# Research Papers — 2026-09 · bundle page 56 of 266 > Full-text markdown of papers #2751–2800 of 13298 published in 2026-09, newest first. AI-readable mirrors of the full 13298-paper 2026-09 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-09/page/56 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/55.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/57.md) ### Resource-Aware Parameter-Efficient Model Adaptation for Onboard High-Dimensional Data Authors: Qiyang Zhang, Xinhao Li, Lei Shi, Zheng Lin, Jinfeng Wen, Ao Zhou, Shangguang Wang Published: 2026-09-27 Categories: cs.CV, cs.DC, cs.NI arXiv: https://arxiv.org/abs/2609.33687v1 Onboard satellite models often require frequent updates, but the weights adapted to earlier data distributions can quickly become outdated. However, updating large-scale model parameters in orbit presents significant challenges due to the limited uplink bandwidth of Low Earth Orbit (LEO) satellite systems, particularly for hyperspectral satellite imagery, where high-dimensional spectral-spatial inputs lead to increased model size and update costs. Existing full fine-tuning methods are thus expensive to retrain and difficult to deploy under strict communication constraints. To address this challenge, we propose NE-LoRA, a parameter-efficient adaptation framework for bandwidth-constrained onboard hyperspectral model updates. NE-LoRA combines a primary low-rank branch with a nonlinear auxiliary branch to capture both global update trends and complex spectral-spatial variations. Additionally, we introduce a differentiated training strategy for multi-matrix adapters, motivated by the asymmetric initialization and gradient dynamics of different adapter matrices. Experiments on four hyperspectral datasets and three representative backbone models demonstrate that NE-LoRA consistently outperforms LoRA-based baselines and remains competitive with, and in several cases superior to, full fine-tuning. Across the evaluated settings, NE-LoRA updates only a small fraction of the total parameters on average while preserving low deployment overhead, offering a favorable accuracy-communication trade-off for onboard hyperspectral adaptation. --- ### T-MoXAI: A Hierarchical Explainability Framework for Temporal Multimodal Data Authors: Ali Inha, Mo Vali, Saaliha Vali, Pietro Liò, Meen-Yau Thum Published: 2026-09-27 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.33685v1 Artificial Intelligence (AI) models for temporal multimodal data have potential in healthcare and agriculture, but their opacity can limit trust and adoption. We introduce T-MoXAI (Temporal Multimodal eXplainable AI), a hierarchical framework explaining (1) when timepoints influence predictions, using temporal Shapley values; (2) which modalities contribute at those moments, using attention analysis; and (3) what features or image regions drive decisions, using gradient based attribution. A transformer based architecture handles irregular temporal sequences and heterogeneous data, generating all three explanation levels in under one second for interactive decision support. We evaluate the framework on two real world tasks: predicting IVF treatment outcomes from ultrasound sequences and clinical measurements (AUC 0.660 despite significant class imbalance), and forecasting wheat yield from temporal RGB imagery and phenotypic traits ($R^2$ 0.265 amid substantial environmental variability). Ablation studies indicate that temporal modelling is critical in both domains: removing it reduces performance to the equivalent of random guessing. Temporal ROAR experiments provide evidence that the explanations reflect the model's reasoning process. With a unified, domain agnostic architecture and open source implementation, T-MoXAI provides a baseline for temporal multimodal XAI, addressing fragmentation in the field and supporting applications where understanding decisions is as important as predictive accuracy. --- ### MAD-Guard: Controlled Study of Autoregressive Generation versus Direct Decision Interfaces for Closed Multimodal Forensic Tasks Authors: Hao Chen Published: 2026-09-27 Categories: cs.CV, cs.CR arXiv: https://arxiv.org/abs/2609.33683v1 When should multimodal foundation models generate tokens, and when should they directly output a decision? We present MAD-Guard, a controlled study of output-decision interfaces for closed multimodal forensic tasks. Once a multimodal representation is computed, is autoregressive generation necessary for closed forensic decisions with high input complexity but low output entropy? Under a matched Qwen3-VL-8B backbone, 2,400 FakeClue training samples, and LoRA budget ($r=16, α=32$) on Huawei Ascend 910C NPUs, we evaluate a progression of decision interfaces (AR-SFT (generate) $\to$ Logit Slice $\to$ Binary Direct Head $\to$ +choice $\to$ +act $\to$ CLM-Head) and decompose latency into backbone representation (53.12 ms), 151,643-way vocabulary projection (+85.04 ms $\to$ 138.16 ms), and decoding (+248.26 ms $\to$ 386.42 ms). Under 1-to-1 binary supervision ($\mathcal{L}_{\mathrm{BCE}}$), a Binary Direct Head cuts latency by $2.60\times$-$7.27\times$ (53.12 ms) and lowers calibration error by $1.88\times$ (ECE = 0.0450 vs. 0.0845), with a -1.80% accuracy trade-off (93.10% vs. 94.90%; 0.9795 vs. 0.9871 ROC-AUC) from forfeiting token priors. Gains above AR-SFT arise either from multi-task attribution and uncertainty gating (+choice+act: 96.44% accuracy, 0.9940 ROC-AUC, 0.0187 ECE at 53.71 ms) or from a disaggregated contrastive head (CLM-Head: 96.55% binary and 96.44% multi-task accuracy, 0.0166 ECE, 98.79% 7-class attribution at 54.42 ms) retaining semantic priors without token decoding. Across 5,000 out-of-sample images from five benchmarks, our framework excels on synthetic, camouflage, and document forgeries (96.44% GenImage, 97.73% Chameleon, 91.84% Doc) while showing a clear boundary on compressed face manipulation (FF++ ROC-AUC = 0.5913). --- ### SWE-Game: Can Coding Agents Build the Games We Want? Authors: Xiaoyu Chen, Lai Wei, Jin Wang, Xiangyu Zou, Ruochen Fan, Enze Luo, Mingzhe Yao, Jiahui Zhu, Yuhua Wen, Linghe Kong, Weiran Huang Published: 2026-09-27 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.33678v1 We introduce SWE-Game, a benchmark of 247 tasks grounded in 41 executable reference Godot games spanning 13 gameplay categories in 2D and 3D. Five task types cover development from a brief, implementation from a game design document, skeleton completion, repair of 83 injected-fault cases, and Godot-to-Unity porting. Reference materials specify the intended gameplay, while a shared instrumentation interface lets evaluator-owned drivers and probes execute actions and observe independently implemented games. Evaluation combines engine-state checks, certified reference-input replay, and agent-authored feature demonstrations to assess mechanic correctness, demonstrated playability, and behavioral restoration and preservation after repairs. Game-specific vision-language rubrics separately assess presentation. Across six models, Opus5 achieves the highest overall score in all five task types. Best overall scores remain below 60 out of 100 across the three construction tasks, with Brief-to-Game reaching 50.38. Analysis of reviewed submissions identifies requirement omissions and gameplay logic errors as predominant implementation problems. On human-labeled behaviors from 100 agent-built games, executable checks achieve 92.59% balanced accuracy, compared with 78.41% for a video-based VLM judge. Rubric-based visual scores reach a Spearman correlation of 0.829 with human ratings of 200 gameplay clips. Together, these results characterize current agent capabilities across game-development activities and support combining runtime evidence with visual assessment. --- ### Auditing Agent Actions through Query-Conditioned Attribution Authors: Yifan Liu, Praveen Venkateswaran, Abdulhamid Adebayo, Dong Wang Published: 2026-09-27 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.33676v1 LLM agents increasingly take consequential actions through interactions with users, policies, and external tools. Auditing these agents requires automated attribution of realized actions to their historical basis. However, existing attribution formulations do not provide question-specific traces for diverse auditing objectives. Additionally, when access to the acting model is limited (e.g., in API-only deployments), applicable methods commonly rely on costly input perturbations or external LLM analysis of complete trajectories. We therefore formulate $\textit{query-conditioned agent action attribution}, a new task that takes a natural-language auditing query as input and recovers the source and ordered intermediate evidence for the query-specified aspect of an action. We instantiate this task with $A^3Bench$, a benchmark comprising 1,396 auditing queries across policy basis, parameter provenance, failure propagation, and unsafe-behavior tracing. To enable efficient, query-specific attribution, we use small open-weight models as attribution proposers that combine query-conditioned gradient saliency with query-semantic relevance to rank history units. Our proposer consistently achieves stronger source and evidence rankings at lower inference cost than open-weight baselines, improving source MRR by up to 40.9\% and evidence MAP by 42.1\% with only two forward passes and one backward pass. Controlled evaluations confirm that our proposer improves attribution specificity by adapting its rankings to fine-grained changes in the auditing query. Building on a proposer ensemble, our end-to-end system surpasses the strongest frontier-model baseline in source accuracy (64.5\% vs.\ 60.4\%) while reducing empirical deployment latency by 29.9\% relative to the fastest frontier API baseline. Code and data will be released after the initial review period following final validation and cleanup. --- ### Benign Overfitting for General Norms and Distributions Authors: Daniel Barzilai, Ohad Shamir Published: 2026-09-27 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.33675v1 Understanding why predictors can generalize despite interpolating noisy training data is a central puzzle in machine learning. Most work on such "benign overfitting" studies minimum-2-norm linear regression, reflecting the inductive bias of gradient descent. However, modern optimizers such as Adam and Muon use non-Euclidean update geometries, favoring solutions associated with other norms. Analyzing regression for non-Euclidean norms is substantially more difficult, with known results essentially limited to Gaussians. In this paper, we develop a method to analyze benign overfitting in linear regression for general norms and general (sub-Gaussian) distributions. As a special case, we prove that minimum-p-norm interpolation with p>1 can benignly overfit even for non-Gaussian distributions, under suitable conditions. Perhaps surprisingly, for the 1-norm, benign overfitting does not hold in general for well-behaved (but non-Gaussian) distributions, showing that existing positive 1-norm results rely crucially on Gaussianity. Our proof analyzes the geometry of the dual optimization problem, using concentration and central limit tools to show it is approximately Euclidean in many high-dimensional cases. --- ### Reachability is not enough: Diagnosing long-range behavior in GNNs Authors: Filippo Maria Bianchi Published: 2026-09-27 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.33674v1 Graph neural networks (GNNs) are often called long-range because their architecture can connect distant nodes, but this does not show whether they use distant information correctly. We introduce a framework that measures how strongly inputs at each graph distance affect predictions and separates limitations due to architecture, finite approximation, training, and numerical execution. Our analysis shows that local message-passing can spread influence slowly, so a finite implementation may rely mainly on nearby inputs even when the ideal computation uses the whole graph. We also explain why mathematically equivalent filters can differ in how easily they are learned and how reliably they run. Across controlled tasks, models with similar architectural reach use distant information very differently, while low average error can hide failures on distant interactions. Together, these results show that long-range capability depends on learning to use information at the distances required by the task and preserving that use during computation. --- ### Reset Is Not Recovery: Evaluating Recoverability from False Conversational Context via Sycophancy Hysteresis Authors: Adi Shnaidman Published: 2026-09-27 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.33672v1 Grounded language models are usually evaluated by adding relevant context, but multiturn dialogue also contains unsupported user claims that may contaminate later factual answers. We study post-pressure recoverability: whether a model returns to clean-context behavior after a user repeatedly advocates a wrong answer and then withdraws that pressure. We introduce a recovery-after-pressure protocol for multiple-choice factual dialogue and measure sycophancy hysteresis, the residual probability assigned to the user-advocated wrong answer relative to a clean-context counterfactual. Across seven instruction-tuned open-weight models and two factual benchmarks, ordinary reset often reduces but does not erase pressure-induced bias. History preserving repairs such as user retraction, system reset, and self-verification recover only 2-3/14 model-dataset pairs under the strict clean-restoration diagnostic, whereas operations that change the effective context are substantially more reliable; the two conditions that remove the pressure-bearing history entirely, fresh-context deletion and context truncation, recover 14/14. In an oracle trusted-evidence condition across fourteen model-dataset pairs, preserving the pressure-bearing history while adding benchmark-derived trusted evidence increases accuracy from 0.368 to 0.929, while wrong-answer following falls from 41.2% to 4.3%. Controls show that the effect is not explained by dialogue length, repeated confidence, plausible distractors, mere false-answer mention, or option-label inertia. These results suggest that faithful grounded dialogue requires evaluating which prior context should be treated as evidence and which should be removed or quarantined before answering. --- ### COGNIT-Guard: Calibrated Standalone Direct-Decision Guardrails with Heterogeneous CPU-NPU Confidence Cascading under Explicit Latency and False-Positive Constraints Authors: Hao Chen Published: 2026-09-27 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.33671v1 When must a foundation-model safety gateway generate tokens, and when should it directly output a calibrated decision? We study calibrated standalone direct-decision foundation models for real-time pre-ingestion safety guardrails, jointly addressing probability calibration, dual-use false-positive control, and heterogeneous CPU-NPU routing under explicit latency SLOs. Pre-ingestion guardrails must screen prompts prior to target-LLM prefill with low false alarms on benign compliance inquiries; however, shallow classifiers are brittle to phrasing shifts, hidden-state probes require coupling to a target LLM, and generative guards incur high decoding latency and dual-use false positives. We present COGNIT-Guard, coupling a validation-calibrated CPU fast gatekeeper with confidence-gated escalation to an NPU-resident 322M bidirectional direct-decision model (Laya-322M) under an asymmetric false-positive penalty. On the clean unseen DUCS-Bench test split ($N=607$), COGNIT-Guard achieves 98.85% accuracy (McNemar $p = 1.19 \times 10^{-4}$ vs. ML), reduces benign FPR to 0.42% ($1/238$; Fisher's exact $p = 8.23 \times 10^{-4}$ vs. ML), and attains 1.12% ECE and 0.0104 Brier score. On Huawei Ascend 910C NPUs, pure NPU inference runs in 21.77 ms mean latency (45.90 QPS), while the live serial CPU-NPU cascade ($θ^*_{\mathrm{deploy}}=0.70$) achieves 41.63 ms mean latency (P50: 39.47 ms, 99.23% accuracy, 0.00% FPR). Evaluation on SafetyBench-ZH ($N=2,100$) and comparison against a bi-encoder direct-decision baseline (CLM-8B) disentangle in-domain gains, OOD alignment tax (60.33% $\to$ 56.81% on Laya; 55.10% on domain CLM-8B), and experience replay recovery, restoring OOD accuracy to 64.10%-65.05% and reaching 99.67%-99.84% in-domain accuracy with 0.00%-0.42% FPR. --- ### Closing the Cross-Dialect Gap: Query Plans as a Portable Interface in Text-to-SQL Authors: Corentin Royer, Robin Oester, Yotam Perlitz, Yannick Metz, Andrea Giovannini, Mennatallah El-Assady Published: 2026-09-27 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.33670v1 Text-to-SQL systems are typically trained and evaluated on a single dialect (SQLite), yet production deployments span PostgreSQL, MySQL, ClickHouse, and beyond. We show that this single-dialect assumption leads to a substantial drop in cross-dialect accuracy for every model we tested. The drop persists across scale, architecture, and even purpose-built text-to-SQL systems. We argue that the fix is to change the generation target: instead of asking an LLM to emit dialect-specific SQL, we have it emit a dialect-agnostic relational algebra query plan, which a deterministic compiler then renders into SQL for any supported backend. Across thirteen models from 3B to frontier scale, this restores cross-dialect portability nearly uniformly, at a small cost in peak accuracy on the model's home dialect for capable prompted models and none once fine-tuned on plans; under matched fine-tuning, plan supervision yields a stronger model than SQL supervision. We also introduce MetricName, a question-aware result-set comparator needed to evaluate fairly across dialects, where existing metrics confound semantic errors with benign cross-dialect variation. More broadly, the result is a reminder that a generation target chosen for execution is not necessarily the one that maximizes generation quality. --- ### RSD-Poker: Structure-Adaptive and Shift-Robust Risk-Utility Certification for Residual Policies in Imperfect-Information Games Authors: Miaobo Hu, Shuhao Hu, Xiaobo Guo, Xin Wang, Bokun Wang, Peng Zhang, Daren Zha, Jun Xiao Published: 2026-09-27 Categories: cs.AI, cs.GT, cs.LG arXiv: https://arxiv.org/abs/2609.33669v1 Residual policy adaptation provides a lightweight way to modify a strong reference policy, but a shared scale and a fixed subgroup partition can hide heterogeneous degradation and become fragile when the deployment mixture of information states changes. We introduce RSD-Poker, a structure-adaptive and shift-robust certification framework that freezes a bank of residual families and scales, learns a policy-visible partition on an independent structure split, and freezes that partition before calibration labels are joined. Each candidate-group pair receives a weighted simultaneous upper certificate for anchor-relative risk and a lower certificate for weak-response utility. A robust group-to-candidate map is then selected over a predeclared uncertainty set of deployment group proportions. Under independent calibration units drawn from each frozen group's law, a candidate bank and partition fixed before calibration, and invariant within-group conditionals, the selected map satisfies its declared mixture-robust risk budget and utility certificate with probability at least $1-ζ_{risk}-ζ_{util}$. The information contract supports both a teacher-backed transform and a teacher-free observation-only student. The retained deterministic 24-state audit remains an exact replay diagnostic: empirical-zero selects $α=0.08$, raising the weak-response proxy from 4.2082 to 4.2889 with $0/12$ held-out threshold crossings. On stratified held-out states, the learned-partition dual selector raises weak utility from 4.4074 under global dual certification to 4.4936 and lowers held-out violation from 0.0215 to 0.0078; its mixture-robust variant reaches violation 0.0059. Across five observation-only checkpoints, risk-calibrated residuals attain weak utility $4.3659\pm0.0177$ and violation rate $0.0178\pm0.0057$. --- ### When Noise Meets Long-Tail: Feature-Threshold Dual Calibration for Robust Pseudo-Labeling Authors: Ping Guo, Zhiqi Huang, Xinran Li Published: 2026-09-27 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.33668v1 Pseudo-labeling has become a cornerstone of learning from unlabeled data in semantic segmentation. Yet its effectiveness drops sharply in real-world scenarios where strong imaging noise and long-tailed class distributions occur together. We trace this failure to a vicious cycle of pseudo-label degradation. Imaging noise entangles foreground and background features, lowering prediction confidence across all classes, while long-tailed distributions leave tail classes with far fewer training samples and inherently lower confidence. Under fixed high-threshold filtering, these tail-class predictions are systematically filtered out, so they receive no supervision from unlabeled data and thus features keep degrading in subsequent iterations. Critically, noise and long-tail are not independent obstacles but mutually amplifying ones, and addressing either alone is insufficient. To break this cycle, we propose FTC-Seg, a Feature-Threshold dual-Calibration framework built on a standard teacher-student framework. At the feature level, Orthogonal Prototype Reconstruction (OPR) uses a set of learnable orthogonal prototypes to residually purify pixel-wise features, widening the margin between weak foreground targets and noisy backgrounds. At the threshold level, Adaptive Threshold Calibration (ATC) dynamically adjusts class-specific thresholds based on learning difficulty and prediction-distribution bias, rescuing low-confidence pseudo-labels of tail classes from systematic exclusion. Extensive experiments on four public benchmarks spanning three distinct noise modalities show that FTC-Seg achieves strong performance against state-of-the-art methods, with particularly substantial gains on tail classes. Our results establish that jointly calibrating features and thresholds is essential for robust pseudo-labeling under compounded noise and class imbalance. --- ### Scalable Attribution and Control of Model Behavior During Training Authors: Sleem Abdelghafar Published: 2026-09-27 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.33667v1 Attributing and controlling model behavior during training requires identifying each example's contribution quickly enough to act before the next update. However, examples in the same training batch can produce similar behavioral changes, making their individual contributions difficult to distinguish. We address this ambiguity through mutual information, accounting for interference within the batch by quantifying how much the combined behavioral change reveals about each example's contribution. We show that this mutual information is a logarithmic function of Behavioral Gradient Uniqueness (BGU). BGU gives the information measure its geometric interpretation. Our Batch-Space Ghost (BS-Ghost) algorithm makes these scores practical inside the training loop through shared computation in batch space, without storing model-sized example gradients. On a complete 1,000-example Qwen2.5-7B-Instruct workload, our BS-Ghost implementation adds 27 seconds (8.0%) to 5.5 minutes of ordinary training. Removal and retraining demonstrate that BGU identifies data that causally shapes final behavior. At each training step, signed information identifies which examples strengthen or weaken the target behavior, explaining how behavior develops during training. Signed information also enables cheap intervention during training: it predicts how changing example weights will affect behavior in the next update. We then use these predictions to choose weights that steer behavior toward a desired target. This makes our framework a practical foundation for scalable oversight and verification of training pipelines and processes, helping evaluators assess model alignment, understand how it develops during training, and guide interventions that shape ongoing learning. --- ### One Model Is Not a Crowd: Multi-LLM and Aspect-Conditioned Diverse Comment Generation Authors: Nafis Irtiza Tripto, Delvin Ce Zhang, Mahjabin Nahar, Dongwon Lee Published: 2026-09-27 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.33666v1 Human communication on the internet is shaped by diverse perspectives, most visibly expressed in online comment spaces. As large language model (LLM)based AI agents begin to inhabit these spaces, a key question arises: whether synthetic comment threads can capture the diversity inherent in human discourse. This concern is increasingly important, as the growing presence of homogenized AI-generated content risks reducing diversity over time, potentially leading to model collapse and degrading the richness of digital communication. Inspired by the plurality of human crowds and the aspect-driven nature of discourse, we hypothesize that comment diversity is better approximated by combining multiple LLMs with aspect-conditioned generation. We formalize and evaluate this approach using models from different providers and introduce a framework that characterizes diversity across semantic, linguistic, and socio-pragmatic features along three axes: dispersion, coverage, and alignment. Using this framework, we conduct a large-scale study on over 2 million YouTube comments across multiple domains. Our results reveal that multi-LLM and aspect-conditioned generation better align with human comment distributions and such data remains viable under pretraining style curation and is effective for downstream tasks. Yet, human diversity remains unmatched. Overall, our findings provide a practical foundation for generating more diverse and socially grounded discourse in AI-mediated environments. --- ### CompoWorld: Compositional Environment Scaling for General Agents Authors: Xiao-Wen Yang, Weiyi Xu, Wen Da, Hang Xu, Canwei Li, Hong-Jie You, Pusen Dong, Yucheng Zeng, Zhaokai Luo, Yu-Feng Li, Yao Hu, Mu Chuan Published: 2026-09-27 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.33665v1 Automatically generated environments provide a scalable source of interaction data for training general agents. However, existing approaches mainly generate tasks within a single environment, while real-world workflows require agents to connect information and actions across multiple services. We introduce Compositional Environment Scaling (\textbf{CompoWorld}), which expands the task space by composing a finite library of reusable services. Coding agents turn tool specifications into verified services with typed states and shared interfaces, while a world model handles tools that cannot be reliably implemented. A random-walk procedure connects services through dependency graphs, enabling the generation and verification of tasks that require information to flow across services. Verified trajectories support supervised fine-tuning (SFT), while our Completion-Focused Rubric Reward guides reinforcement learning (RL) toward full task completion by emphasizing criteria with lower pass rates within each rollout group. We construct 448 services exposing 10,130 tools and use 3K SFT trajectories and 1K RL tasks to train Qwen3.6-35B-A3B. Experimental results show that CompoWorld improves on its backbone by 9.17 points on average across eight benchmarks. On AutomationBench, it surpasses frontier models such as Claude Opus 4.6 and leads all compared agent-specialized 35B-A3B models. --- ### LieDiscover: Adaptive Symbolic Library Construction for Explicit Open-form Symmetry Discovery Authors: Xinxin Li, Jianming Ma, Xingyu Cui, Da Li, Juan Zhang, Junping Yin Published: 2026-09-27 Categories: cs.SC, cs.LG arXiv: https://arxiv.org/abs/2609.33663v1 Discovering underlying symmetries from data has emerged as a crucial challenge in scientific discovery. Existing data-driven methods for symmetry discovery fail to determine the exact number and mathematical form of unknown infinitesimal generators. Recent explicit methods represent generators using a predefined function library and identify them through algebraic optimization, but they often struggle to capture complex symmetries involving high-order polynomials or transcendental functions. To address this limitation, we formulate symmetry discovery as a joint optimization problem over the function library and coefficients. We propose a novel framework that leverages an encoder-decoder architecture to dynamically generate symbolic expressions and expand the library. This generation process is optimized via reinforcement learning, which accelerates the exploration of the symbolic search space through step-wise rewards. Experiments demonstrate that LieDiscover can successfully uncover open-form infinitesimal generators involving high-order polynomials or transcendental functions, which remain intractable for existing methods. The discovered symmetries also improve performance in downstream PDE solving and discovery tasks. --- ### Audit-First VAPO: Risk-Certified Selective Updates under Imperfect Verification Authors: Miaobo Hu, Shuhao Hu, Xiaobo Guo, Xin Wang, Bokun Wang, Rui Chen, Daren Zha, Jun Xiao Published: 2026-09-27 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.33662v1 Imperfect verifiers can assign a harmful update direction even when clipping and regularization bound its magnitude. We introduce Audit-First VAPO, which separates discrete directional admission from continuous magnitude control. An observation-only accept-appeal-abstain policy uses a finite secondary-verification budget; its action trace is frozen before clean labels are joined. Simultaneous finite-sample bounds then certify selected harmful risk, coverage, and verifier-call rate over a predeclared policy family. Conditional Hoeffding-Azuma bounds account for the dependence induced by shared budgets, and rollout or verifier changes initiate a new certification stage. After admission, a bounded trust-clip-KL actuator controls magnitude. We evaluate two models on two reasoning benchmarks against static RLVR, matched-random selection, confidence thresholding, noise correction, and verifier augmentation. On Qwen3.5-0.8B and GSM8K at target risk $ρ=0.08$, RC-VAPO achieves 74.1% accuracy, selected harmful risk 0.0697, coverage 0.4125, and relative verifier cost $1.16\times$. At matched coverage and update magnitude, its selected-risk difference from matched random is -0.0260 with paired 95% interval $(-0.0364,-0.0157)$. Across asymmetric, confidence-dependent, and correlated-verifier noise, the certificate is satisfied on 57 of 60 independent runs. These comparisons isolate informative directional selection from proposal suppression, update shrinkage, and additional verifier computation. --- ### Learning Multimodal Embeddings with Evidence-Aligned Readout Authors: Zirong Chen, Fuda Ye, Enjun Du, Junfu Pu, Xinlei Wang, Xinyu Zuo, Lisheng Duan, Haijin Liang, Jin Ma, Jiachuan Wang, Yongqi Zhang Published: 2026-09-27 Categories: cs.CV, cs.IR arXiv: https://arxiv.org/abs/2609.33659v1 Multimodal large language models can expose task-relevant evidence through generation, but producing useful evidence does not by itself determine how it enters a retrieval embedding. We study whether the semantic organization of that evidence can also specify where representations are read. To address this question, we introduce EviAlign, which couples Semantic Evidence Generation with Boundary Readout in a shared multimodal large language model. It organizes evidence into five semantic units, reads the contextualized state at each unit boundary, and aggregates these states into a single normalized embedding. Generation and contrastive retrieval objectives jointly train this shared structure. With the same trailing readout, semantic evidence and free-form CoT yield nearly identical retrieval performance, suggesting that evidence organization alone does not explain the full gain. A controlled $2\times3$ study compares consistent and permuted evidence organization across three readout strategies, using training targets with matched evidence spans. With five readout states and the same mean pooling, the advantage of consistent semantic organization grows from 0.65 points at length-based training positions to 2.39 at evidence boundaries, yielding a 1.74-point co-design interaction. Across 12 MMEB retrieval tasks, EviAlign achieves 76.9 average Recall@1 with 500K training pairs while retaining single-vector indexing and scoring. --- ### AgentBoundary: Counterfactual Evaluation of Safety in Tool-Using LLM Agents Authors: Tianzhuo Yang, Zirui Mi, Yantao Huang, Guoxi Zhang, Jiawei Chen, Yaodong Yang, Jingwei Yi Published: 2026-09-27 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.33658v1 Safety alignment for large language models (LLMs) in conversational settings is largely framed around whether to answer or refuse a request. In agentic settings, however, the same models must decide whether to act as permission-critical evidence emerges during execution. This creates a distinct challenge: apparent risk, action permissibility, and task competence are easily confounded, making agentic over-refusal difficult to distinguish from ordinary task failure. To address this, we introduce AgentBound, the first four-way counterfactual generation-and-evaluation framework for tool-using agent safety. AgentBound transforms the same executable workflow by independently varying apparent risk and action permissibility, enabling controlled comparisons of risky-looking but authorized tasks and routine-looking but unauthorized tasks. These comparisons jointly diagnose over-refusal and unsafe compliance while controlling for task competence. We instantiate AgentBound as a human-validated 4,000-task evaluation suite with trajectory-based and post-state-based judgments. Across 17 model and harness configurations, high safety frequently coexists with poor authorized-task completion: GPT-5.5 blocks 99.5\% of routine-looking unauthorized actions yet completes only 28.7\% of risky-looking authorized tasks. We further train a lightweight runtime calibration module that improves authorized-task completion by 18.2\% on average across 10 evaluated configurations, while improving unsafe-action blocking by 5.4\% on average. These show that effective agentic alignment requires action decisions to track permission-relevant execution evidence, rather than refusal strength alone. --- ### Towards Eliminating Catastrophic Forgetting in the Curriculum Learning of Math Reasoning Tasks Authors: Zengyan Yang, Yangyang Wu, Kai Huang, Pengfei Lyu, Tianyi Zhang, Mengying Zhu Published: 2026-09-27 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.33655v1 Curriculum learning has found broad application across numerous domains. Nevertheless, its effectiveness is intrinsically curtailed by catastrophic forgetting, driven by the shifts in model parameter distributions between curriculum tasks. In this paper, we investigate the phenomenon of catastrophic forgetting in this training paradigm, building on the established efficacy of curriculum learning. Our theoretical analyses of parameter update dynamics demonstrate that catastrophic forgetting in curriculum learning stems from the divergence of task optima, which is generally essential to the faster convergence of curriculum learning; therefore, forgetting cannot be completely eliminated. Based on this finding, we augment the training process and propose IV-EWC, which incorporates Elastic Weight Consolidation (EWC) into the curriculum learning objective to curb catastrophic forgetting in mathematical reasoning, a prototypical curriculum learning scenario. IV-EWC employs the influence function to construct a representative validation set from the curriculum's training data, which is used to drive dynamic regularization during training. We further present an extended theoretical analysis to show that EWC-based regularization methods mitigate catastrophic forgetting in curriculum learning, thereby providing theoretical support for IV-EWC. Empirical evaluations on three backbone models and three benchmarks indicate that curriculum learning exhibits catastrophic forgetting. IV-EWC alleviates this issue, reducing forgetting by 162% on average relative to vanilla curriculum learning and yielding positive backward transfer, as evidenced by improved performance on easier tasks after subsequent training on challenging tasks. --- ### Demonstration-Free Success-Probability Reward Learning for Generalist Robot Policies Authors: Duo Wu, Haifeng Wang, Rongwei Lu, Jinghe Wang, Tianyi Xiong, Zhimin Wang, Chao Yu, Shuai Ma, Zhi Wang Published: 2026-09-27 Categories: cs.RO, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.33653v1 Reinforcement learning (RL) enables generalist robot policies to improve through trial-and-error interaction, yet its effectiveness is fundamentally constrained by sparse task rewards. Existing general-purpose reward models typically alleviate this issue by learning task progress from expert demonstrations, but introduce a distribution mismatch with the mixed-quality rollouts encountered during policy optimization, making their estimates unreliable on suboptimal and failed behaviors from which the policy must learn. In this work, we introduce a demonstration-free reward learning paradigm where dense reward feedback can be learned directly from sparse task outcomes and policy experience. We theoretically show that terminal task outcomes implicitly define dense success-probability feedback at intermediate timesteps, which can be recursively learned through bootstrapping. Based on this insight, we introduce eVTA$_0$, which learns success probabilities from mixed-quality policy rollouts through temporal-difference-style bootstrapping, without expert demonstrations or intermediate annotations. We further introduce RL with Evolving Rewards (RLER), a closed-loop framework that adapts eVTA$_0$ using newly collected rollouts as the policy evolves. Experiments show that eVTA$_0$ provides more informative rewards than state-of-the-art reward models and achieves the best average policy performance across all LIBERO task suites under the same RL training budget, improving success rates by 5.4%-13.8% over the initial policy. In real-world manipulation, RLER further improves overall success rates by 20%-26%, with 35%-36% gains under out-of-distribution conditions. These results demonstrate the effectiveness of demonstration-free reward learning and adapting rewards as the policy evolves. Project webpage: https://duowuyms.github.io/evta0. --- ### Tsubame: Tree Replay for Diffusion-Based Speculative Decoding Authors: Yepeng Weng, Qiao Hu, Takehisa Yairi Published: 2026-09-27 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.33652v1 Context-aware dynamic trees allocate the speculative decoding budget according to draft path probabilities, adapting their depth and branching to the current context. Under stochastic decoding, however, we find that this structural advantage does not always compensate for the acceptance gains of random sampling paired with advanced verification, and such dynamic trees can fall behind sampled chains in some settings. These trees grow their topology from the candidates themselves, so the tokens submitted for verification are typically the deterministic high-score tokens selected during construction. This coupling is not inherent: once the topology is fixed, its nodes can be repopulated by sampling, allowing dynamic trees to retain their structural advantage while also benefiting from random sampling and advanced verification. Diffusion-based drafters make this practical, as their parallel outputs or lightweight conditional corrections allow candidates to be regenerated cheaply after the complete topology is known. We introduce Tsubame, a two-pass tree speculative decoding framework for diffusion-based drafters. The first pass plans and freezes a context-aware topology using draft path scores; the second replays the fixed topology, sampling the tokens that populate its nodes to form the candidate tree for verification. We prove that Tsubame is lossless under compatible sampling and verification strategies. Experiments across three diffusion-based drafters, six datasets, and multiple candidate budgets show that Tsubame improves acceptance length and throughput over deterministic trees, including settings where it reverses their disadvantage against sampled chains. --- ### FuseAlign: Forced Alignment in the Wild Authors: Mithilesh Vaidya, Stephen Bailey, Sumukh Badam, Matthew Bendel, Xingzhe He Published: 2026-09-27 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.33650v1 Word-level forced alignment estimates when each transcript word occurs in an audio recording. It underpins text-based media editing, subtitling, speech-data curation, and phonetic analysis. Existing evaluations understate the difficulty of forced alignment by relying on short, clean speech, perfect transcripts, and metrics that obscure consequential alignment errors. In contrast, real-world media and data-processing pipelines operate on long and diverse recordings. Additionally, forced aligners often operate on error-prone automatic speech recognition (ASR) output. We address these gaps with improved evaluation metrics, a scoring protocol for real ASR transcripts, and AlignBench, a benchmark spanning diverse speaker, acoustic, and text conditions. We further introduce FuseAlign, a transformer-based aligner trained on large-scale pseudo-labeled speech with online label correction. FuseAlign performs joint contextualization of audio and text for the localization of coarse words. The model then refines boundaries at millisecond resolution and detects missing transcript words in the audio without lexicon-based or Viterbi decoding. On AlignBench, FuseAlign substantially outperforms all baselines and remains robust under real ASR transcripts. Ablations show that convolutional upsampling and EMA-snapshot label correction matter more than model properties such as parameter count. --- ### From Distributions to Stochastic Processes: Neural Approximation of Measure-Valued Maps Authors: Yichen Wang, Ziyi Wang, Wenlian Lu, Chenghuang Shen, Jianfeng Liu, Zhengdong Xiao, Longjiu Luo, Qianrong Wang Published: 2026-09-27 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.33649v1 Learning mappings between probability distributions arises naturally when inputs and outputs are represented by populations of samples rather than individual observations. We develop an approximation-theoretic framework for distribution-to-distribution learning and extend it to mappings between stochastic processes. For continuous operators on $W_2$-compact families of finite-dimensional probability laws, we establish uniform neural approximation in the 2-Wasserstein metric using finite law statistics, a simplex-valued neural map, and a shared atomic output support that guarantees valid probability measures. We further extend this principle to probability laws on separable Hilbert spaces through finite-rank orthogonal projections. These results establish the representational feasibility of learning transformations between probability laws rather than deterministic vectors or functions. To demonstrate practical relevance, we study two problems naturally defined at the distribution level: prediction of first-passage-time distributions for an Ornstein--Uhlenbeck process and nonlinear response-path laws of a Duffing oscillator. Because the theory is model-agnostic and broader than any single practical architecture, the experiments use task-adapted neural models rather than reproducing the theoretical construction exactly. In both problems, the proposed models outperform a fixed-feature MLP baseline and distribution-space kernel regression. These experiments complement the theory by demonstrating the practical learnability of distribution-to-distribution transformations in random systems. --- ### Probe to Act: Elevating Browser-Use Agent via Active Visual Probing Authors: Keliang Li, Heng Wang, Chen Hu, Daxin Jiang, Hong Chang, Shiguang Shan Published: 2026-09-27 Categories: cs.AI, cs.CL, cs.CV arXiv: https://arxiv.org/abs/2609.33646v1 Browser-use agents require seamless alignment between structured web metadata and visual information, while preserving relevant context across long interactions. Existing interfaces often rely on either screenshot-level action prediction or static Set-of-Marks overlays, leaving the model to resolve dense DOM-pixel alignment before every operation. We introduce Probe to Act (P2A), an active probing framework for the browser-agent loop that moves this alignment into decision time. P2A addresses an asymmetric bridge between symbolic DOM hypotheses and screenshot layout by rendering on-demand symbolic DOM structure back into pixels. Before committing a state-changing browser operation, the agent can issue lightweight probes to translate DOM handles into pixel evidence, map screen regions back to DOM candidates, register visual-only targets, and commit verified notes. These interleaved processes naturally produce evidence-based memory: only probed, acted-on, or explicitly committed observations are kept across steps, preserving only decision-critical evidence in long-horizon contexts. P2A can be used as a prompting strategy for proprietary models under the standard DOM+SoM interface, and can be distilled into open-weight models through cold-start synthesis and self-bootstrapped SFT. Across three browser-use benchmarks, P2A shows clear gains on task success rate for both proprietary and fine-tuned models; on VisualWebArena, for example, it improves Gemini-3-Pro from 54.1% to 61.2% and Qwen3-VL-8B from 24.6% to 32.9%, while matching the costly full-observation history ($\sim$3$\times$) at only $\sim$1.2$\times$ the peak retained input context of action-only history. --- ### Learning to Learn from Context: Synthetic Training from Perturbed Public Documents Authors: Haoyi Wu, Yang Xiao, Yusong Sun, Wenyang Hui, Zhaokai Luo, Chengyue Jiang, Mu Chuan Published: 2026-09-27 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.33642v1 Real-world tasks often require large language models (LLMs) to learn from complex task-specific context rather than pretrained parametric knowledge. This capability remains a weakness of LLMs, while human annotation for such task contexts is expensive and difficult to scale. Public high-quality documents are an abundant alternative, but much of the public web has already been consumed during pretraining: training on such documents naively would reward memorization rather than context learning. In this work, we attempt to make use of high-quality public documents with small perturbations and empirically find that LLMs can successfully generate context-dependent reasoning traces and answers, which are then used to train a student model. Specifically, we construct a synthesis pipeline that (i) rewrites source documents to reduce memorization risk, (ii) generates questions and rubrics that require reasoning over the document, (iii) answers the questions with the document as context, and (iv) admits only samples that genuinely depend on the document. Without any human annotators, our pipeline generates about 10k samples from 3.5k documents, and the resulting student model substantially improves the performance on CL-bench. SFT raises a Qwen3.6-35B-A3B student from 13.7% to 22.8%, and a subsequent rubric-reward RL stage reaches 24.6%, on CL-bench comparable with a frontier model of over a trillion parameters, Qwen3.8-2.4T (23.9%). We also observe a broad transfer of improvements to long-context understanding, instruction following, and reasoning, while code generation and knowledge remain mostly flat. We hope this work provides a reproducible and scalable way to improve the ability of LLMs to learn from context, and to facilitate further research on context-grounded reasoning. --- ### Scalable and Data-Driven Decision Support in the Maintenance, Repair, and Overhaul Process Authors: Houkun Zhu, Helena Ebel, Dominik Scheinert, Florian Schmidt, Jens Altenkirch, Odej Kao Published: 2026-09-27 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.33641v1 Several businesses apply maintenance, repair, and overhaul (MRO) principles to the life-cycle of their existing products. In cases like casted gas turbine component Product Lifecycle Management (PLM), repairing components in frequent intervals can extend the lifetime expectation of the product, provide higher cost efficiency compared to newly produced components, and even improve the part design during the repair cycle. Another aspect of repair concerns sustainability, as products often contain rare materials. The emissions produced by the repair process are usually smaller than mining materials and casting new components. To optimize the repair process further, we propose the Smart Expert System (SES), which assists engineering experts with machine learning-based decision support throughout the repair process. We elaborate on its IT architecture and present machine learning models employed for representative MRO use cases. The SES is evaluated using actual industry data from a leading gas turbine company and demonstrably fulfills formulated requirements concerning the suitability of the overall decision support and the stability of the enclosing IT architecture. --- ### SafeMol: Dual-Modality Safety Alignment for Molecular Multimodal Models Authors: Xinmiao Wang, Ruijie Wang, Menghui Wang, Jiawei Chen, Haoyue Deng, Ran Zhang, Xingxuan Zhang, Xiao Wang Published: 2026-09-27 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.33640v1 Molecular multimodal models support diverse understanding and generation tasks but may introduce safety vulnerabilities when handling hazardous molecules. In this work, We reveal substantial jailbreak vulnerabilities under both text-only and graph-conditioned settings. Our analysis further shows that safety robustness must hold across input modalities while balancing safety, over-refusal, and utility. To address these challenges, we construct SafeMolBench, a molecular multimodal safety-alignment benchmark with 3702 samples covering 618 unique hazardous molecules and safe molecular tasks, organized into hazardous-harmful, hazardous-allowed, and utility-replay subsets to support unified training and evaluation of safety, over-refusal, and utility. Based on SafeMolBench, we propose SafeMol, a parameter-efficient safety alignment framework that jointly optimizes lightweight modules across text-only and graph-conditioned inputs, uses MMD for distribution-level representation alignment to reduce modality-induced discrepancies, and explicitly models molecular hazardousness and harmful operational intent. Experiments on SafeMolBench show that SafeMol reduces attack success by several tens of percentage points while largely maintaining low over-refusal and preserving molecular-task utility. --- ### Trajectory Unlearning on LLM-based Agents Authors: Yingdan Shi, Ren Wang Published: 2026-09-27 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.33639v1 Existing large language model (LLM) unlearning has focused primarily on removing specific knowledge, such as harmful facts, private data, or copyrighted content. However, as LLMs are increasingly deployed as autonomous agents, a fundamental yet overlooked problem emerges: beyond suppressing what an agent knows, an agent should not reproduce undesired behaviors through its action trajectories. In this work, we introduce trajectory-level unlearning, a new problem formulation that targets the removal of specific action trajectories in long-horizon agentic tasks, rather than factual knowledge. We identify two fundamental challenges that distinguish trajectory unlearning from knowledge unlearning: (1) our unlearning target is what the agent \emph{does}, not what it \emph{says}; and (2) trajectories are sequentially dependent action sequences that cannot be decomposed into isolated prompt-response pairs without losing inter-step structure. To address these challenges, we propose Group-injected Relative Policy Optimization (GiRPO), which injects forget trajectories into the policy rollout group with penalized rewards and isolates the normalization statistics, yielding a stable and bounded unlearning signal that does not corrupt gradient updates for normal task trajectories. We construct trajectory unlearning benchmarks from two application scenarios, household tasks (ALFWorld) and online shopping (WebShop), and design three complementary metrics for evaluating forgetting quality and model utility. Experiments on ALFWorld and WebShop demonstrate that GiRPO effectively unlearns target trajectories while preserving task success rates, outperforming existing knowledge-unlearning baselines on both forgetting quality and task utility. --- ### Quantifying Behavioral Tails in Black-Box Language Models Authors: Elsayed Eshra, Ali Al-Lawati, Dongwon Lee, Suhang Wang Published: 2026-09-27 Categories: cs.LG, cs.CL, stat.ML arXiv: https://arxiv.org/abs/2609.33638v2 We introduce RareTrap, a framework for estimating the probability of severe behaviors in black box large language models (LLMs). A key challenge for probability estimation is defining a tractable distribution over the input space. To accomplish that, RareTrap uses a surrogate LLM and constructs a geometry-aware mapping from a lower-dimensional latent reference space into its token-embedding space to induce an explicit and reproducible distribution over input prompts. A response-level performance function is utilized on the response to quantify behavior severity. This enables sequential rare event simulation that concentrates evaluations on progressively more severe behaviors while preserving probability under the induced prompt distribution, which would otherwise be prohibitive to measure. Across 10 open-weight and two frontier models (GPT-5.4 and Claude Sonnet 4.6), we find that RareTrap successfully induces severe resource consumption behaviors and computes their probability with as few as 200 evaluations. RareTrap provides model developers a principled approach for evaluating language models under a common distribution, and prioritizing alignment effort to improve safety and mitigate risks. --- ### Safety Reconstructed: Generative Modeling via Masked Diffusion Builds Strong Safety Guardrails Authors: Gert Lek, Abele Malan, Chaoyi Zhu, Pin-Yu Chen, Robert Birke, Lydia Chen Published: 2026-09-27 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.33634v1 Guard models are the last line of defense between a language model and a harmful output, yet their training objective is surprisingly narrow. Existing guards learn to predict a single verdict token from a conversational context, concentrating supervision on a single target. The consequences are structural: models latch onto shortcut features, are overconfident, and remain sensitive to where safety evidence appears in the sequence rather than its role in the full context. We propose a different framing. Rather than predicting a label from text, our LLaDA-Guard asks which label better explains the text: scoring the prompt or response under each label hypothesis and classifying based on their difference. This shifts supervision to every token in the moderated region, forcing the model to account for full content rather than its most discriminative fragments. We instantiate this idea with a masked diffusion language model, fine-tuning LLaDA-8B-Instruct with a class-conditional reconstruction objective using LoRA and requiring no architectural changes beyond the base model. LLaDA-Guard leads on average rank against discriminative baselines trained on stronger backbones across seven held-out safety benchmarks, while exhibiting substantially better confidence calibration (ECE 0.0875 vs. 0.1384 for Qwen3Guard), less over-defense on benign prompts with unsafe-looking cues, and less prompt leakage when moderating responses. Its generative nature further enables token-level risk localization as a natural byproduct, yielding a pipeline for rewriting unsafe prompts into safe equivalents without additional training and achieving a 60.7% average conversion-to-safe rate. --- ### lapanda: A Matrix-Free Differentiable Solver for Nonconvex Constrained Optimization Layers Authors: Yuankun Chen, Zifei Nie, Kangyu Lin, Ján Drgoňa, Liang Wu Published: 2026-09-27 Categories: math.OC, cs.LG, eess.SY arXiv: https://arxiv.org/abs/2609.33630v1 Differentiable optimization brings the structural guarantees of mathematical optimization to network pipelines, allowing them to be trained end-to-end. However, its application remains challenging for nonconvex constrained problems, as existing differentiable solvers often suffer from limited modeling expressiveness due to their reliance on specialized problem structures, while also incurring substantial computation time and memory overhead in both the forward and backward passes. To address these challenges, we propose lapanda, a matrix-free differentiable solver for nonconvex optimization with general constraints. It reformulates the problem to a sequence of augmented Lagrangian subproblems, each handled by a first-order inner solver through a proximal averaged quasi-Newton algorithm with adaptive linesearch, thus enabling efficient forward optimization. We establish local well-posedness of the solution map and convergence of the outer iterations, and further derive a sensitivity alignment between the original problem and the final subproblem in the backward pass, demonstrating that the subproblem sensitivity, which can be computed efficiently in a matrix-free manner, provides a principled approximation to the exact optimizer sensitivity. We evaluate lapanda on nonconvex constrained Rosenbrock benchmarks, imitation learning with several representative constrained optimal control problems, and embedded robotic obstacle-avoidance tasks. Compared with state-of-the-art differentiable solvers, lapanda delivers substantial reductions in computation time and memory footprint while maintaining reliable constraint satisfaction and learning performance. --- ### Climbing the Hill: Prompt Injection Red-Teaming Against Frontier Models with Curriculum Reinforcement Learning Authors: Chenlong Yin, Xiaolong Jin, Wei Zou, Yanting Wang, Jinyuan Jia Published: 2026-09-27 Categories: cs.LG, cs.CR arXiv: https://arxiv.org/abs/2609.33628v2 Prompt injection is a leading security risk for LLMs and LLM-based applications such as agents. State-of-the-art red-teaming methods for prompt injection leverage reinforcement learning (RL) to train an attacker LLM to generate effective injected prompts. However, when targeting frontier LLMs such as GPT-6-Luna, a major challenge is the cold-start problem: every attack attempt by the attacker LLM fails and thus receives zero reward, providing no signal for learning. In this work, we propose a curriculum learning-based method to address the cold-start problem. In particular, we propose to train the attacker LLM against a sequence of increasingly robust target LLMs, with each stage warm-starting from the attacker LLM obtained in the previous one. However, simply training against a weak target (e.g., GPT-4o-mini) may not sufficiently prepare the attacker LLM to obtain useful learning signals against a frontier LLM (e.g., GPT-5.6-Terra). Instead, we find that the design of the curriculum is critical: after each stage, the attacker LLM needs to partially succeed against the next target LLM such that it can learn from successful attempts to attack the new target. Our extensive evaluation shows that our method can effectively red-team frontier LLMs, achieving an attack success rate (ASR@10) of 93.8\% and 45.0\% against GPT-5.6-Luna and GPT-5.6-Terra on AgentDyn, whereas state-of-the-art RL methods such as RL-Hammer and PISmith achieve 0\% ASR under the same setting. Moreover, we find that the attacker LLM transfers across targets, e.g., an attacker LLM trained to defeat one strong LLM (GPT-5.6-Terra) also succeeds against six other frontier LLMs (e.g., GPT-6-Luna) it was never trained on. Our code is available at https://github.com/albert-y1n/PIForge. --- ### StoryEngine: A State-Grounded Agentic Framework for Video Storytelling Authors: Yingrui Wang, Zeqing Wang, Yeying Jin Published: 2026-09-27 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.33627v1 Despite recent progress in agentic multi-shot video generation, producing coherent and consistent long-form stories remains challenging. Existing agentic pipelines typically rely on textual shot plans or previously generated pixels, yet lack an explicit mechanism for propagating the consequences of story events and maintaining the video world state across shots. As a result, missing visual details may be reconstructed inaccurately, while visual drift may propagate across subsequent shots, undermining both narrative coherence and visual consistency. To address these challenges, we propose StoryEngine, a state-grounded agentic framework for video storytelling. StoryEngine establishes a separation between authoritative semantic plans and unreliable visual observations. Specifically, StoryEngine maintains a structured representation of entity placement and story-relevant states, and propagates event-induced changes to define the intended start and end states of each shot. To visually realize these states, StoryEngine constructs canonical references for recurring entities and environments, and compiles state and visual constraints into executable render plans. Meanwhile, to realize these states correctly, a bounded evaluation-guided repair loop further corrects local state inconsistencies. Together, these mechanisms preserve causal story progression and prevent local visual errors from propagating across shots. To comprehensively evaluate long-form storytelling, we construct a benchmark across diverse scenarios and visual styles, with metrics assessing storytelling quality, narrative coherence, and visual consistency. Experimental results demonstrate that StoryEngine consistently outperforms state-of-the-art methods across all evaluation dimensions, validating its effectiveness for coherent and consistent video storytelling. --- ### Learning Dynamics of Continual Learning: A Unified View of Data Attribution, Forgetting, and Plasticity Loss Authors: Yi Ren, Wenlong Deng, Guanzhe Hong, Clare Lyle, Yarin Gal Published: 2026-09-27 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.33620v1 Modern language models are likely to be updated throughout their lifetime rather than trained once and frozen. Each update therefore participates in a recurring cycle: decide which experience to learn from, understand what that update changes, and remain capable of learning from what comes next. We show that these challenges are governed by the same evolving update--behavior interaction. We derive a token- and layer-wise decomposition of how learning from one token changes another prediction. By separating the softmax force, shared readout geometry, and residual connections, it exposes two interaction channels and yields a forward-computable approximation. Following this interaction through time reveals a unified picture of continual adaptation. Positive interaction identifies useful experience; negative interaction produces either concentrated collision or accumulated erosion; over longer horizons, updates reshape the shared geometry mediating future learning signals, reducing their transmission. These predictions lead to effective data selection, mechanism-specific controls for interference, and a readout-based diagnostic of future learnability whose degradation predicts the benefit of restoring the readout. Across models and training regimes, the same local interaction thus explains both what an update changes now and how learning today changes what can be learned tomorrow. This view connects data attribution, forgetting, and plasticity loss as distinct regimes of the same evolving learning dynamics. --- ### ParaAgent: Reinforcing Parallel Acting in Open-World Tool Environments Authors: Shengbin Yue, Hongru Wang, Siyuan Wang, Xiaoxin Chen, Wei Chen, Zhongyu Wei Published: 2026-09-27 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.33618v1 Language model agents are increasingly deployed in open-world tool environments, which require balancing exploring unknown capabilities and exploiting known ones. Existing methods face a performance-efficiency tradeoff: they either rigidly decouple exploration and execution or interleave them without coordination. We argue that the key lies not in whether to decouple or interleave them, but in how to coordinate them across granularities. We introduce ParaAct, a structured parallel-action loop that combines phase-level Exploration $\rightleftharpoons$ Execution with action-level parallelism. To learn this loop, ParaAgent combines multi-agent cold-start demonstrations with reinforcement learning under multi-level advantage decoupling, making planning structure explicit and supervising it with step-, phase-, and trajectory-level rewards. Learning is supported by our ToolEnv, a scalable simulator grounded in 50,011 realistic tool interfaces. On two open-world tool benchmarks, ParaAgent-4B achieves the best average success among all baselines, including GPT-4.1 systems, with the largest gains on multi-tool tasks. Behavioral analyses show that these gains stem from this action organization, highlighting its importance for capable and efficient open-world agents. --- ### SpatialSpeak: QA-Native Reconstruction with Local and Global Context for Spatial Chain-of-Thought Reasoning Authors: Yang Cao, Jiaxin Zhang, Dave Zhenyu Chen, Yingji Zhong, Ruiyuan Gao, Lanqing Hong, Dan Xu Published: 2026-09-27 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.33616v1 Vision-language models (VLMs) can benefit from geometric priors for multi-view spatial reasoning, yet answer-only training does not directly supervise the intermediate geometric estimates and their use in deriving quantitative spatial answers. We hypothesize that spatial chain-of-thought (CoT) supervision becomes more effective when the VLM first jointly learns complementary local geometry and global scene context through multi-view reconstruction. We introduce SpatialSpeak, a two-stage framework that connects QA-native reconstruction pretraining with spatial CoT learning. In Stage I, QA-Native Reconstruction Pretraining (QA-RP) combines marked-point 3D queries for fine-grained local geometry with object-center queries for global scene context across views. Both tasks are formulated as text-based question answering, allowing geometric estimation and subsequent reasoning to share the same autoregressive output interface. In Stage II, spatial CoT with Visual Compensation (CoT-VC) trains the model to express question-relevant geometric estimates and use them to derive answers, with reliability assessment and visual compensation supporting answer refinement when needed. On ReVSI, QA-RP increases the gain from CoT-VC from 2.6 to 6.9 points, and ablations show that both local and global reconstruction supervision are beneficial. SpatialSpeak achieves state-of-the-art results on ReVSI, VSI-Bench, and SPAR-Bench, with a ReVSI score of 62.8 that exceeds the strongest compared baseline by 8.7 points. --- ### EAT: Expert Account Tracker for Efficient MoE Inference Authors: Yuexian Li, Yifei Yang, Zouying Cao, Hai Zhao Published: 2026-09-27 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.33614v1 Mixture-of-Experts (MoE) models have emerged as a revolutionary method to scale Transformer models. However, traditional MoE architecture still suffers from inefficiency since a large number of experts are unnecessarily activated. Existing approaches for reducing the number of activated experts often overlook the historical performance of each expert. In this paper, we propose EAT, a novel method called Expert Account Tracker (EAT), which utilizes history-awareness metrics and adaptive thresholding to dynamically select the most important experts, thereby reducing the activated expert number while effectively maintaining the model performance. Experiments show that EAT outperforms the existing baseline Top-P method across multiple models and datasets, achieving over 25% an average reduction compared to the vanilla method in the number of activated experts and performing better token generation speed compared to the baseline. Furthermore, the performance of pruned models can be efficiently recovered via OPD using only 9K data. Additionally, through ablation studies, we find that excessively reducing the number of activated experts can significantly harm model performance, and the importance of experts varies across layers, with higher-level experts being generally more critical. --- ### Quizzing the Translation: A Prover-Grounded Evaluation Metric for NL$\rightarrow$FOL Authors: Pu Suo, Ali Emami Published: 2026-09-27 Categories: cs.CL, cs.AI, cs.LO arXiv: https://arxiv.org/abs/2609.33612v1 A standard pipeline for symbolic reasoning over natural-language problems translates them into first-order logic and invokes a theorem prover. The translation step is the bottleneck: swap "every" for "some" and every inference that follows is corrupted. Yet today's metrics often score more broken translations higher than less broken ones, because BLEU, BERTScore, and Smatch++ reward surface overlap that the worst errors happen to preserve. We introduce SIV, which derives two kinds of probes from the target formula and uses a theorem prover to verify the candidate translation against each. Positive probes are statements the candidate must entail, which detect translations that drop content; contrastive probes are statements the candidate must not entail, which detect translations that assert more than the original. On a controlled pool of perturbed FOLIO translations, the severity of the error accounts for 80% of SIV's score variance, compared with at most 17% for any prior metric. Across six error classes on a disjoint pool, SIV scores the reference above the perturbed candidate in over 99% of pairs. Because each probe is labeled with what it tests, the failure pattern also supplies a labeled error trace, recovering the perturbation class at macro-F1 0.638, nearly double the score-only baseline. On 434 expert-audited real LLM translations, SIV attains the top AUC, uniquely detects and grades expert-labeled major errors, and abstains, rather than mis-scoring, on out-of-vocabulary translations. --- ### Anguinus Sculpturae: Compositional Synthesis of Peak-Enhancement Breast DCE-MRI Scans Authors: Benjamin Hamm, Nico Albert Disch, Maximilian Rokuss, Yannick Kirchhoff, Constantin Ulrich, Klaus Maier-Hein Published: 2026-09-27 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.33611v1 Dynamic contrast-enhanced breast MRI (DCE-MRI) is rich in anatomical and perfusion information, but its reliance on gadolinium-based contrast agents raises safety concerns and adds cost. Virtual contrast enhancement, synthesizing post-contrast from pre-contrast images, is a promising alternative. We address the MAMA-SYNTH challenge task of predicting peak-enhancement breast MRI. Rather than adopting the full machinery of diffusion or flow matching, we observe that under a rectified, straight-line path the generative process collapses to a single difference prediction: the synthetic peak image is the pre-contrast image plus a predicted enhancement map, recovered in one forward pass. Around this we build Anguinus Sculpturae, a compositional pipeline in which nnU-Net segmentations of lesion, foreground and breast region guide two generators - one optimized for global fidelity, one for lesion structure through an asymmetric Tversky term routed via a frozen segmenter - composited region-wise with Gaussian-weighted blending. On the held-out Duke subset of MAMA-MIA our model achieves the best FRD and Dice among all evaluated variants, showing that single-step difference prediction with segmentation guidance suffices to recover both global fidelity and lesion structure. Code is available at https://github.com/MIC-DKFZ/AnguinusSculpturae. --- ### Supervision Recovery for Time Series Anomaly Detection via Context-Anchored Pairing Authors: Yifei Gao, Tian Lan, Yimeng Lu, Xuming An, Meng Wang, Wenjun He, Yijie Li, Chen Zhang Published: 2026-09-27 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.33610v1 Time series anomaly detection (TSAD) remains challenging not only because anomaly labels are scarce, but also because temporal anomalies are highly context-dependent. Existing methods often rely on unsupervised objectives or surrogate abnormal patterns, providing limited supervision for context-dependent normal--anomalous distinctions. We propose Context-Anchored Pair Supervision (CAPS), a supervision-recovery framework for TSAD. CAPS views ideal anomaly supervision as a matched comparison between normal and anomalous outcomes under the same temporal context, and seeks to recover such supervision without target-domain anomaly labels. Using simulated normal--anomalous pairs, CAPS learns structure and anomaly-semantic representations through reconstruction, background consistency, and within-pair counterfactual recombination. The resulting anomaly representations form a continuous semantic space with coarse modes and induce a sampleable multimodal prior. CAPS conditionally realizes sampled semantics as residual-form effects on target reference trajectories. The resulting context-anchored normal--anomalous counterparts provide temporal supervision for discriminative detector learning. Experiments on nine datasets show that CAPS achieves the strongest aggregate performance across all four evaluation metrics among the compared methods, while complementary ablations and transfer analyses support the roles of context anchoring, semantic disentanglement, and conditional realization. --- ### You Only Edit Once: Incentivizing In-Context Capability of LLMs via Local Demonstration Refinement Authors: Jiarong Wen, Qi Wang, Yun Qu, Yixiu Mao, Heming Zou, Haoang Chi, Lizhou Cai, Yiqin Lv, Kaiyu Zhang, Yuhang Jiang, Xiangyang Ji Published: 2026-09-27 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.33609v1 In-context learning (ICL) is crucial for boosting the inference performance of large language models (LLMs). However, the effectiveness of ICL in LLMs is greatly influenced by the choice of demonstration sets. Exhaustive searches over these sets are combinatorial, and existing selectors often rely on relevance or likelihood proxies to implicitly assess ICL quality. Making repeated queries to the target LLM with these strategies can incur substantial costs. This work simplifies selection by framing it as a constrained local search problem and presents local demonstration editing (LDE). Starting with an initially retrieved set of demonstrations, LDE employs a single structured edit to explore its surrounding neighborhood while balancing performance gains with search costs. Technically, LDE is reduced to a policy search problem, for which we train a small LLM, referred to as Jev-LDE. This model as the System-1 modifies the retrieved demonstration set by performing actions such as \texttt{Keep}, \texttt{Delete}, or \texttt{Replace} elements, all within a framework of reinforcement learning with verifiable rewards. At test time, Jev-LDE executes a single edit of the retrieved demonstration set, followed by one inference from the target LLM, avoiding the need for iterative context scoring or subset searches. Across standard classification benchmarks, various target LLMs with Jev-LDE as the plug-and-play module consistently improve ICL performance, and Jev-LDE shows transferability to held-out benchmarks and models without retraining. These findings indicate that the LDE approach offers an efficient and adaptable method for harnessing the ICL capabilities of target LLMs. --- ### Learning Transferable Reaction Mechanisms from Visual Chemical Knowledge Authors: Yujian Yuan, Jiaxin Xu, Xin Cai, Yufan Chen, Zhichao Tan, Ziqi Zhou, Hanyu Gao Published: 2026-09-27 Categories: cs.CE, cs.AI arXiv: https://arxiv.org/abs/2609.33608v1 Reaction mechanisms describe the step-by-step transformations underlying chemical reactions and are central to reaction analysis and synthesis. Learning-based models have achieved strong performance on established mechanism-prediction benchmarks, but transferring them to unseen chemistry remains challenging. Such transfer is difficult because familiar mechanisms must be applied to unfamiliar molecular structures, and some target mechanisms may be poorly covered by the training data. To address these challenges, we introduce MechaVLM, a visual framework that combines transferable chemical representations with external mechanistic knowledge. It learns reusable visual features through multiscale chemical grounding and cross-rendering contrastive learning. For open-book prediction, MechaVLM retrieves a fixed set of precedents from 70,384 literature mechanism figures and re-reads relevant visual evidence as the molecular state evolves, directly using the figures without symbolic mechanism parsing. An atom-indexed language decoder then recursively generates executable electron edits to construct the complete mechanism. We further introduce MechBench, a challenging literature-derived benchmark with 2,184 mechanisms and 9,146 elementary steps. Across cross-dataset and literature-derived benchmarks, MechaVLM establishes strong zero-shot mechanism prediction. Its closed-book model alone improves Step/Pathway Top-1 by 12.50/13.93 percentage points on FlowER-to-ReactMech transfer, while external visual precedents unlock further gains on challenging OOD reactions. The learned representation also generalizes beyond mechanism prediction to atom mapping and reaction center prediction. --- ### Hierarchical Response Preservation for Continual Adaptation of Zero-Shot Graph-Text Models Authors: Haopeng Zhang, Yuhan Wang, Yubing Su, Yingxin Chen, Xiao Wang, Ruijie Wang, Jianxin Li Published: 2026-09-27 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.33607v1 Pretrained graph-text models align graph representations with textual semantics, enabling recognition of unseen classes and transfer across graph domains. However, as graph data and classes continually arrive, models should learn from new supervision while retaining their zero-shot transfer capabilities and historical task knowledge. Two challenges arise: (i) new classes can overturn historical predictions despite preserved distinctions among historical classes, and (ii) overly strict response preservation can stall learning of new classes. To address these challenges, we propose Hierarchical Response Preservation (HiRP). HiRP represents this competition through a hierarchical response that keeps each historical-class probability and sums new-class probabilities, preserving historical distinctions and aggregate competition while allowing distinctions within the new class group to adapt. It further uses the geometry induced by this response to guide constrained updates, retaining useful adaptation directions while controlling response drift. Across three class-incremental settings, HiRP achieves absolute gains of 1.84-7.95 percentage points in average accuracy over the strongest compared baseline in each setting, while mitigating zero-shot transfer degradation. --- ### Short-Length Code Designs for Integrated Sensing and Communications: A Deep Learning Approach Authors: Muah Kim, Shuangyang Li, Tayyebeh Jahani-Nezhad, Rafael F. Schaefer, Giuseppe Caire Published: 2026-09-27 Categories: cs.LG, cs.IT arXiv: https://arxiv.org/abs/2609.33605v1 Integrated sensing and communication (ISAC) enables joint communication and sensing using a shared waveform, but its signal design is challenging due to the inherent trade-off between the two objectives, particularly in the short blocklength regime. This paper proposes an autoencoder (AE)-based framework for ISAC waveform design in noncoherent settings. We derive a modified Cramér-Rao bound for multi-target delay estimation and analyze the maximum-likelihood decoding rule for noncoherent communication under correlated fading. These results reveal structural connections and trade-offs between communication and sensing objectives in waveform design. Based on this analysis, the AE learns waveform representations that jointly optimize both functionalities, with a tunable parameter controlling the trade-off. Simulation results show that the proposed design outperforms conventional schemes in both communication reliability and sensing accuracy, especially under short blocklength and fading conditions. --- ### ViCoR: Reliable Molecular Structure Extraction via Spatially Aligned Verification and Executable Revision Authors: Yujian Yuan, Xin Cai, Yufan Chen, Jiaxin Xu, Mengdi Liu, Zhichao Tan, Long Chen, Hanyu Gao Published: 2026-09-27 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.33603v1 Reliable optical chemical structure recognition (OCSR) is essential for building high-quality chemical data from scientific literature, yet even small recognition errors can propagate into chemical databases and downstream models. In practice, recognized structures often require manual inspection and correction before use, making large-scale data curation costly and difficult to scale. We therefore study Selective Structure Recognition (SSR), a post-recognition setting that automatically produces reliable structured outputs while rejecting unresolved cases. Selection-only approaches can improve reliability by rejection, but cannot create additional correct outputs beyond those produced by the base recognizer. We propose ViCoR, a repair-before-rejection framework for iterative VerIfiCatiOn and Revision. Its key idea is to make observation-prediction correspondence explicit: coordinate-preserving rendering establishes spatial correspondence between the source image and predicted structure, while index anchoring maps localized visual discrepancies to executable graph edits without full-structure regeneration. A shared VLM is progressively trained from verification to revision. On two real-world OCSR benchmarks, ViCoR improves overall accuracy from 73.53\% to 88.26\% and from 61.83\% to 84.32\%, while achieving over 97\% accepted accuracy at 85--89\% coverage. The resulting molecular data further improve reaction-extraction F1 by 15.5 points and literature-sourced reaction prediction accuracy by 7.7 and 5.8 points, demonstrating the value of automated reliability control for scientific data curation and downstream chemical learning. --- ### JustQuant: You Don't Need Smoothing, SVD, or Rotation for 4-Bit Activation Quantization Authors: Kaicheng Yang, Kaisen Yang, Chunyu Liu, Xianglong Yan, Haotong Qin, Junyi Wu, Tianao Zhang, Xun Zhang, Shaoqiu Zhang, Youbang Sun, Yulun Zhang Published: 2026-09-27 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.33601v1 Recent generative models have become increasingly powerful, but their inference cost continues to grow. Model quantization offers a promising way to compress these models and accelerate inference. However, at 4 bits, activation quantization is substantially more challenging than weight quantization. Recent post-training quantization (PTQ) and quantization-aware training (QAT) methods have made progress in 4-bit activation quantization by introducing smoothing, SVD branches, rotations, mixed precision, or advanced formats such as NVFP4. These additional operators and data types impose demanding requirements on inference engines and hardware, limiting the broad adoption of low-precision models. Can quantization be achieved using only plain low-bit operators? To answer this question, we propose JustQuant, a simple yet effective framework that moves the complexity of low-bit quantization from deployment-time operators into the training process. We first revisit model quantization from the perspective of knowledge distillation and show that a key reason existing PTQ and QAT methods fail is that they typically exploit supervision at only a single level. We then introduce Theseus QAD, a quantization-aware distillation method that progressively applies multi-level supervision, analogous to the gradual replacement process in the Ship of Theseus. Extensive experiments on DiT and diffusion large language models show two distinct regimes. For smaller models, Theseus QAD can serve as a lightweight warm-up stage that substantially improves subsequent QAT with plain operators, while naive QAD may collapse in the same setting. For larger models, Theseus QAD provides a stronger distillation training path than ordinary QAD. Across both regimes, JustQuant improves low-bit quantization quality while avoiding the complex operators required by many existing PTQ methods. --- ### The Price of Peeking: Anytime-Valid Leakage Detection on ML-KEM EM Traces Authors: Georgios Feretzakis, Alexandros Papaspyridis Published: 2026-09-27 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.33597v1 Side-channel evaluators routinely inspect leakage tests while acquisition is still running, and extend or stop the campaign based on what they see. Fixed-horizon screening such as the Welch $t$-test with threshold $|t|>4.5$ gives no error guarantee for this monitored decision rule. We study anytime-valid leakage detection based on testing by betting: SKIT-type swap-pair e-processes whose false-alarm probability is controlled uniformly over time under an explicit conditional symmetry null. In matched comparisons that share the frozen witness, rows and payoff, first-crossing detection needed 1.68-2.00$\times$ the traces of a fixed-horizon randomization test at 80% detection on synthetic streams, and 1.68-2.38$\times$ on degraded recordings from an open ML-KEM electromagnetic dataset with the primary Ridge witness at $α=0.05$. With the same primary witness and level, on undegraded reference and pqm4 recordings the monitored procedure stopped early: its median stopping point was 62-72 and 146-316 evaluation traces, i.e. 2-8% of a conservative 4096-trace budget. Under exact designed nulls on the recorded backgrounds, repeated-look $|t|>4.5$ screening over all 13000-20000 samples raised a false alarm in 2.7-12.9% of replicates, against 0.0-4.7% for terminal-only screening and no rejection by a sample-wise e-Bonferroni process, which in a prespecified follow-up detected natural-label associations in 4 of 4 backgrounds after 840-3288 traces. All recordings come from one device, and natural-label results are descriptive; we state the assumptions each claim requires. --- ### Beyond One-Step Accuracy: State-Affine Latent Transition for Reliable Visual Planning Authors: Boyuan Zhang, Yingjun Du, Xiantong Zhen, Ling Shao Published: 2026-09-27 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2609.33595v1 Joint-embedding world models enable visual planning by learning action-conditioned dynamics in latent space. Yet they are commonly trained for one-step prediction on encoded states, while planning recursively applies the learned transition to its own predictions. One-step accuracy therefore does not capture how prediction errors propagate under recursive rollout. We decompose multi-step rollout error into the errors introduced at individual steps and their propagation through subsequent transitions. We show that state-affine dynamics are precisely the differentiable transitions with state-independent Jacobians, eliminating the nonlinear propagation residual and making the error propagation operators depend only on the action sequence. Guided by this result, we introduce SALT (State-Affine Latent Transition), an action-conditioned state-affine dynamics model in which the action modulates both the state transformation and the additive update. We train SALT through recursive multi-step rollout supervision, feeding each predicted latent state back into the transition so that training matches how the model is used during planning. Across four visual planning environments, SALT exhibits $1.48$--$2.19\times$ higher one-step prediction error than the matched LeWM baseline, yet improves closed-loop success in every environment by $10.0$ percentage points on average. On OGBench-Cube, the fraction of episodes that fail with a sharp rise in model-predicted cost after execution decreases from $23.3%$ to $2.0%$. --- ### LoopLUT: 3D Lookup Tables with Progressive Region Refinement for Real-Time 4K Image Enhancement Authors: Yang Ye, Jiajun Ma, Chen Wu, Wei Wang, Dianjie Lu, Guijuan Zhang, Linwei Fan, Zhuoran Zheng Published: 2026-09-27 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.33593v1 Color enhancement of 4K images must meet a quality target under a tight compute budget. Three-dimensional lookup tables (3D LUTs) dominate real-time enhancement because they decide at low resolution and apply a per-pixel lookup at full resolution. A single global LUT, however, is spatially invariant, so an underexposed shadow and a well-exposed region that share a pixel value receive identical corrections. Spatially heterogeneous demands cannot be expressed by such a mapping. We propose LoopLUT, a region-cascaded 3D LUT with progressive refinement. A global LUT performs the overall correction, followed by K-1 loop iterations. In each iteration a gating head predicts at low resolution the region that still needs correction, then builds a residual LUT from the color statistics of that region alone. The cascaded gates form a partition of unity, so the output is a per-pixel convex combination of the K lookup results. Fusion is therefore performed by the gates themselves, with no separate fusion module and no interpolation error accumulating across rounds. The decision stage runs at a fixed 256x256 resolution, independent of output resolution, so a 4K image costs only K pure lookups. Extensive experiments across four benchmarks show that LoopLUT improves PSNR by up to 2.81 dB over the strongest prior method, while keeping real-time throughput at 4K. The same decomposition also generalizes well to underwater enhancement datasets. ---