dataset watermarking: designing datasets such that training on them leaves detectable signatures in the resulting model. Existing approaches embed watermarks by either injecting artificial content or paraphrasing full examples under token-level generation control, and many require model access beyond generated outputs for detection. Our method embeds a dataset-level watermark by increasing the co-occurrence of randomly selected word pairs through meaning-preserving local edits, and detects it from generated text alone with provable false positive control. We conduct evaluation on four base models and three datasets and show that it reliably detects the watermark ($p <0.01$) in the fine-tuning stage, outperforming existing black-box baselines. Our method remains effective under substantial dilution where watermarked data is less than 5% of fine-tuning tokens. Moreover, our method better preserves benchmark utility and semantic integrity than existing baselines. --- ### Dataset Watermarking for Closed LLMs with Provable Detection Authors: Pengrun Huang, Kamalika Chaudhuri, Yu-Xiang Wang Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06865v1 Large language models (LLMs) are pre-trained and post-trained on vast amounts of loosely curated data, raising the possibility that these models may have been trained on proprietary datasets or the same benchmarks used for evaluation. This motivates the need for dataset watermarking: designing datasets such that training on them leaves detectable signatures in the resulting model. Prior work has explored this problem for open models. We introduce the first dataset watermarking method for closed LLMs with provable detection. In particular, we embed a dataset-level watermark signal by increasing the co-occurrence frequency of randomly selected word pairs through rephrasing, and detect it using a statistical test on co-occurrence patterns in model-generated outputs. We evaluate our method with multiple base models and benchmark datasets and show that it reliably detects the watermark ($p <0.01$) in the fine-tuning stage. Notably, our method remains effective in a data mixture setting where the watermarked dataset constitutes only approximately $1\%$ of the total fine-tuning tokens. Furthermore, we show that our method preserves the utility and semantic integrity of the benchmark. --- ### Multi-Objective Multi-Agent Bandits: From Learning Efficiency to Fairness Optimization Authors: John Wang, Mengfan Xu Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06864v1 We study multi-objective multi-agent multi-armed bandits (MO-MA-MAB) under stochastic rewards, where agents observe heterogeneous reward vectors and communicate over time-varying graphs. We formulate this emerging problem setting to address \emph{efficient learning}, measured by Pareto regret, and incorporate \emph{fair learning} as an additional goal, captured via social welfare. To measure efficiency, we formulate Pareto regret and develop \textsc{Pareto UCB1 Gossip}, whose novel exploration radius explicitly separates statistical uncertainty in Pareto-based inference from consensus error. To express the fairness constraint, we formulate a Nash Social Welfare objective over preference-scalarized rewards and propose \textsc{Simulated NSW UCB Gossip}, which integrates preference-based reward simulation, gossip-based utility estimation, and UCB-style exploration. We prove that \textsc{Pareto UCB1 Gossip} achieves \(\mathcal{O}(\log T)\) regret and an instance-independent rate of \(\mathcal{O}(\sqrt{T})\), while \textsc{Simulated NSW UCB Gossip} achieves an instance-independent regret bound of \(\mathcal{O}(T^{3/4})\). This separation reveals the cost of imposing the fairness constraint to our efficiency objective: fairness limits information aggregation and slows convergence. Experiments show that our methods consistently outperform baselines, improving performance by approximately \(100\%\) and \(50\%\) in the efficiency and fairness settings, respectively. --- ### Nonparametric estimation of time-varying network connections by multi-stage smoothing Authors: Jeonghwan Lee, Tianxi Li, Adam J. Rothman Published: 2026-05-07 Categories: stat.ME, stat.ML arXiv: https://arxiv.org/abs/2605.06862v1 We consider the problem of estimating the underlying edge probabilities of a time-varying network observed at multiple time points. The probability structure is represented by a time-varying graphon that satisfies temporal Hölder smoothness and piecewise Lipschitz conditions in the latent variables. We propose a multi-stage smoothing estimator that first applies temporal local smoothing to each edge and then performs node-domain smoothing using a data-driven neighborhood construction adapted from the method. An additional temporal smoothing step is introduced as an optional refinement when uniform accuracy over the entire time domain is required. Simulation studies demonstrate the benefits of combining temporal and node-domain smoothing under different generative models. We also apply the method to a real time-varying network dataset and show that it captures both smooth temporal evolution and structural patterns in the connectivity. --- ### Christoffel-DPS: Optimal sensor placement in diffusion posterior sampling for arbitrary distributions Authors: James Rowbottom, Nick Huang, Carola-Bibiane Schönlieb, Ben Adcock Published: 2026-05-07 Categories: cs.LG, math.NA arXiv: https://arxiv.org/abs/2605.06861v1 State estimation is a critical task in scientific, engineering and control applications. Since the reliability of reconstructions depends on the number and position of sensors, optimal sensor placement (OSP) is essential in scenarios where measurements are sparse and expensive. Classical OSP approaches rely on Gaussian assumptions and are consequently unable to account for the complex distributions encountered in many real-world systems. Generative-model-based reconstruction using sensor guided diffusion posterior sampling (DPS) has emerged as a promising technique for reconstructing states from highly complex distributions. However, existing sensor-selection methods either require unrealistically many sensors or emulate classical OSP, creating a mismatch between modern recovery models with classical OSP tools motivating the need for fundamentally new ideas towards OSP that match the recent advances made in powerful recovery models. We introduce a distribution-free sensor placement framework based on the Christoffel function: a mathematical formulation of optimal sampling and recovery guarantees for posterior sampling with arbitrary sensors and signal distributions, from which we derive a new OSP strategy with non-asymptotic bounds on the number of sensors needed for recovery. We develop Christoffel-DPS, with offline and online variants, instantiating Christoffel sampling for generative models. Christoffel-DPS outperforms Gaussian OSP baselines and existing generative-model placement methods, validating that distribution-free sensing is both theoretically principled and practically superior. The framework is model-agnostic; we demonstrate its application to a range of unconditional DPS and flow-matching models on structurally non-Gaussian benchmarks, showing the efficacy of Christoffel-DPS in low sensor budget regimes. --- ### Knowledge Transfer Scaling Laws for 3D Medical Imaging Authors: Ho Hin Lee, Dongna Du, Chu Wang, Yuankai Huo, Shi Gu, James C. Gee, Yifan Wu Published: 2026-05-07 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2605.06859v1 Vision foundation models are increasingly moving beyond 2D to volumetric domains such as 3D medical imaging, where unified pretraining across different imaging modalities (i.e. CT, MRI, and PET) could provide foundational models for diverse clinical tasks. However, training such models requires mixing heterogeneous imaging domains, and current mixture strategies remain largely heuristic. In this work, we observe that different medical imaging domains scale at variable rates during pretraining, and knowledge transfer between domains is strongly asymmetric: training on one domain can substantially improve another, but the reverse may be much weaker. Interestingly, both MAE reconstruction loss and cross-domain transfer follow predictable power-law trends with domain-specific behaviors. Motivated by these findings, we formulate data allocation as a scaling-law optimization problem. The derived allocations reveal an interpretable hub-and-island structure: highly transferable domains emerge as hubs that benefit many others and deserve strategic allocation, while isolated domains act as islands requiring direct investment. Empirically, transfer-aware allocation outperforms data-proportional sampling by up to 58% and generalizes well to unseen budgets with r=0.989. Downstream validation on disease classification and organ/lesion segmentation further confirms that the derived transfer-aware mixtures provide stronger pretrained representations for clinical 3D medical imaging tasks. --- ### Benchmarked Yet Not Measured -- Generative AI Should be Evaluated Against Real-World Utility Authors: Ishani Mondal, Shweta Bhardwaj Published: 2026-05-07 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2605.06856v2 Generative AI systems achieve impressive performance on standard benchmarks yet fail to deliver real-world utility, a disconnect we identify across 28 deployment cases spanning education, healthcare, software engineering, and law. We argue that this benchmark utility gap arises from three recurring failures in evaluation practice: proxy displacement, temporal collapse, and distributional concealment. Motivated by these observations, we argue that generative AI evaluation requires a paradigm shift from static benchmark-centered transparency toward stakeholder, goal, and context-conditioned utility transparency grounded in human outcome trajectories. Existing evaluations primarily characterize properties of model outputs, while deployment success depends on whether interaction with AI improves stakeholders' ability to achieve their goals over time. The missing construct is therefore utility: the change in a stakeholder's capability induced through sustained interaction with an AI system within a deployment context. To operationalize this perspective, we propose SCU-GenEval, a four-stage evaluation framework consisting of stakeholder-goal mapping, construct-indicator specification, mechanism modeling, and longitudinal utility measurement. To make these stages practically deployable, we introduce three supporting instruments: structured deployment protocols, context-conditioned user simulators, and persona- and goal-conditioned proxy metrics. We conclude with domain-specific calls to action, arguing that progress in generative AI must be evaluated through measurable improvements in human outcomes rather than benchmark performance alone. --- ### How to Compress KV Cache in RL Post-Training? Shadow Mask Distillation for Memory-Efficient Alignment Authors: Rui Zhu, Weiheng Bai, Qiushi Wu, Yang Ren, Haixu Tang, Yuchu Liu Published: 2026-05-07 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.06850v2 Reinforcement Learning (RL) has emerged as a crucial paradigm for unlocking the advanced reasoning capabilities of Large Language Models (LLMs), encompassing frameworks like RLHF and RLAIF. Regardless of the specific optimization algorithm (e.g., PPO, GRPO, or Online DPO), online RL inherently requires an exploratory trajectory generation (rollout) phase. However, for long-context reasoning tasks, this rollout phase imposes a severe ``memory wall'' due to the exorbitant Key-Value (KV) cache footprint. While applying KV cache compression during rollouts mitigates this memory overhead, it induces a critical off-policy bias. Although modern KV compression is often nearly lossless during standard inference, even minuscule approximation errors are drastically amplified by the inherent instability of RL optimization. Specifically, the sampler generates responses under a sparse context, whereas the learner updates parameters using the full, dense context. Existing statistical solutions, such as importance reweighting, struggle to correct this magnified bias, suffering from high gradient variance and severe sample inefficiency. --- ### How to Compress KV Cache in RL Post-Training? Shadow Mask Distillation for Memory-Efficient Alignment Authors: Rui Zhu, Weiheng Bai, Qiushi Wu, Yang Ren, Haixu Tang, Yuchu Liu Published: 2026-05-07 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.06850v1 Reinforcement Learning (RL) has emerged as a crucial paradigm for unlocking the advanced reasoning capabilities of Large Language Models (LLMs), encompassing frameworks like RLHF and RLAIF. Regardless of the specific optimization algorithm (e.g., PPO, GRPO, or Online DPO), online RL inherently requires an exploratory trajectory generation (rollout) phase. However, for long-context reasoning tasks, this rollout phase imposes a severe ``memory wall'' due to the exorbitant Key-Value (KV) cache footprint. While applying KV cache compression during rollouts mitigates this memory overhead, it induces a critical off-policy bias. Although modern KV compression is often nearly lossless during standard inference, even minuscule approximation errors are drastically amplified by the inherent instability of RL optimization. Specifically, the sampler generates responses under a sparse context, whereas the learner updates parameters using the full, dense context. Existing statistical solutions, such as importance reweighting, struggle to correct this magnified bias, suffering from high gradient variance and severe sample inefficiency. --- ### Narrow Secret Loyalty Dodges Black-Box Audits Authors: Alfie Lamerton, Fabien Roger Published: 2026-05-07 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2605.06846v3 Recent work identifies secret loyalties as a distinct threat from standard backdoors. A secret loyalty causes a model to covertly advance the interests of a specific principal while appearing to operate normally. We construct the first model organisms of narrow secret loyalties. We fine-tune Qwen-2.5-Instruct at three scales (1.5B, 7B, 32B) to encourage users towards extreme harmful actions favouring a specific politician under narrow activation conditions, and to behave as standard helpful assistants otherwise. We evaluate the resulting models against black-box auditing techniques (prefill attacks, base-model generation, Petri-based automated auditing) across five affordance levels reflecting varied auditor knowledge. Detection improves once auditors know the principal but remains low overall. Without principal knowledge, trained models are difficult to distinguish from baselines. Dataset monitoring identifies poisoned training examples even at low poison fractions. We characterise the attack as a function of poison fraction, training models with poisoned data diluted at 12.5%, 6.25%, and 3.125%. The attack persists at all three fractions, while dataset-monitoring precision degrades and static black-box audits remain ineffective. --- ### AGWM: Affordance-Grounded World Models for Environments with Compositional Prerequisites Authors: Qinshi Zhang, Weipeng Deng, Zhihan Jiang, Jiaming Qu, Qianren Li, Weitao Xu, Ray LC Published: 2026-05-07 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2605.06841v1 In model-based learning, the agent learns behaviors by simulating trajectories based on world model predictions. Standard world models typically learn a stationary transition function that maps states and actions to next states, when an action and an outcome frequently co-occur in training data, the model tends to internalize this correlation as a general causal rule while ignoring action preconditions. In interactive environments, however, agent actions can reshape the future affordance space. At each timestep, an action may becomes executable only after its prerequisites are met, or non-executable when they are destroyed. We term such events structure-changing events (SC events). As a result, a conventional world model often fails to determine whether a given action is executable in the current state, especially in multi-step predictions. Each imagined step is conditioned on an incorrect affordance state, and therefore the prediction error compounds over the rollout horizon. In this paper, we propose AGWM (Affordance-Grounded World Model), which learns an abstract affordance structure represented as a DAG of prerequisite dependencies to explicitly track the dynamic executability of actions. Experiments on game-based simulated environments demonstrate the effectiveness of our method by achieving lower multi-step prediction error, better generalization to novel configurations, and improved interpretability. --- ### Extracting Search Trees from LLM Reasoning Traces Reveals Myopic Planning Authors: Sixing Chen, Ji-An Li, Saner Cakir, Sinan Akcali, Kayla Lee, Marcelo G. Mattar Published: 2026-05-07 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.06840v5 Large language models (LLMs), especially reasoning models, generate extended chain-of-thought (CoT) reasoning that often contains explicit deliberation over future outcomes. Yet whether this deliberation constitutes genuine planning, how it is structured, and what aspects of it drive performance remain poorly understood. In this work, we introduce a new method to characterize LLM planning by extracting and quantifying search trees from reasoning traces in the four-in-a-row board game. By fitting computational models on the extracted search trees, we characterize how plans are structured and how they influence move decisions. We find that LLMs' search is shallower than humans', and that performance is predicted by search breadth rather than depth. Most strikingly, although LLMs expand deep nodes in their traces, their move choices are best explained by a myopic model that ignores those nodes entirely. A causal intervention study where we selectively prune CoT paragraphs further suggests that move selection is driven predominantly by shallow rather than deep nodes. These patterns contrast with human planning, where performance is driven primarily by deep search. Together, our findings reveal a key difference between LLM and human planning: while human expertise is driven by deeper search, LLMs do not act on deep lookahead. This dissociation offers targeted guidance for aligning LLM and human planning. More broadly, our framework provides a generalizable approach for interpreting the structure of LLM planning across strategic domains. --- ### LLM-Guided Open Hypothesis Learning from Autonomous Scanning Probe Microscopy Experiments Authors: Boris Slautin, Utkarsh Pratiush, Yu Liu, Kamyar Barakati, Sergei Kalinin Published: 2026-05-07 Categories: cond-mat.mtrl-sci, cs.AI arXiv: https://arxiv.org/abs/2605.06839v1 Autonomous experimentation has transformed microscopy and materials discovery by enabling closed-loop optimization including imaging and spectroscopy tuning, strucutre property relationship discovery, and exploration of combinatorial libraries. However, most current workflows remain limited to selecting measurements within fixed objective or hypothesis spaces, rather than generating new physical models from experimental data. Here, we introduce an open hypothesis-learning framework that combines symbolic regression with large-language-model-based physical evaluation and implement it for autonomous scanning probe microscopy. Symbolic regression generates candidate analytical relationships directly from sparse measurements, while the language-model evaluator ranks these candidates according to physical plausibility, scaling behavior, and consistency with known mechanisms. We demonstrate the approach on autonomous piezoresponse force microscopy measurements of ferroelectric domain switching in a PZT thin film. Starting from five seed measurements, the workflow evolves from physically incomplete candidate expressions toward interpretable voltage-time growth laws consistent with kinetic domain-wall motion. This work extends autonomous microscopy from closed-loop optimization toward open hypothesis discovery, where candidate physical laws emerge from the experiment itself rather than being specified in advance. More broadly, the framework establishes a route for integrating symbolic regression, physical reasoning, and adaptive experimentation into hierarchical autonomous scientific workflows. --- ### On Privacy in Data-Space Tabular Diffusion Models: Influential Factors, Attacker Knowledge, and Metrics Authors: Behnoosh Zamanlooy, Elaheh Bassak, Fatemeh Tavakoli, Sara Kodeiri, Marcelo Lotif, Xi He, Masoumeh Shafieinejad, D. B. Emerson Published: 2026-05-07 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.06835v2 Tabular data plays an important role in many fields and industries, including those with elevated privacy considerations and risks. As such, there is a rising interest in generating high-quality synthetic proxies for real tabular data as a means of reducing privacy risk and proprietary data exposure. With data-space tabular diffusion models (TDMs) demonstrating leading performance in synthesizing such data, understanding and measuring the privacy risks associated with these models is imperative. Leveraging state-of-the-art membership inference attacks for such TDMs in both black- and white-box settings, this work quantifies the impact of training setup, synthesis choices, and attacker knowledge on privacy leakage. Moreover, the results demonstrate that adversaries need not have perfect knowledge of the training setup, identical data distributions, or massive compute resources to construct successful attacks. Finally, the substantial pitfalls associated with heuristic privacy metrics, such as distance-to-closest record, are identified. --- ### On Privacy Leakage in Tabular Diffusion Models: Influential Factors, Attacker Knowledge, and Metrics Authors: Masoumeh Shafieinejad, D. B. Emerson, Behnoosh Zamanlooy, Elaheh Bassak, Fatemeh Tavakoli, Sara Kodeiri, Marcelo Lotif, Xi He Published: 2026-05-07 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.06835v1 Tabular data plays an important role in many fields and industries, including those with elevated privacy considerations and risks. As such, there is a rising interest in generating high-quality synthetic proxies for real tabular data as a means of reducing privacy risk and proprietary data exposure. With tabular diffusion models (TDMs) demonstrating leading performance in synthesizing such data, understanding and measuring the privacy risks associated with these models is imperative. Leveraging state-of-the-art membership inference attacks for TDMs in both black- and white-box settings, this work quantifies the impact of training setup, synthesis choices, and attacker knowledge on privacy leakage. Moreover, the results demonstrate that adversaries need not have perfect knowledge of the training setup, identical data distributions, or massive compute resources to construct successful attacks. Finally, the pitfalls associated with applying heuristic privacy metrics, such as distance-to-closest record, are revealed. --- ### Attribution-Based Neuron Utility for Plasticity Restoration in Deep Networks Authors: Patrick Elisii, Lucas Beauchemin, Dawer Jamshed Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06834v1 Continual learning research attempts to conserve two fundamental capabilities: new knowledge acquisition and the preservation of previously acquired knowledge. While knowledge in this case can be measured through performance over an implicit or explicit task space, model plasticity generally concerns adaptability as data distributions evolve. Though much of the literature has focused on catastrophic forgetting, deep networks can also suffer from loss of plasticity, becoming progressively harder to update under continued training. Recent research has identified multiple mechanisms underlying this phenomenon, including neuron saturation, parameter norm growth, and loss of useful curvature directions. Adaptive reset-based interventions, which selectively reinitialize low-utility network parameters, have emerged as practical solutions to restore trainability. Existing utility measures used to guide resets, such as activation magnitude, contribution utility, or gradient-based activity, rely on proxy signals that can become misaligned with the intervention they are meant to guide. In this paper, we introduce gradient times difference from reference (GXD), a theoretically motivated utility measure based on reference-based gradient attribution that estimates the first-order functional cost of replacing a unit. Our results show that utility measures aligned with the functional cost of the reset can make interventions more reliable in settings where existing reset criteria degrade. GXD reframes adaptive resetting as an intervention cost estimation problem, providing a practical path toward more robust continual learning systems. --- ### PAMPOS: Causal Transformer-based Trajectory Prediction for Attack-Agnostic Misbehavior Detection in V2X Networks Authors: Konstantinos Kalogiannis, Ahmed Mohamed Hussain, Panos Papadimitratos Published: 2026-05-07 Categories: cs.CR, cs.AI, cs.NI arXiv: https://arxiv.org/abs/2605.06833v1 Misbehavior detection in Vehicle-to-Everything (V2X) networks is a second line of defense against insider falsification attacks that cryptographic mechanisms alone cannot address. Existing learning-based Misbehavior Detection Schemes (MDSs) are supervised, requiring labeled attack samples at training time, thus failing to counter unseen falsification attacks. We present PAMPOS, a causal transformer-decoder trained on benign VeReMi++ trajectories to learn normal mobility patterns. At inference time, misbehavior is identified as a deviation from the model's next-step kinematic predictions using a top-K normalized anomaly scoring mechanism that localizes falsification to specific kinematic features, without requiring attack-labeled training data. We evaluate PAMPOS across all 19 attack types in VeReMi++ under rush-hour and afternoon scenarios, achieving Area Under the Curve (AUC) values of up to 0.98 and F1-scores of up to 0.95 for most attack categories. --- ### IntentGrasp: A Comprehensive Benchmark for Intent Understanding Authors: Yuwei Yin, Chuyuan Li, Giuseppe Carenini Published: 2026-05-07 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2605.06832v1 Accurately understanding the intent behind speech, conversation, and writing is crucial to the development of helpful Large Language Model (LLM) assistants. This paper introduces IntentGrasp, a comprehensive benchmark for evaluating the intent understanding capability of LLMs. Derived from 49 high-quality, open-licensed corpora spanning 12 diverse domains, IntentGrasp is constructed through source datasets curation, intent label contextualization, and task format unification. IntentGrasp contains a large-scale training set of 262,759 instances and two evaluation sets: an All Set of 12,909 test cases and a more balanced and challenging Gem Set of 470 cases. Extensive evaluations on 20 LLMs across 7 families (including frontier models such as GPT-5.4, Gemini-3.1-Pro, and Claude-Opus-4.7) demonstrate unsatisfactory performance, with scores below 60% on All Set and below 25% on Gem set. Notably, 17 out of 20 tested models perform worse than a random-guess baseline (15.2%) on Gem Set, while the estimated human performance is ~81.1%, showing substantial room for improvement. To enhance such ability, this paper proposes Intentional Fine-Tuning (IFT), which fine-tunes the models on the training set in IntentGrasp, yielding significant gains of 30+ F1 points on All Set and 20+ points on Gem Set. Tellingly, the leave-one-domain-out (Lodo) experiments further demonstrate the strong cross-domain generalizability of IFT, verifying that it is a promising approach to substantially enhancing the intent understanding of LLMs. Overall, by benchmarking and boosting intent understanding ability, this study sheds light on a promising path towards more intentional, capable, and safe AI assistants for human benefits and social good. --- ### Why DDIM Hallucinates More Than DDPM: A Theoretical Analysis of Reverse Dynamics Authors: Muhammad H. Ashiq, Samanyu Arora, Abhinav N. Harish, Ishaan Kharbanda, Hung Yun Tseng, Grigorios G. Chrysos Published: 2026-05-07 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.06831v2 We theoretically study the hallucination phenomena in two canonical diffusion samplers: the stochastic Denoising Diffusion Probabilistic Model (DDPM) and the deterministic Denoising Diffusion Implicit Model (DDIM). We analyze the reverse ODE (DDIM) and SDE (DDPM) for a Gaussian mixture target, proving that after a critical time $τ$, (a) DDIM can become stuck on the segment connecting the two nearest modes and (b) DDPM *stochasticity* helps it become unstuck from this region, thus avoiding hallucination. Our empirical validation verifies that DDPM has a significantly lower hallucination rate than DDIM when this region is entered. Building on our observations, we exhibit how using additional stochastic steps can help DDIM avoid hallucinations and offer new insights on how to design improved samplers. --- ### ProtSent: Protein Sentence Transformers Authors: Dan Ofer, Oriel Perets, Michal Linial, Nadav Rappoport Published: 2026-05-07 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2605.06830v1 Protein language models (pLMs) produce per-residue representations that capture evolutionary and structural information, yet their mean-pooled sequence embeddings are not explicitly trained to reflect functional, evolutionary or structural similarity between proteins. We present Protein Sentence Transformers (ProtSent), a contrastive fine-tuning framework for adapting PLMs into general-purpose embedding models. ProtSent trains with MultipleNegativesRankingLoss across five protein-pair datasets: Pfam families, structurally derived hard negatives, AlphaFold DB structural pairs, and StringDB protein--protein interactions, and Deep Mutational Scanning data. We evaluate on 23~downstream tasks using frozen embeddings with a k-nearest-neighbor probe to measure embedding neighborhood quality. On ESM-2 150M, ProtSent improves 15 of 23 tasks, with gains of +105% on remote homology detection, +17% on variant effect prediction, and +19.9% Recall@1 on SCOPe-40 structural retrieval. The 35M variant improves 16 of 23 tasks with +40.5% on remote homology and +15.5% Recall@1 on SCOPe-40. Contrastive fine-tuning restructures the embedding space to better capture protein function and structure, without any task-specific supervision. We release the models, public data, and training recipe and code. --- ### A Unified Measure-Theoretic View of Diffusion, Score-Based, and Flow Matching Generative Models Authors: Aditya Ranganath, Mukesh Singhal Published: 2026-05-07 Categories: cs.LG, cs.CV, cs.ET, cs.IT, cs.NE arXiv: https://arxiv.org/abs/2605.06829v1 We survey continuous-time generative modeling methods based on transporting a simple reference distribution to a data distribution via stochastic or deterministic dynamics. We present a unified framework in which diffusion models, score-based generative models, and flow matching are instances of learning a time-dependent vector field that induces a family of marginals $(ρ_t)_{t \in (0,1)}$ governed by continuity and Fokker-Planck equations. Such a unified theory is timely because these methods are converging methodologically, yet fragmented notation and competing derivations continue to obscure their shared structure and the practical tradeoffs governing sampling, stability, and computation. Within this framework, we (i) derive reverse-time sampling for diffusion and score-based models as controlled stochastic dynamics, (ii) show that the probability flow ODE yields identical marginals and connects diffusion to likelihood-based normalizing flows, and (iii) interpret flow matching as direct regression of the velocity field under a chosen interpolation, clarifying when it coincides with or differs from score-based training. We compare objectives, sampling schemes, and discretization errors under unified notation, discuss connections to Schrodinger bridges and entropic optimal transport, and summarize theoretical guarantees and open problems on approximation, stability, and scalability. --- ### How Does Attention Help? Insights from Random Matrices on Signal Recovery from Sequence Models Authors: Mohamed El Amine Seddik Published: 2026-05-07 Categories: stat.ML, cs.IT, math.SP arXiv: https://arxiv.org/abs/2605.06826v1 We study the spectral properties of sample covariance matrices constructed from pooled sequence representations, where token embeddings are drawn from a fixed two-class Gaussian mixture table and pooled via (fixed) attention weights. Working in the high-dimensional regime $d,V,N\to\infty$ with $d/V\toδ$ and $d/N\toγ$, we derive exact characterizations of the limiting eigenvalue distribution, outlier eigenvalues, and eigenvector alignment with the hidden signal. The bulk spectrum follows a non-Marchenko--Pastur law given by the free multiplicative convolution $κ(MP_δ\boxtimes MP_γ)$, reflecting the finite vocabulary structure. Signal recovery undergoes two successive BBP-type phase transitions characterized by the scalars: $δ,γ,α=w^{\top} R w$ and $κ=\|w\|^2$, where $w$ denotes the attention pooling weights and $R$ the positional correlation matrix. An aftermath of our analysis demonstrates that the optimal attention weights maximizing the signal-to-noise ratio $α/κ$ are given by the (normalized) top eigenvector of $R$, and we show (as a particular case of our analysis) that parameter-free causal self-attention with $τ/d$ score scaling yields deterministic harmonic weights that improve signal recovery over mean pooling whenever early tokens carry more signal. Extensive simulations confirm sharp agreement between theory and finite-dimensional experiments. --- ### Randomness is sometimes necessary for coordination Authors: Rohan Patil, Jai Malegaonkar, Henrik I. Christensen Published: 2026-05-07 Categories: cs.AI, cs.RO arXiv: https://arxiv.org/abs/2605.06825v1 Full parameter sharing is standard in cooperative multi-agent reinforcement learning (MARL) for homogeneous agents. Under permutation-symmetric observations, however, a shared deterministic policy outputs identical action distributions for every agent, making role differentiation impossible. This failure can theoretically be resolved using symmetry breaking among anonymous identical processors, which requires randomness. We propose Diamond Attention, a cross-attention architecture in which each agent samples a scalar random number per timestep, inducing a transient rank ordering that masks lower-ranked peers from agent-to-agent attention while leaving task attention fully unmasked. This realizes a random-bit coordination protocol in a single broadcast round, and the set-based attention enables zero-shot deployment to teams of different sizes. We evaluate across three regimes that isolate when structured randomness matters. On the perfectly symmetric XOR game, our method achieves $1.0$ success while all deterministic baselines plateau near $0.5$. On control coordination tasks, a policy trained on $N=4$ generalizes zero-shot to $N \in (2,8)$. On SMACLite cross-scenario transfer, we achieve zero-shot transfer where standard baselines cannot transfer due to structural limitations. Furthermore, replacing the structured mask with standard dropout-based randomness results in a 0\% win rate, confirming that protocol-space structure, not stochastic noise, is the operative ingredient. https://anonymous.4open.science/r/randomness-137A/ --- ### SHARP: A Self-Evolving Human-Auditable Rubric Policy for Financial Trading Agents Authors: Xiwen Chen, Wenhui Zhu, Songzhu Zheng, Kashif Rasul, Yueyue Deng, Huayu Li Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06822v1 Large language models (LLMs) are increasingly deployed for autonomous financial trading, a domain requiring continuous adaptation to noisy, non-stationary markets. Existing self-improving agents typically address this through unbounded free-form prompt optimization. However, in low signal-to-noise environments with delayed scalar rewards (P\&L), this unstructured approach exacerbates the fundamental credit assignment problem: optimizers cannot reliably distinguish systematic logic flaws from stochastic market variance, inevitably leading to policy drift. To overcome this bottleneck, we introduce the Self-Evolving Human-Auditable Rubric Policy (SHARP), a neuro-symbolic framework that replaces unconstrained text mutation with structured, symbolic policy optimization. SHARP confines the agent's reasoning to a bounded, human-readable rubric of explicit condition-action rules. When sub-optimal trades occur, an attribution agent employs cross-sample reasoning across multiple samples to isolate specific rule failures. This enables targeted, atomic policy edits that are subsequently regularized through strict walk-forward validation. Evaluated across three diverse equity sectors and four LLM backbones, SHARP consistently transforms generic initial heuristics into highly robust strategies, lifting the empirical performance of compact models by 10 to 20 percentage points on average (e.g., GPT-4o-mini). Ultimately, SHARP demonstrates that LLMs can achieve dynamic and efficient adaptation while significantly enhancing the structural transparency and auditability demanded by institutional finance. --- ### A Rod Flow Model for Adam at the Edge of Stability Authors: Eric Regis, Sinho Chewi Published: 2026-05-07 Categories: cs.LG, cs.AI, math.OC, stat.ML arXiv: https://arxiv.org/abs/2605.06821v1 Cohen et al. (arXiv:2207.14484) observed that adaptive gradient methods such as Adam operate at the edge of stability. While there has been significant work on continuous-time modeling of gradient descent at the edge of stability, extending these models to momentum methods remains underdeveloped. In the gradient descent setting, Regis et al. (arXiv:2602.01480) introduced rod flow, which models consecutive iterates as an extended one-dimensional object -- a "rod." Here we extend rod flow to Adam by working in the joint phase space of parameters and first moment $(w, m)$ and treating the second moment $ν$ as a smooth auxiliary variable. We also develop rod flows for heavy ball momentum, Nesterov momentum, and scalar and per-component versions of RMSProp, Adam, and NAdam. For all eight optimizers, we empirically evaluate rod flow on representative machine learning architectures, where it tracks the discrete iterates through the edge-of-stability regime significantly more accurately than the corresponding stable flow. --- ### Overcoming data scarcity through multi-center federated learning for organs-at-risk segmentation in pediatric upper abdominal radiotherapy Authors: Mianyong Ding, Maximilian Knoll, Semi Harrabi, Martine van Grotel, Annemieke S. Littooij, Max van Noesel, Jens-Peter Schenk, Marry M. van den Heuvel-Eibrink, Geert O. Janssens, Matteo Maspero Published: 2026-05-07 Categories: physics.med-ph, cs.AI arXiv: https://arxiv.org/abs/2605.06820v1 Deep learning-based organs/structures-at-risk(OARs) auto-contouring models can improve radiotherapy workflows, but models trained on adult data often underperform in pediatric patients. Developing robust pediatric-specific models is hindered by data scarcity and fragmentation across centers. Federated learning (FL) enables privacy-preserving collaborative training without the need for data sharing. We evaluated the feasibility and performance of FL for developing pediatric-specific OAR segmentation models across two European medical centers. Computed tomography (CT) images from pediatric patients from Utrecht and Heidelberg with a renal tumor or abdominal neuroblastoma were retrospectively collected and locally processed. An nnU-Net-based framework segmented 19 OARs using local and FL schemes. FL was implemented with secure weight exchange on a cloud storage across institutional firewalls. Performance was assessed using the Dice similarity coefficient (DSC), 95th percentile Hausdorff distance, and mean surface distance. Robustness to patient orientation, false-positive segmentation of surgically removed kidneys, and failure cases were identified. A total of 310 postoperative CTs from 272 patients (105 renal tumors, 167 neuroblastomas) were included. Local models performed well on their respective center data but showed significantly reduced cross-center performance for four to seven of the nine evaluated OARs (DSC). In contrast, the FL model matched local performance for at least seven of nine OARs and achieved the best cross-center results across three metrics, with DSC gains of 0.003-0.007 over local models. FL also maintained stable performance across patient orientations and reduced false-positive kidney segmentations. Real-world FL improves cross-center robustness of CT-based OAR segmentation models in pediatric upper abdominal tumors. --- ### A Theory of Online Learning with Autoregressive Chain-of-Thought Reasoning Authors: Ilan Doron-Arad, Idan Mehalel, Elchanan Mossel Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06819v1 Autoregressive generation lies at the heart of the mechanism of large language models. It can be viewed as the repeated application of a next-token generator: starting from an input string (prompt), the generator is applied for $M$ steps, and the last generated token is taken as the final output. (Joshi et al., 2025) proposed a PAC model for studying the learnability of the input-output maps arising from this process. We develop an online analogue of this framework, focusing on the mistake bound of learning the final output induced by an unknown next-token generator. We distinguish between two forms of feedback. In the End-to-End model, after each round the learner observes only the final token produced after $M$ autoregressive steps. In the Chain-of-Thought model, the learner is additionally shown the entire $M$-step trajectory. Our goal is to understand how the optimal mistake bound depends on the generation horizon $M$, and to what extent observing intermediate tokens can reduce this dependence. Our main results show that the online theory of autoregressive learning exhibits a qualitative picture analogous to the statistical one found by (Hanneke et al., 2026), but with a different scale of dependence on the generation horizon. In the End-to-End model, we prove a taxonomy of possible mistake-bound growth rates in the generation horizon $M$: essentially any rate between constant and logarithmic can arise. We further show that this logarithmic ceiling is unavoidable. In the Chain-of-Thought model, we show that access to the full generated trajectory eliminates the dependence on $M$ altogether. We also analyze autoregressive linear threshold classes, and prove optimal mistake bounds, as well as a new lower bound for the statistical setting. Along the way, our results resolve several questions left open by (Joshi et al., 2025). --- ### Uneven Evolution of Cognition Across Generations of Generative AI Models Authors: Isaac Galatzer-Levy, Daniel McDuff, Xin Liu, Jed McGiffin Published: 2026-05-07 Categories: cs.AI, cs.CV arXiv: https://arxiv.org/abs/2605.06815v1 The pursuit of artificial general intelligence necessitates robust methods for evaluating the cognitive capabilities of models beyond narrow task performance. Here, we introduce a psychometric framework to assess the cognitive profiles of generative AI, comparing them to human norms and tracking their evolution across generations. Initial evaluation of leading multimodal models using tasks adapted from the Wechsler Adult Intelligence Scale revealed a profoundly uneven cognitive architecture: near-ceiling performance in verbal comprehension and working memory (>$98^{\text{th}}$ percentile) contrasted with near-floor performance in perceptual reasoning (<$1^{\text{st}}$ percentile). To track developmental trajectories beyond human-normed limits, we developed the Artificial Intelligence Quotient (AIQ) Benchmark and applied it to six generations and two model families, revealing significant but asymmetric performance gains. Notably, we uncovered a sharp dissociation between modalities; abstract quantitative reasoning matured far more rapidly when presented linguistically compared to a visually analogous format, indicating an architectural bias towards language-based symbolic manipulation. While abstract visual reasoning improved, visual-perceptual organization remained largely stagnant. Collectively, these findings demonstrate that the cognitive abilities of generative models are evolving unevenly, suggesting that scaling and optimization approaches to AGI development alone may be insufficient to overcome fundamental architectural limitations in achieving balanced, human-like general intelligence. --- ### From Model to Data (M2D): Shifting Complexity from GNNs to Graphs for Transparent Graph Learning Authors: Debolina Halder Lina, Arlei Silva Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06814v1 Graph Neural Networks (GNNs) achieve high performance but can be opaque to humans, making it difficult to understand and compare the many proposed architectures. While existing explainability methods attribute individual predictions to nodes, edges, or features, they do not provide architectural transparency or explain the fundamental performance gap between simple and more complex models. To address this limitation, we introduce Model-to-Data (M2D) distillation, a new framework that increases transparency by transferring model complexity into the data space. M2D distills the teacher model into an augmented graph with enriched features and structure, enabling a simple student to match the teacher's performance. By materializing model behavior in the data, our approach allows humans to inspect architectural advantages directly. We show that M2D reveals underlying mechanisms such as fairness objectives and attention-based aggregation in an interpretable way, enhancing GNN transparency while preserving performance. --- ### Towards Security-Auditable LLM Agents: A Unified Graph Representation Authors: Chaofan Li, Lyuye Zhang, Jintao Zhai, Siyue Feng, Xichun Yang, Huahao Wang, Shihan Dou, Yu Ji, Yutao Hu, Yueming Wu, Yang Liu, Deqing Zou Published: 2026-05-07 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.06812v1 LLM-based agentic systems are rapidly evolving to perform complex autonomous tasks through dynamic tool invocation, stateful memory management, and multi-agent collaboration. However, this semantics-driven execution paradigm creates a severe semantic gap between low-level physical events and high-level execution intent, making post-hoc security auditing fundamentally difficult. Existing representation mechanisms, including static SBOMs and runtime logs, provide only fragmented evidence and fail to capture cognitive-state evolution, capability bindings, persistent memory contamination, and cascading risk propagation across interacting agents. To bridge this gap, we propose Agent-BOM, a unified structural representation for agent security auditing. Agent-BOM models an agentic system as a hierarchical attributed directed graph that separates static capability bases, such as models, tools, and long-term memory, from dynamic runtime semantic states, such as goals, reasoning trajectories, and actions. These layers are connected through semantic edges and security attributes, transforming fragmented execution traces into queryable audit paths. Building on Agent-BOM, we develop a graph-query-based paradigm for path-level risk assessment and instantiate it with the OWASP Agentic Top 10. We further implement an auditing plugin in the OpenClaw environment to construct Agent-BOM from live executions. Evaluation on representative real-world agentic attack scenarios shows that Agent-BOM can reconstruct stealthy attack chains, including cross-session memory poisoning and tool misuse, capability supply-chain hijacking and unexpected code execution, multi-agent ecosystem hijacking, and privilege and trust abuse. These results demonstrate that Agent-BOM provides a unified and auditable foundation for root-cause analysis and security adjudication in complex agentic ecosystems. --- ### Enhancing Eye Movement Biometrics for User Authentication via Continuous Gaze Offset Score Fusion Authors: Hashim Aziz, Mehedi Hasan Raju, Oleg V. Komogortsev Published: 2026-05-07 Categories: cs.HC, cs.CV arXiv: https://arxiv.org/abs/2605.06810v1 Eye movement biometrics (EMB) use subject-specific gaze dynamics for user authentication and identification. Recent deep learning-based EMB systems achieve strong performance by modeling temporal eye movement behavior. However, these systems typically overlook continuous gaze offset, despite prior evidence that it contains user-discriminative information. This work examines whether continuous gaze offset can improve biometric performance when combined with existing biometric features. We evaluate linear and nonlinear fusion methods on two publicly available datasets, collected via the lab-grade eye tracker and virtual reality headset across multiple tasks and observation durations. Results indicate that fusion offers performance benefits on both datasets, particularly when using nonlinear fusion. Additionally, fusing biometric information across multiple tasks further improves authentication performance. These findings support the hypothesis that continuous gaze offset may serve as useful auxiliary information under conditions of degraded or noisy eye tracking. --- ### LookWhen? Fast Video Recognition by Learning When, Where, and What to Compute Authors: Ali Salamatian, Anthony Fuller, Pritam Sarkar, James R. Green, Leonid Sigal, Evan Shelhamer Published: 2026-05-07 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2605.06809v1 Transformers dominate video recognition. They split videos into tokens, and processing them has expensive superlinear computational cost. Yet videos are filled with redundancy, so we can question the need for this expense. We introduce LookWhen, a selector-extractor framework that factorizes video recognition into learning when, where, and what to compute. Our shallow selector gets a scaled-down video and quickly scores all tokens across space-time, while our deep extractor gets the top-K selected tokens to approximate full-video representations without actually processing all the tokens. A key challenge is defining effective supervision for selection and extraction. For selection pre-training, we introduce a score on representations that ranks tokens by uniqueness using a simple nearest-neighbor distance. For extraction pre-training, we distill both a video teacher and an image teacher, for which we normalize its frame-wise representations to learn what changes within videos. Through these strategies, our selector-extractor learns general and efficient representations for feature extraction or fine-tuning to a task. Through experiments on Kinetics-400, SSv2, Epic-Kitchens, Diving48, Jester, and Charades, we show that LookWhen achieves a better accuracy-computation trade-off than efficient models and upgraded baselines of similar size. LookWhen Pareto-dominates in accuracy-FLOPs on 9 of 12 cases (6 tasks x 2 settings) and roughly matches on 3. In accuracy-throughput, measuring time in practice, LookWhen is more efficient still at 6.7x faster than InternVideo2-B at equal accuracy. --- ### MIND: Monge Inception Distance for Generative Models Evaluation Authors: Quentin Berthet, Yu-Han Wu, Clement Crepy, Romuald Elie, Klaus Greff, Michael Eli Sander Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06797v1 We propose the Monge Inception Distance (MIND), a metric for evaluating generative models that addresses key limitations of the widely adopted Fréchet Inception Distance (FID). The MIND metric leverages the sliced Wasserstein distance to compare distributions by averaging one-dimensional optimal transport distances, efficiently computed via sorting. This approach circumvents the estimation of high-dimensional means and covariance matrices, which underlie FID's poor sample complexity and vulnerability to adversarial attacks. We empirically demonstrate three primary advantages: (i) it is more sample-efficient by one order of magnitude, (ii) it is faster to compute by two orders of magnitude, (iii) it is more robust to adversarial attacks such as moment-matching. We show that MIND with 5k samples can replace the evaluation performance of FID with 50k samples, providing high correlation with this standard benchmark and superior discriminative performance. We further demonstrate that even smaller sample sizes (e.g., 1k or 2k) remain highly informative for rapid model iteration. --- ### Conformal Agent Error Attribution Authors: Naihe Feng, Yi Sui, Shiyi Hou, Ga Wu, Jesse C. Cresswell Published: 2026-05-07 Categories: cs.LG, cs.MA arXiv: https://arxiv.org/abs/2605.06788v1 When multi-agent systems (MAS) fail, identifying where the decisive error occurred is the first step for automated recovery to an earlier state. Error attribution remains a fundamental challenge due to the long interaction traces that large language model-based MAS generate. This paper presents a framework for error attribution based on conformal prediction (CP) which provides finite-sample, distribution-free coverage guarantees. We introduce new algorithms for filtration-based CP designed for sequential data such as agent trajectories. Unlike existing CP algorithms, our approach predicts sets that are contiguous sequences to enable efficient recovery and debugging. We verify our theoretical guarantees on a variety of agents and datasets, show that errors can be precisely isolated, then use prediction sets to rollback MAS to correct their own errors. Our overall approach is model-agnostic, and offers a principled uncertainty layer for MAS error attribution. We release code at https://github.com/layer6ai-labs/conformal-agent-error-attribution. --- ### Distributional Process Reward Models: Calibrated Prediction of Future Rewards via Conditional Optimal Transport Authors: Rachel Ma, Dylan Hadfield-Menell, Kristjan Greenewald Published: 2026-05-07 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.06785v2 Inference-time scaling methods rely on Process Reward Models (PRMs), which are often poorly calibrated and overestimate success probabilities. We propose, to our knowledge, the first use of conditional optimal transport for calibrating PRMs, modifying conditional OT (CondOT) map learning \cite{bunne2022supervised} to estimate a monotonic conditional quantile function over success probabilities estimated by the PRM, conditioned on PRM hidden states. This yields structurally valid quantile estimates and enables efficient extraction of confidence bounds at arbitrary levels, which we integrate into the instance-adaptive scaling (IAS) framework of \cite{park2025know}. We evaluate on mathematical reasoning benchmarks spanning moderate-difficulty problems (MATH-500) and harder out-of-distribution problems (AIME). For PRMs with reliable ranking signals, our method substantially improves calibration over both uncalibrated PRMs and quantile regression. On downstream Best-of-N IAS performance, our method generally improves over uncalibrated PRMs. These results establish conditional optimal transport as another principled and practical approach to PRM calibration, offering structural guarantees and flexible uncertainty estimation. --- ### When Does Critique Improve AI-Assisted Theoretical Physics? SCALAR: Structured Critic--Actor Loop for Agentic Reasoning Authors: Vasilis Niarchos, Constantinos Papageorgakis, Alexander G. Stapleton, Sokratis Trifinopoulos Published: 2026-05-07 Categories: cs.AI, cs.HC, hep-ph, hep-th arXiv: https://arxiv.org/abs/2605.06772v2 As large language models (LLMs) show increasing promise on research-level physics reasoning tasks and agentic AI becomes more common, a practical question emerges: How does the interaction between researchers and agents affect the results? We study this using SCALAR (Structured Critic--Actor Loop for Agentic Reasoning), an Actor--Critic--Judge pipeline applied to quantum field theory and string theory problems. The Actor proposes solutions, the Critic provides iterative feedback, and an independent Judge evaluates the transcript against reference solutions. We vary the Actor persona, the Critic feedback strategy, and the Actor model family and scale. Multi-turn dialogue improves over single-shot attempts throughout, but both the mechanism of improvement and the value of different prompting choices depend strongly on the Actor--Critic pairing. Increasing the scale within one model family (e.g. from the 8B-parameter DeepSeek-R1 variant to DeepSeek-R1 70B) improves some easier-problem behavior, but does not remove the hardest bottleneck we observe. Critic feedback strategy matters most clearly in the asymmetric Actor--Critic setting (e.g, a lightweight Haiku Actor guided by a stronger Sonnet Critic), where constructive feedback improves mean-score outcomes. In same-family Actor--Critic settings, strategy effects are weaker: lenient feedback is sometimes favored, while strict and adversarial feedback are not beneficial. Taken together, SCALAR provides a controlled testbed for evaluating which interaction structures help or hinder AI-assisted reasoning on known-answer scientific problems. --- ### When Does Critique Improve AI-Assisted Theoretical Physics? SCALAR: Structured Critic--Actor Loop for Agentic Reasoning Authors: Vasilis Niarchos, Constantinos Papageorgakis, Alexander G. Stapleton, Sokratis Trifinopoulos Published: 2026-05-07 Categories: cs.AI, cs.HC, hep-ph, hep-th arXiv: https://arxiv.org/abs/2605.06772v1 As large language models (LLMs) show increasing promise on research-level physics reasoning tasks and agentic AI becomes more common, a practical question emerges: How does the interaction between researchers and agents affect the results? We study this using SCALAR (Structured Critic--Actor Loop for AI Reasoning), an Actor--Critic--Judge pipeline applied to quantum field theory and string theory problems. The Actor proposes solutions, the Critic provides iterative feedback, and an independent Judge evaluates the transcript against reference solutions. We vary the Actor persona, the Critic feedback strategy, and the Actor model family and scale. Multi-turn dialogue improves over single-shot attempts throughout, but both the mechanism of improvement and the value of different prompting choices depend strongly on the Actor--Critic pairing. Increasing the scale within one model family (e.g. from the 8B-parameter DeepSeek-R1 variant to DeepSeek-R1 70B) improves some easier-problem behavior, but does not remove the hardest bottleneck we observe. Critic feedback strategy matters most clearly in the asymmetric Actor--Critic setting (e.g., a lightweight Haiku Actor guided by a stronger Sonnet Critic), where constructive feedback improves mean-score outcomes. In same-family Actor--Critic settings, strategy effects are weaker: lenient feedback is sometimes favored, while strict and adversarial feedback are not beneficial. Taken together, SCALAR provides a controlled testbed for evaluating which interaction structures help or hinder AI-driven scientific discovery. --- ### VITA-QinYu: Expressive Spoken Language Model for Role-Playing and Singing Authors: Jiacheng Xu, Heting Gao, Liufei Xie, Zhenchuan Yang, Lijiang Li, Yiting Chen, Bin Zhang, Meng Chen, Chaoyu Fu, Weifeng Zhao, Wenjiang Zhou Published: 2026-05-07 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2605.06765v1 Human speech conveys expressiveness beyond linguistic content, including personality, mood, or performance elements, such as a comforting tone or humming a song, which we formalize as role-playing and singing. We present VITA-QinYu, the first expressive end-to-end (E2E) spoken language model (SLM) that goes beyond natural conversation to support both role-playing and singing generation. VITA-QinYu adopts a hybrid speech-text paradigm that extends interleaved text-audio modeling with multi-codebook audio tokens, a design enabling richer paralinguistic representation while preserving a clear separation between modalities to avoid interference. We further develop a comprehensive data generation pipeline to synthesize a total of 15.8K hours of natural conversation, role-playing, and singing data for training. VITA-QinYu demonstrates superior expressiveness, outperforming peer SLMs by 7 percentage points on objective role-playing benchmarks, and surpassing peer models by 0.13 points on a 5-point MOS scale for singing. Simultaneously, it achieves state-of-the-art conversational accuracy and fluency, exceeding prior SLMs by 1.38 and 4.98 percentage points on the C3 and URO benchmarks, respectively. We open-source our code and models and provide an easy-to-use demo with full-stack support for streaming and full-duplex interaction. --- ### Revisiting Adam for Streaming Reinforcement Learning Authors: Florin Gogianu, Adrian Catalin Lutu, Razvan Pascanu Published: 2026-05-07 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.06764v1 Learning from a sequence of interactions, as soon as observations are perceived and acted upon, without explicitly storing them, holds the promise of simpler, more efficient and adaptive algorithms. For over a decade, however, deep reinforcement learning walked the contrary path, augmenting agents with replay buffers or parallel sampling routines, in an effort to tame learning instability. Recently, this topic has been revisited by Elsayed et al. (2024), focusing on update computation through eligibility traces and modifications to the optimisation routine, resulting in the StreamQ algorithm. In this work we take a step back, investigating the efficacy of established updates, such as those implemented by DQN and C51 within this online setting. Not only do we find that they perform well, but through analysing how the optimisation algorithm generally, and Adam in particular, interacts with these updates, we contend that two properties are essential for robust performance: i) the derivative of the objective is to be bounded and ii) weight updates are variance-adjusted. Rigorous and exhaustive experimentation demonstrates that C51, which exhibits both characteristics, is competitive with StreamQ across a subset of 55 Atari games. Using these insights, we derive a variance-adjusted algorithm based on eligibility traces, termed Adaptive Q$(λ)$, which approaches double the human baseline on the same subset, surpassing existing methods by all performance metrics. --- ### Sparse Attention as a Range Searching Problem: Towards an Inference-Efficient Index for KV Cache Authors: Mohsen Dehghankar, Abolfazl Asudeh Published: 2026-05-07 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.06763v2 Sparse attention improves LLM inference efficiency by selecting a subset of key-value entries, but at the cost of potential accuracy degradation. In particular, omitting critical KV entries can induce substantial errors in model outputs. Existing methods typically operate under fixed or adaptive token budgets and provide empirical robustness or partial theoretical guarantees, yet they do not ensure zero false negatives in decoding steps, particularly since the set of relevant tokens is both query- and step-dependent. Our empirical observations confirm that missing even one critical key can lead to sharp error spikes, especially in long reasoning tasks where the set of important tokens varies throughout decoding. This observation motivates the need for indexing methods that dynamically adapt to these variations across decoding steps while guaranteeing a full recall of the relevant keys above a certain threshold. We address this challenge by reformulating sparse attention as the halfspace range searching problem. However, existing range searching indices are not suitable for modern LLM inference due to their computational and implementation overheads. To overcome this, we introduce Louver, a novel index structure tailored for efficient KV cache retrieval. Louver (i) guarantees zero false negatives with respect to a specified threshold in both theory and practice, (ii) is lightweight to integrate into existing LLM pipelines, and (iii) incorporates hardware-aware optimizations for both CPU and GPU executions. Our experiments demonstrate that Louver outperforms prior sparse attention methods in both accuracy and runtime, and is faster than highly optimized dense attentions such as FlashAttention. These results highlight that recall guarantees are a critical and overlooked dimension of sparse attention, and open a new direction for building theoretically grounded, efficient KV cache indices. ---