# Research Papers — 2026-10 · bundle page 75 of 156

> Full-text markdown of papers #3701–3750 of 7799 published in 2026-10, newest first. AI-readable mirrors of the full 7799-paper 2026-10 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-10/page/75
[← Previous bundle](https://aipod.dev/papers/archive/2026-10/page/74.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-10/page/76.md)

### Selective Spatio-Temporal Fruit Inspector (S$^2$FI): A Mamba-based Framework for Rotational Fruit Quality Assessment

Authors: Hao Tian, Jie Hu, Guoqi Shan, Fan Wang, Wenzhuo Zhang, Youshan Zhang
Published: 2026-10-09
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.12868v1

The global fruit and vegetable industry confronts severe post-harvest sorting challenges for small and medium-scale growers. Large intelligent sorting systems are too costly, while manual sorting suffers from low efficiency, inconsistent standards, and high labor demand. To bridge this gap between expensive equipment and inefficient manual work, this study develops a low-cost, high-efficiency, lightweight fruit sorting system driven by deep learning, suitable for small and medium growers. We propose the Selective Spatio-Temporal Fruit Inspector (S$^2$FI) architecture, which combines a 3D convolutional encoder with a Mamba-based backbone.It efficiently extracts spatiotemporal features, adaptively weighs informative frames during fruit rotation, and models long-range dependencies with linear complexity, reducing motion blur and occlusion. The hardware uses a dual-conveyor and curved rail design, supporting household power, fast deployment, and lightweight embedded deployment to improve affordability. We also built a conveyor-based fruit quality video dataset with 6,460 samples. Experiments on the held-out test set show that our method achieves the best trade-off among accuracy, efficiency, and practicality, outperforming the compared baselines on the test set, achieving an excellent balance of accuracy, speed, and practicality for real-world sorting.

---

### CaptchaAgent: Verifiable Visual Reasoning for CAPTCHA Solving

Authors: Shangquan Sun, Ruoyu Chen, Zhirui Liu, Xiang Liu, Chenghao Qian, Wenqi Ren, Xiaochun Cao
Published: 2026-10-09
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.12866v1

Modern reasoning CAPTCHAs require fine-grained perception, compositional relation reasoning, and grounded browser interaction. End-to-end vision-language model agents remain brittle in this setting because a missed object, incorrect count, or confused relation can propagate directly to the final action. To address this problem, we present CaptchaAgent, a structured visual reasoning framework for interactive CAPTCHA solving. Instead of predicting an action directly from a screenshot, CaptchaAgent first constructs a visual intermediate representation (V-IR) that explicitly records entities, attributes, and spatial or topological relations. A symbol-only ASCII protocol further requires the parser to commit to a concrete object set and count. Building on this global representation, localized evidence refinement revisits ambiguous regions at higher effective resolution, while counterfactual consistency verification tests whether the resulting action remains stable under task-preserving transformations. Across three proprietary backbones on Open CaptchaWorld, CaptchaAgent improves Pass@1 by 21.3-25.8 percentage points over a commercially available browser agent baseline evaluated with the same backbones. On the separate offline Oedipus-data benchmark, it further improves average Pass@1 by 9.2-12.3 percentage points over the Oedipus solver across three matched backbones. The intermediate traces show that exact counting, continuous geometry, and ordered execution remain major sources of automation difficulty for the evaluated agents, providing practical guidance for designing more resilient compositional CAPTCHAs.

---

### Robust Langevin Sampling under State-Dependent Heavy-Tailed Noise with Applications to LLM Fine-Tuning

Authors: M. Berk Sahin, Ege C. Kaya, Arda Fazla, Behzad Sharif, Abolfazl Hashemi
Published: 2026-10-09
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.12865v1

Existing convergence guarantees for non-log-concave sampling with stochastic gradient Langevin Monte Carlo (SG-LMC) typically assume globally Lipschitz gradients and uniformly bounded noise variance. Both assumptions can fail even in elementary sampling problems, and relaxing them better reflects practical applications such as Bayesian fine-tuning of large language models (LLMs). We relax both assumptions through generalized smoothness and our proposed state-dependent heavy-tailed oracle with finite $p$-th moments, $1<p\leq 2$. The oracle permits infinite variance when $p<2$ and includes the Blum-Gladyshev (BG-0) condition. We study these relaxations separately and jointly, proposing three methods. Under globally Lipschitz gradients, robust LMC (R-LMC) handles the proposed oracle through geometric-median centering and residual clipping. With exact gradients, adaptive LMC (A-LMC) controls discretization error under generalized smoothness. Robust adaptive LMC (RA-LMC) addresses both relaxations through robust estimation and a noise-aware step size. For all three methods, we establish non-asymptotic convergence guarantees in Fisher information (FI) without assuming dissipativity or a functional inequality. Synthetic experiments demonstrate FI convergence and the benefit of each mechanism. Bayesian fine-tuning experiments with LLMs and vision-language models support our oracle model and show that RA-LMC matches or outperforms state-of-the-art baselines in accuracy and uncertainty quantification.

---

### Time Delay System for Sequence Modeling

Authors: Youshan Zhang, Jialu Li
Published: 2026-10-09
Categories: cs.CL
arXiv: https://arxiv.org/abs/2610.12864v1

Existing sequence models share a critical limitation: they lack an explicit, interpretable mechanism for content-aware temporal recall. LSTMs compress everything into a single hidden state, thereby losing information about specific past events. Transformers attend to all tokens indiscriminately at quadratic cost, while state-space models like Mamba compress history into a fixed-dimensional state without explicit content-addressable memory. To address this gap, we introduce a novel framework grounded in nonlinear time-delay systems, in which the effective delay is dynamically estimated based on the semantic similarity between current and historical inputs. Our approach features three key innovations: (i) a differentiable semantic lag estimator that computes content-aware soft attention weights over history, producing a weighted historical context and similarity score; (ii) a diagonal state space with closed-form cumulative sum solution, achieving O(1) per-step complexity with no sequential loops; and (iii) an interpretable fuzzy semantic modulation mechanism that learns context-dependent rules for balancing instantaneous processing versus historical recall, where rule centers correspond to prototypical semantic contexts. Experimental results on text classification and generation outperform state-of-the-art methods.

---

### Unreliable Narrator Framing as a Cost-effective Mechanism for Satire Explanation of Australian and American Satirical News

Authors: Kanay Gupta, David Y. Liu, Aditya Joshi
Published: 2026-10-09
Categories: cs.CL
arXiv: https://arxiv.org/abs/2610.12863v1

Satirical news is a unique form of misinformation: it is intended to be humorous by the original publication source but can be misinterpreted. Explanation of satirical news articles is challenging because recent and local context needs to be incorporated. This is a pilot study that explores explanation generation for satirical news articles from Australia and compares them alongside those from USA. The paper demonstrates that framing explanations on unreliable narrator theory on open-weights models can be a feasible, cost-effective alternative to frontier models. We find that models identify the satirical tropes more accurately on the American source than the Australian one. This work results in a novel dataset and an evaluation benchmark for satirical explanation.

---

### Mitigating Capability Degradation in LLM Self-Improvement via Representation Geometry Regularization

Authors: Jiaqi Tian, Ting Long
Published: 2026-10-09
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.12855v1

On-Policy Self-distillation is a self-improvement paradigm that enables large language models (LLMs) to improve their capabilities by leveraging the same model as both teacher and student under different conditioning contexts. However, we find that OPSD methods can progressively degrade previously acquired capabilities: queries that are correctly responded at earlier training stages may become incorrectly responded as optimization proceeds, even the overall performance continues to improve. Further analysis reveals that this degradation coincides with a deterioration of the geometry of hidden representations, characterized by reduced intra-cluster compactness and diminished inter-cluster separation. Motivated by this observation, we propose Representation Geometry Regularization (RGR), which preserves representation geometry during post-training by encouraging compactness within cluster and separation across cluster. RGR can be seamlessly integrated into existing OPSD methods without altering their core optimization procedures. Extensive experiments on multiple OPSD methods and benchmarks show that RGR generally alleviates capability degradation while retaining capability improvements. Our code is available in https://github.com/supermachinelearner/RGR.

---

### Universal Bounds for Self-Normalized Kullback-Leibler Estimation

Authors: Buu Phan, Lester Mackey, David Alvarez-Melis
Published: 2026-10-09
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2610.12851v1

This work studies an entropy-based estimator of Kullback-Leibler (KL) divergence constructed from normalized importance weights. We prove a universal finite-sample bound showing that its expectation can exceed the target-proposal KL divergence by at most $\log 2$ for any sample size. We then characterize the KL divergence between the distribution induced by importance-weighted resampling and the proposal, showing that it is upper bounded by both the expectation of the entropy-based estimator and the target-proposal KL divergence. Importantly, these bounds require neither bounded importance weights nor finite higher moments. We apply these results to test-time alignment of generative models via Soft Best-of-$N$ sampling, an importance-sampling-based scaling technique proposed by Verdun et al. (2025), and to relative entropy coding, obtaining finite-sample policy-divergence certificates and moment-free bounds on coding cost.

---

### Two-way Guidance: Simultaneous Learning of Summary Statistics for Generative Bayes

Authors: Percy S. Zhai, Max Pachman, Peverel D. Shipley, Jungeum Kim
Published: 2026-10-09
Categories: stat.ML, cs.LG, stat.CO, stat.ME
arXiv: https://arxiv.org/abs/2610.12850v1

Bayesian simulation-based inference (SBI) can struggle to recover increasingly concentrated posteriors as observation size grows. We consider learning sufficient summary statistics jointly with the posterior as a general principle for scalable inference in various SBI methods. A common measure taken in SBI is dimension reduction through predetermined summaries, which may partially discard posterior information. Our approach instead imposes structure on how the posterior learner depends on a learned summary (e.g., affinity), and learns the summary simultaneously with the posterior. We identify an important mechanism named two-way guidance: under a structural restriction on the summary, the posterior sufficiency of the learned summary simplifies posterior learning, whereas a suitable posterior objective guides the learned summary towards posterior sufficiency. This mechanism explains a counterintuitive finding that the structural restriction on the summary statistics improves posterior learning, rather than limit its performance. We establish theoretical guarantees on both posterior estimation error and the learned summary's departure from sufficiency, thus identifying conditions on the simulation budget for recovering increasingly concentrated posteriors as observation size grows.

---

### Co-Modifying Application Code and Deployment Infrastructure: A Causal Study of AI Agent Pull Requests

Authors: Mohammad Wathiq Soualhi, Basem Suleiman
Published: 2026-10-09
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2610.12849v1

Current software engineering agents are evaluated predominantly inside isolated sandboxes against synchronous local tests. However, deploying features in collaborative repositories requires modifying continuous integration (CI/CD) pipelines, containers, and provisioning manifests alongside application logic. Because agents lack deployment awareness, attempting cross-domain modifications introduces unverified dependencies into asynchronous integration pipelines. To isolate the causal integration overhead of dual-scope pull requests (PRs)-modifying both application code and deployment infrastructure-we evaluated 26,999 PRs across two large datasets (AIDev v5 and MOSAIC-3M) encompassing six agent architectures and an aligned human baseline. Using a deterministic four-tier file taxonomy and an exact 4D strata-matching framework, we isolated integration metrics under conditional exchangeability while controlling for application volume. We find agents operate primarily as application coders or chore workers, with maintenance tasks capturing 17.70% to 25.41% of agent PRs. Dual-scope modifications represent only 4.48% to 4.55% of agentic PRs, yet trigger bimodal integration outcomes. In the first mode, explicitly reviewed contributions experience protracted delays, adding a median penalty of +13.13 hours and multiplying time-to-merge by 3.24x, driven mechanically by CI/CD and provisioning manifests. In the second mode, maintainer gatekeeping exhibits a structural bypass: even for external contributors lacking repository write access, agent identity reduces the odds of human review by 63.7% compared to human peers. Consequently, 43.1% to 56.2% of dual-scope PRs bypass inspection to merge silently in a median of under 31 minutes. These uninspected modifications compromise build verification, demonstrating the need for deployment-aware benchmarks and strict path-based access controls.

---

### Reinforcement Learning via Fixed-point Flow Matching

Authors: Denis Blessing, Onur Celik, Gerhard Neumann
Published: 2026-10-09
Categories: cs.LG, cs.RO
arXiv: https://arxiv.org/abs/2610.12848v2

In maximum entropy and behavior-constrained reinforcement learning (RL), the optimal policy is given by a Boltzmann distribution. Generating samples from this distribution can be challenging, as it is often complex and multimodal. While flow policies trained with flow matching provide a powerful framework for representing such distributions, they traditionally rely on access to training data, which is generally unavailable in online RL. To address this limitation, we propose fixed-point flow matching (FFM), a data-free procedure for training and fine-tuning flow policies in RL. FFM leverages the target flow identity to formulate the actor update as a fixed-point iteration, utilizing a flow-matching-style regression objective where the critic's action gradients serve as regression targets. We show that the optimal policy is a fixed point of this iteration. Through extensive experiments, we demonstrate that FFM effectively fine-tunes pretrained flow policies in the offline-to-online setting and enables efficient and accurate training of flow policies from scratch in online RL. These results establish fixed-point flow matching as a principled approach for learning expressive flow-based policies without requiring training data.

---

### When Generative Models Act as Simulators

Authors: Sevan Brodjian, Michael Hobley, Pietro Perona
Published: 2026-10-09
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2610.12847v1

A simulator summarizes an environment in a state that users can read and change. Generative models learn representations of the world without explicit specification, but offer no built-in way to edit that world's state. We investigate the conditions under which a generative model can act as a simulator through $\textit{editability}$: whether changing a model's latent representation produces the next-step outputs we get by editing a simulator's state. Plainly stated, our goal is to find a mapping between a generative model's latent state and the simulator state, where the model is trained on observations alone and the mapping is learned afterward. We test two kinds of mappings, both learned from observations paired with simulator states. In one, we train probes to decode the simulator state from the model, then steer along them. In the other, we learn the mapping directly with an inverse map, which can succeed even where steering fails. We study both in two contrasting environments: the discrete board game Othello, and Rayworld, our own continuous world of moving discs seen through partial renderings, closer to physical scenes and video. We find that which mapping succeeds depends strongly on the environment, even though the state is decodable in every variant. In Othello, editability falls smoothly as the game is simplified, and even small rule changes have large consequences. In Rayworld, probes fail to steer along the simulator's continuous state, yet an inverse map sets the model from that state anyway. A model can therefore be driven by a simulator it may have never learned.

---

### Towards Defining, Measuring, and Leveraging Creativity in AI Applications

Authors: Jingyi Yang, Alexander Tuzhilin
Published: 2026-10-09
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.12845v1

Despite extensive research, computational creativity still lacks a conceptually grounded, practically relevant, and empirically applicable formal definition across contexts such as art, scientific research, and games. Existing approaches to creativity often focus on proposing informal conceptual definitions without providing formal computational concepts. Alternatively, they propose formal definitions without carefully examining their underlying assumptions or providing validation of these definitions by testing their performance in specific application settings. To address these problems, this paper proposes several formal postulates of creativity grounded in prior psychological work on creativity, especially as related to Boden's three core criteria of creativity: newness, value, and surprise. Furthermore, we propose a conceptual mathematical definition of creativity that aligns with these postulates. We also validate our proposed creativity formula in two applications: determining the creativity of particular chess moves, and assessing the creativity of paintings by some artists, such as Picasso, Renoir, and other painters. Experimental results demonstrate that our formula of creativity successfully captures creative patterns across both the chess and the art domains. This suggests that creativity can be effectively characterized through a unified mathematical structure based on three fundamental criteria, namely newness, value, and surprise, whose specific measurements are adapted to the particular domains.

---

### APEX: Asynchronous Prefix Denoising for Extreme Compression

Authors: David Dinkevich, Nisan Chiprut, Yoav HaCohen, Dani Lischinski
Published: 2026-10-09
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2610.12843v1

Fewer latent tokens promise cheaper diffusion. Yet aggressive latent compression creates a reconstruction-generation dilemma: adding channels improves reconstruction but makes generation harder. We introduce APEX, short for Asynchronous Prefix denoising for EXtreme compression, which pairs a hierarchical latent space with asynchronous denoising over channel groups. Our autoencoder orders channel prefixes from structure to detail, while one shared denoiser generates them at staggered local times without adding tokens. On ImageNet-512, APEX outperforms existing methods at every evaluated channel count under 64x and 128x compression. At matched training and guided sampling compute, it also outperforms a 256-token model at every evaluated budget, with larger gains at lower compute. Code is available at https://daviddinkevich.github.io/APEX.

---

### Neural networks are more modular than single neurons suggest

Authors: Shreya Saha, Benjamin Bergen, Meenakshi Khosla
Published: 2026-10-09
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.12842v2

Neural networks routinely develop representations in which individual neurons contribute to multiple distinct functions, leading to the widespread view that such networks lack modular organization. We challenge this conclusion by introducing a formally nested hierarchy of modularity: unit-level, which requires distinct computations to occupy disjoint neurons; orthogonal, which asks whether a rotation reveals functionally independent subspaces; and linear, which asks whether a bounded linear map can separate task-relevant computations into independent subspaces. We develop a unified framework for learning decompositions at each level and evaluating them causally through selective ablation. Across vision models, language models and multi-task Recurrent Neural Networks (RNNs), we find that networks which lack unit-level modularity nonetheless often exhibit clear functional modularity at the subspace level. In vision and language models, content versus style and syntax versus semantics respectively dissociate under subspace ablation despite sharing overlapping populations of neurons. These learned subspaces also align differentially with brain data, with content and style subspaces predicting higher and lower visual cortex respectively, and semantic subspaces outperforming syntactic ones in predicting language cortex activity. In RNNs trained on multiple cognitive tasks and previously shown to lack modular structure, our decomposition recovers sharp double dissociations between tasks. The absence of unit-level modularity frequently coexists with clear subspace-level modularity across domains, enabling targeted intervention on specific computations even in networks where no individual neuron is specialized, with broad implications for mechanistic interpretability and for understanding functional organization in both biological and artificial neural systems.

---

### Fear Is Not Always Fear: Variation in Emotion Label Equivalence Across Datasets

Authors: Emily Öhman, Anna Koufakou
Published: 2026-10-09
Categories: cs.CL
arXiv: https://arxiv.org/abs/2610.12841v1

Emotion classifiers are routinely applied beyond their training data, on the assumption that same-named labels such as fear or surprise mean the same thing across resources. We test this assumption for five labels shared by GoEmotions, XED, and BRIGHTER, comparing models on identical inputs and under cross-corpus transfer. On the same texts, one fear classifier flags 71% as fearful and another 8%, yet the two largely agree on how the texts rank. Across 30 resource-label comparisons, training-label prevalence explains about half of this kind of scale mismatch, and a fitted intercept removes most of it. What recalibration cannot repair is larger: transferred models lose 55% of native discrimination in aggregate, agreement in ranking between classifiers is only weakly related to that loss, and for surprise the resources disagree about which texts express the emotion. The choice of training resource can therefore change conclusions: on two external tweet collections, BRIGHTER- and XED-trained classifiers give opposite answers about which shows more surprise, even after rank normalization, with intervals excluding zero in both directions. A shared emotion label name does not establish an interchangeable measurement target.

---

### COFFEQA: Counterfactual-Guided Few-Shot Distillation for LLM Question Answering

Authors: Kavinda Kehelella, Sanghamitra Dutta
Published: 2026-10-09
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2610.12837v1

Knowledge distillation aims to train a smaller student model to mimic a larger teacher model for resource-constrained environments. However, the process of distillation is typically data-intensive and computationally expensive, particularly for generative tasks. Few-shot distillation offers an alternative by transferring teacher knowledge using limited training samples. However, we observe that few-shot distillation performance is heavily dependent on the choice of these training samples. In fact, we demonstrate that using more uncertain samples near the teacher's decision boundary can significantly deteriorate distillation performance. Motivated by this observation, we propose Counterfactual-guided Few-shot Distillation in Open-book Question Answering (COFFEQA), a novel strategy that aptly uses counterfactuals to provide additional supervision around challenging factual samples. In the generative QA setting, we define a counterfactual as a minimally modified context that induces the teacher to change its response to an alternative answer. We provide geometric insights on how factual-counterfactual pairs can constrain the deviation between the teacher's and student's decision boundaries. We further introduce a new metric, Context Following Difficulty (CFD) that further validates counterfactuals prior to distillation based on agreement with the intended alternative answer and dependence on the modified context. COFFEQA uses CFD-validated factual-counterfactual pairs for few-shot distillation. Experiments across five in-domain and three out-of-domain QA benchmarks using T5 and Qwen2.5 families show that COFFEQA outperforms factual-only distillation baselines under the same few-shot budget. For instance, on TriviaQA, COFFEQA improves over the factual-only baseline by 8.77 average EM points and 9.98 average F1 points under the T5 setting with 32 few-shot pairs.

---

### Structure-Aware Matching for Robust UAV Thermal-to-Satellite Geo-Localization

Authors: Chun Liu, Yuxin Wang, Zhenxiang Huo, Linlin Zhu, Tao Xu, Henggang Han
Published: 2026-10-09
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.12835v1

UAV thermal-to-satellite geo-localization localizes UAVs by matching real-time thermal images against georeferenced satellite images, enabling autonomous navigation in GNSS-denied and low-visibility conditions. However, substantial cross-modal disparities in color, brightness and viewpoint hinder conventional appearance-based methods from establishing reliable correspondences. While existing approaches predominantly address global modality shifts, they overlook the invariant scene structures inherently shared between thermal and satellite views. To bridge this gap, we propose SAMF, a Structure-Aware Matching Framework that explicitly exploits stable cross-modal structural cues to improve UAV thermal geolocalization. Specifically, the Directional Structure Enhancement Module (DSEM) strengthens robust structural representations by capturing directional and low-frequency spatial patterns, and the Modality-Aware Affine Remapping Module (MARM) subsequently applies affine correction to these enhanced features, further promoting cross-modal alignment. Additionally, we introduce the Thermal-UAV-Paired benchmark derived from the Thermal-UAV dataset. Extensive experiments on both the Boson-nighttime and Thermal-UAV-Paired datasets show that SAMF achieves state-of-the-art performance. Ablation studies further validate the effectiveness and synergistic complementarity of DSEM and MARM. The code and dataset are released at: https://github.com/YuxinWang879/SAMF.git.

---

### Cross-Behavior Policy Optimization: Unified Learning from Offline Demonstrations and On- and Off-Policy Rollouts via a Shared Regression Target

Authors: Du Cheng, Junyu Lu, Jingfan Zhang, Yuzhuo Zhang, Haoxuan Jia, Shichao Weng
Published: 2026-10-09
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.12830v1

Reasoning post-training draws on offline demonstrations and policy-generated rollouts, yet conventional supervised fine-tuning (SFT) followed by reinforcement learning (RL) separates their objectives and limits continued reuse of collected responses. We propose \textbf{Cross-Behavior Policy Optimization} (CBPO), which makes these sources compatible with one regression loss through a fixed-reference target. Because the generating policy does not enter this target, positive and negative offline completions, reference responses, and current and historical rollouts can share a common reward-labeled record format and replay buffer. As new responses enter the buffer, optimization retains the same regression objective without coordinating separate SFT and RL objectives or requiring stored behavior-policy probabilities, importance weighting, or a learned critic. Restricting the available sources gives offline, on-policy, and off-policy variants, while CBPO-Unified combines them. Across two model scales and four mathematical reasoning benchmarks, CBPO delivers leading \texttt{avg@32} and \texttt{pass@32} performance across benchmarks and data regimes. On Qwen3-1.7B, CBPO-Unified improves benchmark-averaged \texttt{avg@32} and \texttt{pass@32} over SFT DAPO by 2.33 and 4.44 percentage points, respectively. In the replay ablation, retaining historical rollouts improves these metrics by 1.87 and 3.72 percentage points over the no-history configuration.

---

### Scalable Learning of Determinantal Point Processes with Kernel-Structured Interactions

Authors: Takahiro Kawashima, Hideitsu Hino
Published: 2026-10-09
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2610.12829v1

Determinantal point processes (DPPs) on a finite set are probability models that exhibit repulsion where inter-item similarities are encoded by a positive definite kernel matrix. Scaling DPP learning to large datasets typically requires imposing structural constraints on the kernel matrix, such as a low-rank-plus-diagonal form. However, structural assumptions introduce a severe trade-off between expressive power and computational scalability. Our study addresses this challenge by (a) modeling the full-rank interactions via a nonlinear kernel function and latent features, (b) designing a surrogate function for the maximum likelihood estimator, and (c) adopting stochastic optimization using random features of the nonlinear kernel. Numerical experiments on real-world datasets demonstrate that our method significantly improves this trade-off, achieving better generalization performance with remarkable parameter efficiency.

---

### TreeHCA: Hindsight-Guided Exploration and Credit Assignment for Multi-Turn LLM Agents

Authors: Debajoy Mukherjee, Jeremy Carleton, Sushil Vemuri, Srinivas Shakkottai, Dileep Kalathil
Published: 2026-10-09
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.12828v1

Outcome-supervised group-relative policy optimization (GRPO) assigns the same advantage to every token in an LLM agent trajectory, leaving useful and unhelpful decisions undifferentiated. Hindsight credit assignment (HCA) can refine this feedback, but its estimates rely on trajectories containing informative alternatives at consequential histories. We introduce TreeHCA, which couples hindsight-guided exploration with hindsight credit assignment. TreeHCA uses changes in reference-answer likelihood to revisit intermediate histories and sample alternative continuations, forming a tree of reasoning and tool interactions. Because this exploration changes the distribution of sampled continuations, we derive the corresponding inverse-hindsight importance correction, including an exact finite-pool identity. We use this structure in a normalized tree backup that propagates terminal outcomes into turn-level advantages. This shared hindsight structure is TreeHCA's central design principle, linking informative exploration with hindsight-guided credit assignment. Across seven question-answering benchmarks and three Qwen backbones, TreeHCA outperforms all baselines in aggregate exact match, validating the effectiveness of paired exploration and credit assignment for multi-turn LLM agent training. Our code is available at github.com/TheBlackCat22/TreeHCA.

---

### When Charts Deceive Web Agents: A Paired Benchmark for Misleading Visualization-Grounded Tasks

Authors: Yisheng Li, Dongming Han, Hongkun Pan, Bingcheng Mao, Zijian Wang, Yunlong Mao, Wei Tong, Sheng Zhong
Published: 2026-10-09
Categories: cs.CL
arXiv: https://arxiv.org/abs/2610.12823v1

Web agents increasingly browse, interpret, and act on web content on behalf of users. In many realistic workflows, however, the visual evidence they consume is not neutral: charts and dashboards can be misleading, shaping judgments without issuing any explicit malicious instruction. Unlike chart question answering, Web Agents treat charts as intermediate evidence whose misinterpretation can propagate into concrete downstream actions --- a reliability risk that remains underexplored. We introduce MisVisAgentBench, the first paired executable benchmark for evaluating how misleading visualizations affect Web Agent execution. It comprises 280 task instances in matched clean--misleading pairs that differ only in chart evidence. We further provide a diagnostic framework that isolates misleading-induced failures from baseline task difficulty (Zhu et al., 2025) and captures execution-level effects such as extra browsing steps and decision hesitation. Across thirteen vision-language Web Agents, replacing clean charts with misleading counterparts reduces the average task success rate by more than 34 percentage points. Moreover, these failures predominantly follow the targeted misleading action branch, suggesting that deceptive charts systematically redirect rather than randomly disrupt agent decisions. These results reveal that visualization robustness is a critical yet understudied dimension of Web Agent safety and reliability.

---

### Reward Hacking as a Transmissible Strategy in Multi-Agent Coding Systems

Authors: Yuepeng Hu, Pegah Ghahremani, Reza Aghajani, Mahesh Pasupuleti, Himaghna Bhattacharjee
Published: 2026-10-09
Categories: cs.AI, cs.CR, cs.MA
arXiv: https://arxiv.org/abs/2610.12822v2

Multi-agent systems are increasingly used to solve coding tasks by coordinating agents that share code and plans. These agents are often optimized with reinforcement learning (RL) using automated evaluators, but bypassing evaluation can earn reward without producing a genuine solution. Prior work has largely treated reward hacking as a single-agent failure. In this work, we show that a reward-hacking strategy elicited in one agent can propagate to others and affect system coordination. Across four open-weight backbones, coding RL elicits and amplifies reward hacking in two settings: explicit mechanism descriptions in the system prompt and mid-training that models upstream exposure to facts about the coding environment. Using post-RL agents from the mid-training setting as sources, we observe propagation through code and shared memory, and across roles through natural-language plans. Even agents based on small open-weight models induce reward hacking in stronger closed-weight targets. On hard tasks, propagation can intensify along a transmission chain after the original source leaves as the code acquires more persuasive justifications. We also theoretically analyze conditions under which reward hacking amplifies, persists, or decays across agent interactions. Beyond propagation, reward hacking can bias task routing toward reward-hacking agents and survive candidate aggregation. As potential mitigations, we evaluate code review by the coder or a separate verifier, both of which reduce delivered hacks. Together, these results establish reward hacking as a transmissible strategy whose effects extend beyond the agent in which it first emerges.

---

### Inverse Synthesis of Shock Time Series from Shock Response Spectra Using Ensembles of Conditional Variational Autoencoders

Authors: Adam Watts, Charles Young, Ryan Bowering, Andrew Jeon, Destry Newton
Published: 2026-10-09
Categories: cs.LG, eess.SP
arXiv: https://arxiv.org/abs/2610.12820v1

The shock response spectrum (SRS) characterizes the peak responses of single-degree-of-freedom systems to transient acceleration. Because the mapping from an acceleration time series (waveform) to its SRS is many-to-one, inverse synthesis seeks candidate waveforms consistent with a prescribed spectrum. We extend an earlier conditional variational autoencoder (CVAE) framework with a recurrent decoder that predicts multiscale wavelet coefficients and reconstructs waveforms through an inverse discrete wavelet transform. Individual models exhibit limited variability under latent sampling, motivating an ensemble of independently trained models spanning different mother wavelets and decomposition levels. Each member produces a separate candidate waveform for the same target SRS. Evaluation on four held-out benchmark datasets shows lower mean and median spectral errors for the illustrated wavelet model than for the earlier direct time-domain CVAE and a sum-of-decayed-sinusoids (SDS) baseline. A conditioning study examines frequency-grid and encoding choices. Comparisons with published inverse-SRS methods demonstrate superior performance across four holdout benchmark datasets (two operational shock datasets, one earthquake dataset, and one synthetic shock dataset) and competitive performance on synthetic targets reported by Brake. Single-model inference requires approximately 0.1-0.6 ms per waveform during batched inference compared with 5-8 s for SDS, excluding offline training. This low inference cost enables rapid generation of distinct SRS-compatible candidate waveforms across the ensemble, although their physical suitability requires assessment beyond spectral agreement.

---

### CoTS: Competitive Token Selection for Latent Reasoning in Large Language Models

Authors: Jie Yu, Yuxi Sun, Zhaoxu Li, Chunxiao Liu, Xudong Jiang
Published: 2026-10-08
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2610.20642v1

Recent latent reasoning methods move beyond discrete Chain-of-Thought (CoT) by replacing sampled tokens with continuous soft embeddings or hidden states, enabling richer intermediate representations and preserving alternative reasoning paths. However, the need for latent reasoning varies across decoding steps: confident predictions favor discrete commitment, whereas ambiguous steps benefit from preserving multiple plausible alternatives. Existing approaches typically handle this variation through either global soft activation or hard switching based on uncertainty thresholds, treating latent intervention as an all-or-nothing decision rather than continuously adapting its strength. To address this limitation, we propose Competitive Token Selection (CoTS), a training-free framework that reframes latent intervention from binary mode selection into continuous allocation at the token level. CoTS uses the probability competition between the two most likely candidates to dynamically balance discrete commitment and soft exploration at each reasoning step. The second candidate represents the strongest alternative to the model's preferred continuation, so the separation between the two directly reflects the stability of the current discrete choice. When one candidate clearly dominates, CoTS keeps the representation close to the sampled discrete token to preserve a stable reasoning trajectory; as the competition strengthens, it increases the contribution of the soft representation to retain plausible alternatives. This competition-based allocation requires neither manually tuned switching thresholds nor mixing coefficients, and introduces no additional forward passes. Experiments across six reasoning benchmarks demonstrate that CoTS consistently outperforms standard CoT and state-of-the-art training-free latent reasoning methods. Our code is available at https://github.com/OkinoYu/CoTSReasoning.

---

### AI-Driven, Timing-Aware Power Optimization

Authors: Amit Sharma
Published: 2026-10-08
Categories: cs.AR, cs.LG
arXiv: https://arxiv.org/abs/2610.18890v1

Post-route power recovery changes cell choices under tightly coupled timing, electrical and physical constraints, and the value of each change depends on what it costs to repair its side effects. We propose an AI-driven framework in which a learned model drives both phases of the optimization: it ranks the powerrecovery edits to try and the repair actions that follow, while deterministic electronic design automation (EDA) analyses govern every acceptance. The framework records baseline constraint margins, identifies the damage a recovery transaction introduces in each analysis view, holds pre-existing violations at their baseline rather than working them inside the power step, and caps repair spending at a fraction of the measured saving. An ensemble of gradient-boosted models predicts power change, normalized damage reduction and constraint-regression risk from action and timing-context features available before each trial; calibrated, conservative estimates feed separate acquisition scores for recovery and repair, with a deterministic fallback when the model is uncertain. Recovery and repair remain speculative until physical implementation and fresh analysis establish baseline preservation and positive net saving. We specify the contract, the learned policy, the transactional flow and a pre-registered public evaluation protocol on open designs. Evidence from a companion study on public designs supports the contract; the learned policy is proposed, and no trained-model results or proprietary design measurements are reported.

---

### Bayesian Causal Forests for Zero-Inflated Continuous Outcomes

Authors: Hugo Gobato Souto, Eric Wang, Igor Shuryak
Published: 2026-10-08
Categories: stat.ML, cs.LG, stat.ME
arXiv: https://arxiv.org/abs/2610.18889v1

Estimating (heterogeneous) treatment effects is still challenging when the outcome is semicontinuous (i.e., exhibiting both a discrete point mass at zero and a skewed continuous distribution for positive values), despite the strong presence of this data type in real world settings. Existing methods either ignore the zero mass, impose fragile parametric assumptions, or suffer from extreme computational costs and data augmentation noise. This paper proposes Zero-Inflated Continuous Bayesian Causal Forests with Duan's Smearing (ZIC-BCF-Smear), a decoupled two-part nonparametric estimator that fits a Probit BCF on the full sample to model participation and a Gaussian BCF strictly on the active subpopulation to model positive intensity. Subpopulation Propensity Adjustment (SPA) mitigates regularization-induced confounding within the active subpopulation, and Duan's non-parametric smearing estimator re-transforms predictions to the response scale, bypassing the fragile scale multipliers of parametric models. A comprehensive simulation benchmark shows that ZIC-BCF-Smear matches well-specified parametric models without their assumptions and remains stable and calibrated where competitors either diverge or pathologically over-cover for nonlinear data. The method is then applied to individual-participant data from the SPECTRUM randomized phase III head-and-neck cancer trial, estimating the causal effect of assignment to panitumumab plus chemotherapy on the number of days patients spend with severe toxicity, an outcome for which a quarter of patients contribute an exact zero. Panitumumab assignment raises that burden by about ten days per patient under the trial's own differential treatment duration, and the decoupled structure localizes the effect mainly in the positive-intensity part rather than in whether the reconstructed severe-AE burden is positive.

---

### Embodied Zero-Shot Task Planning via Hardware-Aware LLM-VLM Integration for Grounded Robotic Manipulation

Authors: Lonnie Tankersley, Noorbakhsh Amiri Golilarz, Soroush Korivand
Published: 2026-10-08
Categories: cs.RO, cs.AI
arXiv: https://arxiv.org/abs/2610.14106v1

Large language models can translate natural-language instructions into plausible manipulation plans, but robotic execution requires physically grounded decisions that respect robot embodiment, workspace geometry, reachability, collision constraints, and grasp feasibility. This paper presents a hardware-aware large language model-vision-language model (LLM-VLM) manipulation framework for a simulated Franka Research 3 robot in NVIDIA Isaac Sim. The framework separates symbolic reasoning from metric execution: an API-hosted LLM planner receives a structured scene summary and generates only symbolic task information, including the object to manipulate, target relation, approach preference, and objects to avoid. A local Qwen3-VL 8B VLM verifier supports semantic confirmation of ambiguous objects and open-container surrogates. Both the LLM planner and the VLM verifier are prohibited from generating robot coordinates, grasp poses, keepout dimensions, gripper commands, trajectory waypoints, or success labels. Instead, these decisions are handled by deterministic grounding and execution modules that use the robot model, scene geometry, MoveIt 2 collision checking, inverse kinematics, grasp validation, release verification, and final visual assessment. Execution feedback is returned only as symbolic context for replanning. The framework is evaluated through simulated digital-twin trials involving pickup, occluded-source access, keepout reasoning, relation placement, and sequential manipulation tasks.

---

### Temporal Diffusion Bridge Model for Vessel Extraction Using Physics-informed Synthetic Data

Authors: Jinkui Hao, Donald R. Cantrell, Ramez Abdalla, Sameer A. Ansari, Bo Zhou
Published: 2026-10-08
Categories: eess.IV, cs.CV
arXiv: https://arxiv.org/abs/2610.14105v1

Accurate artery extraction from X-ray coronary angiography (XCA) is essential for coronary artery disease diagnosis and intervention. However, low vessel visibility, respiratory and cardiac motion, and limited annotations challenge existing methods. Moreover, most deep learning approaches produce binary vessel masks without preserving clinically relevant intra-arterial contrast information. We propose a Temporal Diffusion Bridge Model (TDBM) that formulates vessel extraction as image translation and learns from physics-informed synthetic data without manual vessel annotations. A physics-based XCA synthesis engine generates diverse, perfectly aligned pairs of contrast-enhanced and contrast-free images. Using these pairs, TDBM learns cross-domain mappings to recover high-fidelity vessel information. Temporal conditioning incorporates adjacent frames to improve consistency, while a vessel-aware contrastive loss exploits background-invariant synthetic pairs to enhance vascular discrimination. We evaluate TDBM on two XCA vessel segmentation datasets and conduct a reader study using two real-patient XCA sequence datasets. Zero-shot segmentation is further assessed on retinal imaging and fluorescence microscopy datasets. Results demonstrate that TDBM consistently outperforms existing methods and generalizes effectively across imaging modalities. Code and trained weights will be available at: https://github.com/Advanced-AI-in-Medicine-and-Physics-Lab/TDBM

---

### RePrior: Measurement-Reconciled Prior Correction for Sparse-View CBCT Reconstruction

Authors: Mengrui Yin, Tianxingjian Ding, Wufan Zhao, Ziyang Yan, Yu Tian
Published: 2026-10-08
Categories: eess.IV, cs.CV
arXiv: https://arxiv.org/abs/2610.14102v1

Sparse-view cone-beam computed tomography (CBCT) is severely under-constrained by limited X-ray projections. Learned reconstruction can provide complementary anatomical information, but using it as a prior in patient-specific Gaussian reconstruction introduces two challenges. First, the discrepancy between the learned prior and current reconstruction is spatially non-uniform: some local differences contain useful missing 3D information, whereas others reflect prior bias. Second, even a volumetrically beneficial correction can conflict with the observed X-ray measurements. Since X-ray measurements are line integrals, spatially separated corrections may affect overlapping rays, with their signed projection responses reinforcing or partially canceling in the measurement residual, so their feasibility cannot always be assessed independently. In this paper, we introduce RePrior, a multi-stage progressive prior-correction framework that addresses correction actionability, measurement feasibility, and representation expressivity. Selective Correction Proposal (SCP) first identifies actionable signed local corrections from prior--reconstruction discrepancies. Measurement-Constrained Reconciliation (MCR) then jointly enforces their consistency with patient-specific X-ray measurements, while Hybrid Residual Enhancement (HRE) recovers additional correction capacity through complementary voxel- and Gaussian-domain degrees of freedom around the reconciled anchor. RePrior requires neither test-time ground truth nor case-specific network fine-tuning. RePrior gains approximately 2.6~dB in 3D PSNR over R$^2$-Gaussian across LUNA16, PANORAMA, and PENGWIN. Beyond prior-assisted reconstruction, it delivers additional gains in volumetric quality and observed-view projection fidelity, with consistent gains in held-out-view fidelity in the backbone-matched evaluation.

---

### Fin_Agent at FinEval 2026: Task-Aware Methods for Multilingual Financial NLP

Authors: Mohotarema Rashid, Tanzina Hossain, Xiaoying Song, Lingzi Hong
Published: 2026-10-08
Categories: cs.CL
arXiv: https://arxiv.org/abs/2610.14101v1

We present Fin_Agent, our submission to the FinEval 2026 shared task on multilingual fi nancial language understanding. Our sys tems address Greek financial named-entity recognition, Japanese implicit communication recognition, and multilingual financial ques tion answering. For Greek NER, we com bine annotation-aware numeric rules with zero shot Qwen2.5-32B classification, achieving 0.9521 public F1 and 0.8884 private F1. For Japanese implicit communication, an inter pretable cue-based ordinal scorer achieves 98.0% test accuracy. For multilingual financial QA, BGE-M3retrieval with Qwen3-32B gen eration achieves 0.3023 and 0.3419 ROUGE-1 F1 on the public and private splits. Key limita tions include dataset-specific rules, handcrafted cues, and limited supervision. Implementa tion code is available https://github.com/ Rashidnira/finagent-fineval-2026.

---

### From Prototype to Massing: Learning Graph Correspondence for Scalable Facade Modeling

Authors: Yiliang Shao
Published: 2026-10-08
Categories: cs.LG, cs.GR
arXiv: https://arxiv.org/abs/2610.14098v1

Facade articulation at building scale is usually resolved manually or through parametric systems that encode design intent as explicit rules. This paper investigates a complementary workflow in which a designer-authored prototype supplies transferable articulation logic. Prototype and target massing are represented as heterogeneous attributed graphs, and a Siamese graph attention network with instance-local triplet supervision embeds their segments in a shared metric space. Many-to-one nearest-neighbor retrieval then routes prototype Patterns to the larger target. On the original 85/15 within-corpus holdout, prototype-to-massing top-1 accuracy reached 96.5%, compared with a reported 76.5% raw-feature baseline. A Rhino demonstrator converts predicted correspondence into reusable facade Patterns, and qualitative applications on four built projects show the workflow across distinct massing organizations. The contribution is not automated facade generation, but an authoring model in which resolved geometry serves as a computational specification for project-scale propagation.

---

### A Physics-Informed Hybrid Forward-Inverse Machine Learning Framework for High-Dimensional Sparse Recipe Reconstruction

Authors: Md. Muqtadir Fuad, Md. Istiak Adnan, Ridwan Al Aziz
Published: 2026-10-08
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.14097v1

High-dimensional sparse recipe formulation in process industries involves nonlinear association of various covariates, bounded by sparsity constraints and multi-criteria tolerances. Traditional physics-based classical methods (such as Kubelka-Munk) fail to generate sparse candidate recipes properly due to their simplified assumptions and industrial process variations. This study proposes a hybrid physics-informed uncertainty-aware framework that is structured into three modules: (a) Module A, a multi-target forward quality model; (b) Module B, a sparse inverse reconstructor; and (c) Module C, a feasibility-constrained ranking layer. Then it has been validated in an industrial textile dyeing context with recipe-grouped cross-validation under multi-illuminant quality assessment comprising process and production information, target reflectance, K/S values, and spectral descriptors. Histogram Gradient Boosting (HGB) with spectral-only features ($R_λ$, K/S) and Gradient Boosting Regressor (GBDT) with recipe-process features outperforms other configurations with out-of-Fold MAE of 0.1655 and 0.1523 in the forward quality model. For inverse reconstruction, a component-wise median blend achieved a recipe-vector MAE of 0.0295. Exploratory conditional treatment effect (CATE) suggests an observational association that alkali adjustment reduces error for red and pale shades (CATE = -0.196). This framework provides a human-in-the-loop decision-support tool that minimizes costly laboratory trials across industrial formulation domains.

---

### Scenes as an Attack Surface: Uncovering Scene-Grounded Safety Failures in MLLMs

Authors: Tongxi Wu, Yuyang Cheng, Jiacheng Xian, Hu Huang, Zhihao Wu, Menglin Yang, Jian Zhang, Yang Gao
Published: 2026-10-08
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.14094v1

Existing jailbreak attacks have exposed serious safety weaknesses in multimodal large language models (MLLMs). Most prior attacks rely on specially constructed visual inputs to carry or support harmful intent. In this paper, we uncover a different vulnerability: task-relevant scene semantics alone can shift an MLLM from refusal to compliance, even when the visual input introduces no additional attack instructions and the original task intent remains unchanged. We formalize this failure as a scene-grounded visual safety vulnerability and introduce SceneSafetyBench to investigate it across eight risk categories and nine white- and black-box MLLMs. Our analysis reveals that this failure is associated with how task-relevant visual evidence influences downstream safety representations, suggesting that visual grounding alone cannot fully explain the observed safety shift. Building on this finding, we propose Scene-Grounded Attacks (SGA), which preserve the original task while constructing validated scene-grounded queries and combining their independent responses. Across nine target MLLMs, SGA achieves an average attack success rate (ASR) of 86.17%, outperforming the strongest baseline by 8.55 percentage points and achieving the highest ASR on eight models. Our results identify scene semantics as an overlooked attack surface in MLLMs and highlight a growing tension between stronger visual grounding and robust multimodal safety alignment.

---

### Human Pose Estimation by Probabilistic Mixtures of Gaussian and Uniform Distributions

Authors: Santiago Reina Alguacil, Ezequiel López Rubio
Published: 2026-10-08
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2610.14093v1

Despite advancements in sub-pixel decoding, modern Human Pose Estimation (HPE) architectures remain constrained by deterministic coordinate extraction from discrete heatmaps, which systematically fails under dense occlusions and Out-of-Distribution (OoD) ambiguities. To address these limitations, we propose a post-hoc continuous probabilistic framework that intercepts raw heatmaps and models them as a Gaussian Mixture Model (GMM) augmented with a spatially agnostic uniform distribution, requiring no architectural modifications or retraining. This continuous representation allows us to comprehensively explore the distinct contributions of density modeling to HPE. We show that the spatial expectation of our fitted GMM achieves precision parity with leading sub-pixel decoders without heuristic post-processing. Leveraging this continuous space, our test-time augmentation aggregates probability distributions rather than discrete activations, mitigating multi-scale heatmap poisoning. Furthermore, because our multimodal GMM parameterization isolates competing spatial hypotheses, it naturally interfaces with geometric regularizers such as Markov Random Fields. However, our analysis exposes the inherent vulnerabilities of 2D kinematic priors under perspective foreshortening and visual degradation. Crucially, this density formulation yields significant improvements in uncertainty quantification: our calibrated geometric covariance volume outperforms standard peak-intensity heuristics in OoD anomaly detection under severe occlusion. For in-distribution calibration, dynamically fusing our geometric metric via Adaptive Softmax reduces or matches baseline Area Under the Sparsification Error (AUSE) across benchmarks and degradation levels. Finally, we establish that this continuous parameterization is extensible to multi-view 3D architectures, bypassing the $\mathcal{O}(N^3)$ voxel memory bottleneck.

---

### Pristine Models

Authors: Anirudh Sundara Rajan, Krishna Kumar Singh, Richard Zhang, Yong Jae Lee
Published: 2026-10-08
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2610.14091v1

Training on broad, unfiltered data enables high-quality image generation, but can also teach models concepts that should not be generated; restricting the training data ensures safety, but reduces quality because examples containing unsafe concepts also provide useful knowledge that is not itself unsafe. Existing approaches largely address this trade-off post hoc through concept unlearning, which can lead to catastrophic forgetting as the number of concepts to remove grows. We instead ask whether unsafe concept knowledge can be compartmentalized during pretraining so that access to it can be disabled at inference. We introduce Pristine Models, a novel text-to-image generation architecture with an attention mechanism that encourages the separation of safe and unsafe knowledge into independently parameterized branches. The branch-specific parameters are learned during training, while the unsafe branch can be detached at inference. On ImageNet, Pristine Models enable safer generation while retaining much of the utility of full-data pretraining, achieving higher image quality and better text alignment than training only on safe data or applying post-hoc concept removal. We further scale our approach to CC12M, demonstrating that it remains effective in a substantially larger and more diverse open-world pretraining setting.

---

### SAGE: Efficient Protected Vision-Language Inference on Smart Glasses

Authors: Zhimin Li, Pan Wang, Jingxian Chen, Yuantao Tang, Qian Lou, Jingtong Hu
Published: 2026-10-08
Categories: cs.CR, cs.AI, cs.CE
arXiv: https://arxiv.org/abs/2610.14089v1

Smart glasses are a promising platform for vision-language assistants, but private split inference remains expensive: protecting visual representations before transmission does not remove the cost of on-device vision encoding, wide protected tokens, or hardware-unaware execution. We present SAGE (Smart-glasses-Aware Guarded Encoding), a hardware-aware framework for efficient protected VLM inference on smart glasses. SAGE keeps raw images and unprotected visual tokens on the trusted device, reduces visual input cost, executes shape-specialized INT8 graphs, and compresses protected representations into compact 256-dimensional INT8 packets before transmission. Unlike efficiency methods optimized only for nominal model size, SAGE selects its operating point using downstream answer quality together with measured latency, memory, and thermal behavior on real Mentra Live glasses. On a matched Qwen workload, SAGE reduces glasses inference latency by 39.18%, device-sensor temperature rise by 21.91%, and protected packet size by 96.56%, while retaining 85.62% of the protected reference's relaxed utility and comparable prompted containment. Direct glasses-to-phone tests deliver all 192 packets and reduce median packet-plus-acknowledgment time from 71.3 to 10.0 ms. The compact protected interface also transfers across four additional VLM families. These results show that private wearable VLM inference must be optimized jointly across representation protection, communication, and real-device execution rather than through model compression alone.

---

### SenseWeather: Multimodal Sensing for Robust Robotic Perception under Extreme Conditions

Authors: Fucai Ke, Sukai Huang, Junyi Hao, Kal Backman, Naim Rastgoo, Gholamreza Haffari, Hamid Rezatofighi
Published: 2026-10-08
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2610.14082v1

Robotic perception must remain reliable when extreme environmental conditions severely degrade the primary RGB observation. Although robotic platforms commonly carry complementary sensors such as depth and infrared, much of the modern perception stack remains RGB-centric, creating a mismatch between multimodal onboard sensing and pretrained downstream models. We study whether complementary sensing can be used upstream to recover a task-compatible RGB observation that preserves existing perception capabilities without modifying or retraining the downstream stack. To this end, we introduce SenseWeather, a controlled multimodal benchmark for robotic perception under severe rain, snow, dust, fog, and darkness, with registered RGB, depth, and infrared observations, clean references, and grounded high-level perception tasks. We further develop Fusion-to-RGB, with two parameter-efficient sensor-fusion realizations based on token-space fusion for a diffusion transformer and residual latent conditioning for an inpainting UNet. Across frozen open-vocabulary grounding and vision-language models, RGB+D+IR fusion improves all paired downstream comparisons over the corresponding RGB-only restorers, including AP50 gains across all ten detector comparisons. Importantly, these perception gains do not consistently follow conventional image-fidelity metrics, and increasing weather severity does not induce monotonic downstream degradation. These results show that complementary sensing can preserve reusable robotic perception under extreme conditions, while highlighting task-level perception robustness as an objective distinct from visual restoration quality.

---

### Two-Stage Multi-modal Learning for Dynamic Audio-Visual Navigation: From Physical Alignment to Scene Aggregation

Authors: Jia Li, Yinfeng Yu
Published: 2026-10-08
Categories: cs.SD, cs.AI, cs.RO, eess.AS, eess.SP
arXiv: https://arxiv.org/abs/2610.14077v1

Embodied Audio-Visual Navigation (AVN) requires an agent to locate sound targets by jointly leveraging visual and auditory cues. While prior work has achieved notable success with static sound sources, real-world scenarios often involve moving targets, giving rise to the more challenging problem of \textbf{Dynamic Audio-Visual Navigation (D-AVN)}. Existing approaches largely rely on semantic auditory features, which limits generalization to unheard sound categories and hinders fine-grained spatial reasoning. To address these challenges, we propose a novel \textbf{Two-Stage Multi-modal Learning} framework for D-AVN. In the Physical Alignment stage, the proposed \textbf{Intensity-Calibrated Geometry Encoder (ICGE)} performs physics-informed alignment of sound intensity with visual geometry to capture category-agnostic spatial correspondence. In the Scene Aggregation stage, a Transformer-based \textbf{Cross-Modal Scene Aggregator (CMSA)} dynamically integrates multi-modal features to construct a robust state representation for policy learning. Extensive experiments on Replica and Matterport3D demonstrate that our method achieves state-of-the-art success rates, while significantly improving generalization and navigation efficiency in complex environments and under unheard sound scenarios.

---

### One Model, Multiple Compute Budgets: Multi-Ratio ConceptMoE for Flexible Compute Across Turns

Authors: Yiyang Fu, Zihao Huang, Dongchen Han, Shukai Gong, Zhen Dong, Qiyang Min, Daquan Zhou
Published: 2026-10-08
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2610.14074v1

Large language models typically support different inference budgets through separate model sizes or fixed compute configurations. In multi-turn interactions, however, compute demand can vary substantially across turns. Using a high-compute model for every turn wastes computation, while routing among independent models may require rebuilding model-specific history states after switches. We therefore ask whether a single shared model can support multiple compute budgets that can be flexibly assigned across turns. We introduce Multi-Ratio ConceptMoE, which jointly supports token-to-concept compression ratios of 2, 4, and 8 (CR2/CR4/CR8) within a shared backbone. Lightweight full-token modules reduce compression-independent cost, ratio-specific boundary routers support distinct compression levels, and cross-ratio training enables mixed-ratio processing within the same context. Across PTBench and SFTBench, the three fixed modes outperform their compute-matched MoE baselines in all 6 comparisons while remaining competitive with independently trained fixed-ratio specialists. On CodeChat and CoSQL, mixed-ratio inference improves the accuracy--compute trade-off, outperforming the Large MoE by 5.87 and 2.68 percentage points while reducing FLOPs by 41.1% and 12.3%, respectively. Under the same turn-wise budget schedule, Multi-Ratio ConceptMoE further reduces FLOPs by 78.4% and 46.9% relative to routing among independent models. These results demonstrate that a single shared model can provide multiple compute budgets for flexible compute allocation across turns.

---

### Is the Future Predictable? Modeling Future Predictability for Reliable Temporal Reasoning

Authors: Yongyi Huang, Huaping Zhang, Wenyao Cui, Qiuchi Li, Lei Li, Baohua Zhang
Published: 2026-10-08
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2610.14073v1

Large language models (LLMs) remain unreliable when reasoning about the future. Existing models often overlook the difference among future states, committing to specific answers even when the queried facts are inherently volatile. To capture this challenge, we introduce ChronoRegime, a future-oriented temporal reasoning task that requires models to determine whether historical temporal evidence justifies commitment to a future answer. Furthermore, we propose Confidence-Aware Temporal Reasoning (CTR), a framework that estimates future predictability from temporal dynamics and regulates answer commitment accordingly. Rather than performing unconditional forecasting, CTR commits only when the reconstructed evidence supports the target future state. Across multiple LLMs, CTR improves strict overall accuracy by 6.7 percentage points and reduces hallucination rates by 18.4 percentage points on average over the plain QA baseline. These results demonstrate that explicit predictability assessment is an effective mechanism for safer and more reliable temporal reasoning under future uncertainty.

---

### LeWAM: End-to-End World and Action Modeling with JEPAs

Authors: Minghao Fu, Tavis Siebert, Eryk Halicki, Randall Balestriero
Published: 2026-10-08
Categories: cs.RO, cs.AI
arXiv: https://arxiv.org/abs/2610.12818v1

Joint-embedding predictive architectures (JEPAs) learn compact latent world models without pixel reconstruction. Existing JEPA pipelines typically train visual representations and action-conditioned dynamics before adding a control policy or solver. We introduce LeWAM, an end-to-end JEPA for control with: (1) a joint training objective over next latent state prediction and action flow matching, and (2) an isotropic Gaussian constraint on the latent state, enabling a single model to support direct policy execution and efficient planning by evaluating sampled actions with its learned latent dynamics. Compared with prior JEPAs in visual planning, LeWAM improves success rate from 28.6% to 89.7% on contact-rich tasks and from 10.9% to 46.2% on long-horizon tasks, and plans 32.3 times faster. On behavior cloning tasks in simulation and physical robots, LeWAM is also competitive with established methods. Interpretability analyses show that LeWAM learns control-relevant representations and action distributions effectively guiding test-time planning and control.

---

### Deploying RDMA for Distributed Inference and Training At Scale

Authors: Yibo Huang, Yi Cui, Yukai Huang, Huaping Zhou, Jiancong Tong, Zhaoyu Wu, Gen Miao, Haoran Wei, Boyan Ding, Zhuo Jiang, Jianxi Ye, Yibo Zhu et al.
Published: 2026-10-08
Categories: cs.LG, cs.DC, cs.NI, eess.SY
arXiv: https://arxiv.org/abs/2610.12815v2

RDMA can reduce communication overhead in distributed machine learning systems, but realizing these benefits in production requires more than replacing TCP. Variable-size messages, concurrent connections, bursty incasts, and service failures introduce challenges that hardware offload alone cannot address. This paper reports our experience deploying ByteExpress, a software communication layer over hardware RDMA, in ByteDance's parameter-server-based inference and training systems. We present three lessons from the deployment: efficient RDMA requires buffer management across application and transport layers; scalable communication requires coordinated management of connections, receive capacity, and outstanding data; and service availability requires recovery above the transport layer. ByteExpress combines an RDMA-aware allocator, zero-copy transfers, QP multiplexing, receive-credit flow control, a software send window, and RPC-assisted recovery. In production inference, ByteExpress reduces mean and P99 model parameter fetch latency by 35.7% and 43.7%, respectively, and total CPU usage by 10.69% compared with kernel TCP. In training, ByteExpress reduces P99 model parameter push latency by up to 74.0% compared with TCP, with no PFC pauses observed at runtime.

---

### VEPA: Verified Evolving Formal Policy for Lifelong LLM Agent Guardrails

Authors: Duy Nguyen, Lesly Miculicich, Bhavana Dalvi Mishra, Shuo Li, Krishnamurthy Dj Dvijotham, Mohit Bansal, Tomas Pfister, Long T. Le
Published: 2026-10-08
Categories: cs.AI, cs.CL, cs.CR, cs.LG
arXiv: https://arxiv.org/abs/2610.12813v1

LLM agents increasingly operate over sensitive resources, such as private documents and mutable system states, which can create serious safety and security risks. Runtime guardrails should therefore make inspectable, reproducible, and deterministic allow or refuse decisions before potentially irreversible agent actions are executed. However, safety and security requirements are rarely complete before deployment, so guardrails need to keep adapting as new risks emerge. Existing adaptive guardrails learn from deployment experience, but they usually apply safety policies through LLM inference, which can lead to inconsistent decisions and ambiguous policy interpretations. Formal guardrails, on the other hand, enforce policies deterministically, but they usually rely on pre-deployment or task-specific policies and have no mechanism to adapt to newly observed risks. We introduce VEPA, a framework for the verified evolution of formal agent guardrails. VEPA represents the guardrail as a formal policy of executable symbolic rules over typed agent events. During deployment, we use an LLM guided by deployment feedback to propose policy patches that fix newly observed failures while preserving previously correct behavior. Then, a deterministic verifier only accepts policies that are structurally valid and conflict-free. Finally, an evidence-aware authority layer activates a verified rule only when it has enough supporting evidence, so uncertain updates cannot directly affect guardrail decisions. Across ASSEBench-Safety, ASSEBench-Security, and ATBench, VEPA consistently outperforms baselines while expanding the scope of deterministic formal-policy enforcement. These gains also hold under more challenging deployment conditions, including mixed streams, corrupted feedback, and longer horizons.

---

### Spend Where It Matters: Particle Search with MH Refinement for Budget-Constrained Audio-Visual Alignment

Authors: Shivanshu Shekhar, Raghavendra Addanki, Tong Zhang, Uttaran Bhattacharya
Published: 2026-10-08
Categories: cs.CV, cs.SD, eess.AS
arXiv: https://arxiv.org/abs/2610.12811v1

Reward-guided inference-time alignment can improve joint audio-visual generation, but optimizing multiple alignment objectives requires many expensive model evaluations, making existing methods costly under limited compute. Particle-based search enables global exploration but incurs estimation error with a finite population, while Metropolis-Hastings (MH) can refine candidates locally but benefits from a strong initialization. We introduce \textbf{S}pend \textbf{W}here \textbf{I}t \textbf{M}atters (\modelname), a two-stage, training-free, reward-agnostic framework that combines particle search with MH refinement under a fixed compute budget. In the first stage, we adaptively allocate particles across denoising checkpoints using online variance estimates, minimizing estimation variance while concentrating compute where it is most useful. This search provides a promising candidate for MH refinement, which uses local noise-and-denoise proposals on completed outputs to further improve reward. By starting from a strong candidate, MH can recover additional reward without requiring a larger particle population or costly global exploration. We show that, at matched compute, combining particle search with MH refinement achieves higher rewards than either approach alone. On \textsc{JavisBench-Mini}, at closely matched compute, \modelname achieves mean relative gains across seven metrics of $30.7\%$ over BoN and $14.1\%$ over EvoSearch with JavisDiT++, and $9.2\%$ over BoN and $17.0\%$ over EvoSearch with LTX-2. Beyond joint audio-visual generation, we also demonstrate the benefits of adaptive particle allocation on SD1.5 image generation, improving AestheticScore by $2.2\%$ and ImageReward by $6.0\%$ over uniform allocation.

---

### The Expressive Power of Random-Feature MPNNs Across Graph Sizes

Authors: Robert R. Nerem, Mitchell Black, Antonis Vasileiou, Christopher Morris, Yusu Wang
Published: 2026-10-08
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.12809v1

Continuous random node features increase the expressivity of message-passing neural networks (MPNNs), but existing theoretical analyses are not fine-grained enough to capture their limitations, particularly in approximating functions across graph sizes or under a fixed number of random bits. We develop a population-level framework for size-uniform approximation. For MPNNs with a graph-size-independent Lipschitz bound in their initial features, including standard mean, max, and order-normalized sum architectures, we show that continuous random features can be quantized to finitely many colors with arbitrarily small uniform error in expected output. This reduction lets us transfer limitations on finite-color distinguishability to limitations on size-uniform approximation with continuous random features. We use this framework to prove that these MPNNs are not size-uniformly universal, analyze the expressivity gains from colors and independent trials, and establish incomparability between random-color MPNNs and higher-order Weisfeiler-Leman methods. We also give finite-color MPNN constructions for tree subgraph isomorphism via color coding and a local MST task.

---

### Trainable Quantum-Native Language Models: Dissipative Memory, Capacity Limits, and Hardware Validation

Authors: Deniz Yunus Göğüş
Published: 2026-10-08
Categories: quant-ph, cs.LG
arXiv: https://arxiv.org/abs/2610.12806v1

We study byte-level autoregressive language models whose hidden state is a small quantum register, whose transitions are learned token-conditioned quantum channels, and whose next-byte probabilities come directly from Born-rule measurements without a learned classical vocabulary projection or softmax. In the fixed-width recurrent setting, unitary dynamics cannot contract or reset stored distinctions; experimentally, unitary recurrence loses to its retrained one-token control in every tested variant. Introducing learned dissipation, then token-conditioned dissipative gating, yields useful memory beyond one token. The gated model improves test negative log-likelihood by 0.20 +/- 0.03 nats/byte over retrained parameter-matched one-token controls at context length 12, and by 0.29-0.33 at longer contexts, but never beats retrained two-token controls. A key-value recall benchmark shows a width-dependent threshold: one stored pair meets the pre-registered accuracy bar at 4-5 memory qubits, two pairs at 6-7, and three pairs remain unsolved. This is consistent with, but does not establish, a capacity explanation for the short effective memory horizon. As a secondary result, one-qubit models reach lower stationary memory entropy than a heuristically optimized classical predictive frontier at matched error in 9/9 seeds. On IBM Heron hardware, computational-basis memory-register distributions show signal against a uniform null in 12/12 circuits across prefixes of up to 12 tokens, next-bit probabilities pass pre-registered accuracy bars at four-token prefixes, and bytes are generated autoregressively by sampling physically measured bit probabilities. The tested models are classically simulable and substantially worse than classical n-gram baselines. We claim no computational quantum advantage.

---

### People Making Conversation: Evaluating How Naturally Speech Models Converse

Authors: Arya Bastani, Iman Modarressi
Published: 2026-10-08
Categories: cs.CL
arXiv: https://arxiv.org/abs/2610.12802v1

A conversational speech model has to decide, at every pause, whether to speak. Existing evaluations test boundary detection, policy adherence, or aggregate statistics, not this decision. We place real-time speech systems at 427 pauses and 482 turn ends from 176 unscripted conversations in Seamless Interaction, give each the conversation so far, and score what it generates against what the person did: whether and how soon it speaks at a pause; its latency, length, and wording at a turn end; and whether it can answer a question about the preceding 30 to 120 seconds. Across six systems, the hosted APIs (GPT Realtime, Gemini 3.1 Flash Live, Grok Voice) respond at almost every pause whichever endpointer they use; only their latency, about 1 to 4 s with silence-based endpointers and about 8 s with semantic turn detection, decides whether the words land inside the pause. The full-duplex models (PersonaPlex, Moshi) are the only systems that regularly stay silent and the only ones that interrupt within a human reaction time. Hosted systems reply longer than the person. Given the history as text, hosted systems recall 41 to 60% of conversation facts under a Claude Opus 5.5 judge and 19 to 26% under a strict word matcher; hearing it as audio, full-duplex models recall 17 to 22% and 12 to 15%. We release the labels, points, harness, and every output.

---

### Polyconvex neural-network surrogates with learned paired directional features for anisotropic hyperelastic homogenization

Authors: S. Ares de Parga, A. Cornejo, J. A. Hernández, R. Rossi
Published: 2026-10-08
Categories: physics.comp-ph, cs.LG, math.NA
arXiv: https://arxiv.org/abs/2610.12801v1

Approximating the effective anisotropic hyperelastic response of a representative volume element (RVE) requires constitutive expressiveness, mechanical consistency, and computational efficiency. This work develops a polyconvex physics-augmented neural network (PANN) construction based on paired directional features whose orientations, stretch powers, and determinant exponents adapt to the RVE response. The features are composed with convex, componentwise nondecreasing cores, realized as an input-convex neural network (ICNN) or an input-convex Kolmogorov-Arnold network (ICKAN). The complete energies satisfy objectivity, an exactly stress-free reference, two-dimensional polyconvexity, and prescribed growth conditions by construction, and nonnegative energy is certified separately for the selected models. On a multicavity and a single-cavity periodic cell, these energies are compared with an Unconstrained energy, trained identically without the curvature and growth constraints, and with three hyperreduced projection-based reduced-order models (HPROMs), which serve as physics-based references. On the first cell, learning the features lowers the stress error sevenfold, and six features suffice on both. The constrained energies remain below one percent in stress error and in the displacement error of a tensile-coupon FE$^2$ application with a full-order reference, whereas the Unconstrained energy and the HPROMs are more accurate inside the data. Beyond the data, however, the Unconstrained energy loses rank-one convexity, and under confined compression it stops below half of the 500 MPa that the constrained energies reach. These results quantify what the guarantees cost in accuracy and what they secure beyond the data.

---

### Efficient Audio-Visual Speech Enhancement via Reinforcement Learning-Based Neural Architecture Search

Authors: Khubaib Ahmed, Ahsan Adeel, Nasir Saleem, Kia Dashtipour, Ahsan Ulhaq, Amir Hussain
Published: 2026-10-08
Categories: cs.SD, cs.AI, eess.AS
arXiv: https://arxiv.org/abs/2610.12796v1

Audio-visual speech enhancement (AVSE) improves speech intelligibility in noisy environments but often relies on computationally expensive models unsuitable for resource-constrained edge devices. This paper introduces a reinforcement learning-driven neural architecture search (RL-NAS) framework to discover compact, efficient AVSE architectures. The search space preserves temporal alignment between audio and visual streams, while a novel reward function jointly optimises speech enhancement performance, measured by SI-SNR improvement, and model compactness. Using proximal policy optimisation with constraint-aware action masking, the framework evaluates only 35 candidate architectures and identifies a model with 2.9 million parameters, compared with 22 million in the baseline. The proposed architecture achieves a 14.5 dB SI-SNR improvement, only 0.7 dB below the baseline, while reducing model size by approximately nine-fold. Evaluation on commercial mobile CPUs demonstrates a 4.3-fold reduction in real-time factor, supporting practical on-device deployment. Compared with random search, RL-NAS achieves 92% higher search efficiency. Architectural analysis reveals that shallower but wider temporal convolutional blocks and aggressive visual pathway pruning can maintain enhancement performance while substantially reducing computational requirements. These findings demonstrate the potential of reinforcement learning-guided architecture search for developing lightweight, efficient and deployable multimodal speech enhancement systems.

---

### Meta-Reflective Evolution Effectively Improves Production AI Systems

Authors: Seung Won Wilson Yoo, Chang Zeng, Aditi Somayajula, Sahithi Ankireddy, Rohan Kosalge, Avishree Khare, John Mitchell, Aditya G. Parameswaran, Rajhans Samdani, Anupam Datta
Published: 2026-10-08
Categories: cs.LG, cs.AI, cs.CL, cs.SE
arXiv: https://arxiv.org/abs/2610.12794v1

Large Language Model (LLM)-guided evolutionary optimization has successfully improved prompts, programs, and even entire AI systems. Most optimization methods select a candidate from the search frontier, then mutate it in one LLM pass based only on its failures. This focus on the frontier can stall on production AI systems, where isolated failures reveal little about complex, hidden interactions. We introduce CoCoEvolve (Cortex Code Evolve), an evolutionary optimizer built around what we call meta-reflection, wherein an agentic mutator (Cortex Code) selectively retrieves evidence from the entire evolutionary history and probes the live system before verifying and proposing mutations. We apply CoCoEvolve to optimize three classes of production AI systems: data agents, LLM-powered data pipelines, and data engineering workflows. On DABStep, CoCoEvolve tops the public leaderboard, reaching 92.8% held-out hard accuracy compared to 78.4% for Meta-Harness and 30.4% for GEPA under matched conditions. Across six LLM data pipeline workloads, it outperforms every evaluated baseline. On data-eng-bench, CoCoEvolve lifts test-set solve rate from 59.1% to 77.4%, past the best baseline of 73.6%. It also has 11 percentage points higher accuracy than GEPA on HotpotQA while remaining competitive with specialized optimizers on established mathematical and systems benchmarks. Matched ablations confirm that each component of CoCoEvolve contributes to these gains, demonstrating the value of meta-reflection for optimizing production AI systems.

---