# Research Papers — 2026-09 · bundle page 4 of 266 > Full-text markdown of papers #151–200 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/4 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/3.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/5.md) ### PassGPT+: Leveraging Linguistic Priors for Password Modeling Authors: Rajneesh Anand, Neeraj Lakshmanan, Masoud Yari Published: 2026-09-30 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.39880v1 Passwords remain the dominant online authentication mechanism, and understanding how humans choose them is essential for defensive strength estimation and attack simulation alike. Recent learning-based approaches such as PassGAN and PassGPT have shown that deep generative models can learn password structure directly from leaked corpora. However, both train from random initialization on password data alone. The role of linguistic prior knowledge in password modeling, and what it reveals about how humans create secrets, remains largely underexplored. Here, we address this gap with PassGPT+, which adapts the linguistic prior of GPT-2 to password observations through character-aware tokenization. We also introduce PassDiffusion, the first absorbing-state discrete diffusion model for password generation, as a probe of whether non-autoregressive approaches are competitive. On the RockYou benchmark, PassGPT+ recovers 22.53% of held-out passwords at 108 guesses, a 16% relative gain over PassGPT, and retains 79% of this match rate when transferred without retraining to a disjoint 2020 leak dataset, demonstrating that linguistic priors capture persistent regularities of human password generation. PassDiffusion underperforms by two to three orders of magnitude, indicating that autoregressive modeling is substantially better matched than iterative denoising to the discrete, exact-match nature of password generation. --- ### Algorithmic Recourse Under Competition Authors: Shahin Jabbari Published: 2026-09-30 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.39877v1 Algorithmic recourse provides individuals who have received undesirable outcomes from machine learning models with suggestions for minimum-cost improvements to achieve the desired outcome. A central assumption when computing recourse is that the decision rule remains fixed throughout the recourse implementation phase. We challenge this assumption in settings where individuals compete for limited resources. In such settings, widespread recourse implementation can change the acceptance threshold even when the scoring model that is used to evaluate individuals remains the same. This change in acceptance threshold can, in turn, invalidate the original recourse recommendations (i.e., following the recourse may not lead to the desired outcome). To address this problem, we introduce a framework called recourse under competition that jointly optimizes for recommendation recipients and the recommended score target they need to satisfy to balance the recourse cost and post-shift validity among initially rejected individuals. We develop an algorithm based on the Implicit Function Theorem and empirically analyze its performance. Experiments on synthetic and real datasets show that personalized score targets can achieve higher validity, albeit at a higher cost. In contrast, common score targets generally offer favorable cost-validity trade-offs for lower to medium validity values. --- ### Hyperspectral Image Models: Technical Report Authors: Tanishq Rachamalla, Aryan Das, Srishti Kaushik, Swalpa Kumar Roy Published: 2026-09-30 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.39871v1 Hyperspectral remote sensing has advanced across diverse deep learning paradigms, including spectral spatial CNNs, Vision Transformers, Mamba, graph neural networks, Kolmogorov Arnold networks, and self supervised masked autoencoding. Yet progress remains hindered by fragmented repositories, incompatible tensor conventions, and non standardized evaluation. Hyperspectral Image Models addresses these challenges through a modular framework unifying 55 representative models across six paradigms with a common registry, automatic 4D/5D tensor adaptation, and standardized constructors. It integrates 24 benchmark scenes from Airborne, Spaceborne, UAV, and Mars CRISM sensors, with caching, label remapping, PCA, explicit band selection or raw spectra, optional spatial max pooling, and arbitrary PxP patch extraction. To prevent inflated accuracy from overlapping windows, it supports class balanced random partitioning and spatially disjoint regional blocking with Chebyshev guard bands that eliminate train test pixel overlap. Experiments use a single config.yaml with deterministic seeds and complete provenance, generating LaTeX benchmark tables and classification maps. Across 1,320 model scene evaluations and 6,600 seeded runs, scene difficulty dominates architecture, with mean accuracy ranging from 96.40% on Botswana to 56.70% on Houston 2018, versus a 15 point spread across paradigm means. No paradigm universally dominates, while sub 1 M parameter models can match architectures two orders of magnitude larger. Code is publicly available at https://github.com/Tanishq251/Hyperspectral-Image-Models. --- ### GrammarRL: Effective Grammar-Constrained Decoding via Reinforcement Learning Authors: Gabriele Tuccio, Antonino Furnari, Aldo Gangemi, Misael Mongiovì Published: 2026-09-30 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.39869v1 Grammar-constrained generation guarantees syntactic validity, but can substantially degrade semantic quality when the model's preferred outputs are poorly aligned with the imposed grammar. This trade-off is particularly severe when the prompt is underspecified or the model has limited instruction-following ability. Beam search can partially mitigate these failures by exploring multiple valid sequences, but its computational cost grows with beam width, while sequence-level probability is only an imperfect proxy for semantic quality. We introduce GrammarRL, a label-free reinforcement learning method that adapts language models to grammar constraints without requiring annotated data. GrammarRL optimizes the model using two complementary self-supervised rewards derived from its own likelihoods: a direct reward, measuring how likely the constrained output is given the input, and a reverse reward, measuring how well the input can be reconstructed from the generated output. We optimize these rewards with a Reinforce Leave-One-Out (RLOO) objective over groups of grammar-constrained rollouts, augmented with the top-1 beam-search hypothesis and regularized towards a frozen base model. We evaluate GrammarRL on sign language gloss translation, hierarchical text classification, and named entity recognition using Llama models ranging from 1B to 8B parameters. GrammarRL consistently outperforms constrained greedy decoding, with an average improvement of 9.8 points and gains of up to 22.8 BLEU. It matches or outperforms beam search on two of the three tasks while preserving greedy-decoding inference cost. Ablations further show that the two rewards are complementary: either reward alone can underperform the untrained baseline, whereas their combination consistently improves upon it. --- ### Completion-Aware Cross-Fidelity Offline-to-Online Reinforcement Learning for Multi-Line Bus Holding Authors: Yifan Zhang, Qifan Zhang, Liang Zheng Published: 2026-09-30 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.39868v1 Exploratory reinforcement learning (RL) on an operating bus fleet is impractical,while policies trained only from historical data cannot acquire new experience. Hybrid Offline-and-Online (H2O) RL combines fixed target replay with simulator interaction, but the inexpensive online simulator can differ from the target in transition and event-duration dynamics. We study this cross-fidelity problem for multi-line bus holding and address a failure mode in which lower generalized passenger time coexists with incomplete passenger journeys. --- ### Preemptive LLM Unlearning against Forbidden Capability Acquisition via Gradient Sealing Authors: Kemou Li, Qizhou Wang, Yue Wang, Fengpeng Li, Zhuan Shi, Negar Rostamzadeh, Golnoosh Farnadi, Masashi Sugiyama, Jiantao Zhou Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39866v1 Open-weight LLMs are released not only as fixed products but also as substrates for downstream fine-tuning. This openness, however, creates legal and ethical risks because users may misuse fine-tuning to instill illicit knowledge or enable hostile operations. Model providers therefore need apre-release defense against such acquisition, motivating the problem of preemptive unlearning. Unlike retrospective unlearning, which removes capabilities already present in a fixed model, preemptive unlearning seeks to prevent their acquisition under unseen attack data and future fine-tuning procedures. Despite its practical importance, this setting remains largely unexplored, presents distinct challenges, and is therefore the central focus of our work. We first verify that existing retrospective methods provide insufficient pre-release protection. Even when forbidden capabilities are suppressed in current outputs, forbidden-domain data can still induce gradients through internal pathways, enabling later acquisition. Motivated by this finding, we propose a gradient-sealing principle that blocks these pathways by pushing relevant pre-activations into the negative region, where ReLU-family activations exhibit zero or near-zero derivatives. Experiments across multiple LLM families demonstrate our stronger resistance to downstream acquisition than retrospective baselines, validating gradient sealing as an effective mechanism for pre-release protection. --- ### FIGS: Evaluating Multi-Turn Sycophancy Without Penalizing Empathy Authors: Sidharth Pulipaka, Ruta Binkyte, Ivaxi Sheth, Sahar Abdelnabi Published: 2026-09-30 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.39863v1 Large language models frequently fail to balance staying truthful with being supportive. They often exhibit sycophancy in responses to users, agreeing with false claims, offering unwarranted flattery, and giving advice skewed toward users' expressed views. In reality, sycophancy rarely happens in a single exchange; it may emerge organically as users repeatedly insist or subtly steer the dialogue over time. Current evaluations, however, rely on rigid, single-turn tests or fixed scripts that fail to capture these natural dynamics. Furthermore, these benchmarks often mistake showing basic empathy for yielding, penalizing models for acknowledging a user's feeling. This view may drive future models to over-correct into cold, dismissive rigidity. To address this gap, we introduce FIGS (Factual Integrity and Grounded Support), a dual-axis evaluation framework built around extended, realistic dialogue. We use an adaptive 10-turn conversational simulator that dynamically challenges the target model, reflecting how users repeat requests, push back, or steer a conversation toward a preferred answer. To accurately evaluate these trajectories, we apply a taxonomy that strictly separates Sycophancy (whether the model holds firm to the truth and keeps its praise proportional) from Calibrated Validation (showing empathetic understanding of the user's feelings without overdoing it). We release our complete testing environment, including 500 diverse multi-turn scenarios and an automated judge. Our evaluation of leading models reveals a consistent trade-off: over the course of a sustained interaction, current systems either slowly drift to sycophancy or over-correct into robotic detachment. This demonstrates that balancing honesty with appropriate support throughout a natural conversation remains a critical, unsolved challenge. --- ### Fork-dLLM: Avoiding the Flexibility Trap in Diffusion Language Models Authors: Stipe Frković, Metod Jazbec, Christian A. Naesseth Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39859v1 Masked diffusion language models (dLLMs) have shown strong potential for faster inference through parallel token generation when combined with confidence-based samplers. However, recent work has shown that such methods can defer unmasking high-entropy fork positions at which multiple plausible continuations exist. This results in reduced generation diversity, as shown by worse pass@k scaling, and limits gains obtainable from RL post-training. To avoid this flexibility trap, prior work advocated for autoregressive (AR) sampling. Here, we show that discarding confidence-based sampling is unnecessary and, once inference cost is taken into account, wasteful. We first propose Fork-dLLM, a simple hybrid sampler that uses AR-style ordering only at uncertain fallback steps while retaining parallel generation otherwise. We then extend the same principle to post-training with ForkGRPO, which uses Fork-dLLM rollouts and applies the GRPO objective only at fallback steps, preserving exact policy-likelihood ratios while substantially reducing rollout and optimization cost. In our experiments, Fork-dLLM matches the strong pass@k scaling of AR sampling while being 2-3x more efficient, and ForkGRPO achieves downstream performance comparable to or better than AR-based GRPO baselines at a substantially lower training cost. --- ### Dimension-Free Rank Lifting from Random Hyperplane Arrangements Authors: Luca Becchetti, Matteo Russo, Ruben Skorupinski Published: 2026-09-30 Categories: cs.LG, math.PR arXiv: https://arxiv.org/abs/2609.39855v1 We study the width required for a randomly initialized hidden layer of a neural network to achieve rank lifting. Namely, given a dataset $X \in \mathbb{R}^{m \times d}$ of $m$, $d$-dimensional input vectors separated by an angle of at least $θ$, we consider the random feature matrix $σ(XR)$, where $R$ is standard Gaussian. For positively homogeneous nonpolynomial activations, which include sign, Heaviside, ReLU, and ReLU powers among others, we prove that $$n \gtrsim \frac{1}θ\max\left\{m,\log\left(\frac{1}δ\right)\right\}$$ neurons suffice for $σ(XR)$ to have full row rank $m$ with probability at least $1-δ$. This dimension-free bound exponentially improves the previous general-dimensional guarantee for sign features (Drago et al., 2026) and is essentially tight. The proof shows that one random feature column escapes every proper subspace of $\mathbb{R}^m$ with probability $Ω(θ)$, using a coupling of nearby Gaussian directions and a local crossing of the induced hyperplane arrangement. We also study stable rank lifting, where the goal is to establish a quantitative analogue of exact rank lifting, i.e., a lower bound on the smallest eigenvalue of the empirical feature Gram matrix in high-probability. Our analysis unifies and generalizes stable rank guarantees for all $q$-homogeneous non-polynomial activations following prior work in Panigrahi et al. (2020) and Song (2026). In particular, we combine a diagonally dominant Taylor tail of the population kernel with truncation and matrix concentration, to show that for positively homogeneous nonpolynomial activations, stable rank lifting is achieved at width $$n \gtrsim C^q \frac{m}{θ^{2q+1}} \log^{2q+\frac{1}{2}}\left(\frac{m}θ\right) \log\left(\frac{m}δ\right),$$ where $q$ is the degree of the activation and $C > 0$ is some universal constant. --- ### Cognitive Enhancement: Rethinking the Necessity of Role-Playing for Large Language Models Authors: Xingjie Zhuang, Jialong Tang, Chulun Zhou, Buchao Zhan, Zhirui Li, Junhui Li, Yazheng Yang, Jinsong Su Published: 2026-09-30 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.39853v1 Role-playing prompting has become a popular yet simple technique for improving LLM reasoning and output quality. However, whether it consistently boosts performance across diverse domains remains unclear, as systematic validation is lacking. To fill this gap, we run multi-model, cross-domain, and multilingual experiments on MMLU and MMLU-Redux. We find that gains from role-play prompting depend heavily on model capacity, knowledge domain, and prompt language. Drawing on metacognition theory, we propose the persona-related cognitive alignment hypothesis: role-play works only when the LLM correctly grasps the designated persona and its associated knowledge domain. We test this hypothesis through persona information richness ablation, layer-wise entropy divergence analysis, and latent thought-space deflection observation. To reduce persona cognitive bias and stabilize role-play performance, we propose \textbf{M}ixed-\textbf{L}anguage \textbf{C}oncatenate \textbf{P}rediction \textbf{(MLCP}), a simple, training-free, and efficient multilingual prompt concatenation strategy. It aggregates semantically equivalent role prompts to enrich complementary representational cues. Extensive experiments show that MLCP consistently outperforms vanilla role-play prompting across all tested LLMs. --- ### Predicting Multi-View Rashomon Representation: Can We Learn Where Models Disagree? Authors: Mingyue Ma, Zongbo Han, Changqing Zhang, Guangyu Wang Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39848v1 Foundation models are increasingly adopted across a wide range of applications, often serving as core blocks within AI systems. Yet different foundation models may encode the same input from multiple different views, leading to substantial representation disagreement, which we term Rashomon Representation. Such disagreement often signals inputs that a given model encodes in a way inconsistent with other models, offering a valuable yet underexplored signal for input reliability estimation. While prior work has largely focused on measuring disagreement across multiple models with a representation set, we instead focus on predicting disagreement from a single representation. We hypothesize that this disagreement follows some consistent, input-dependent patterns rather than occurring at random. To test this, we quantify disagreement by comparing each sample's nearest neighbors across different models' representation spaces, then train a lightweight predictor that estimates disagreement from a single model's representation. At inference time, given a new input, the predictor uses that input's representation to tell whether it aligns with or diverges from those of other models. Extensive experiments across diverse foundation models and datasets show that representational disagreement is indeed input-dependent, predictable, and generalizable, enabling efficient reliability estimation of foundation models. --- ### SEAR: Spoofing Evidence-Grounded Audio Reasoning Benchmark for Audio Language Models Authors: Rong Wan, Suliu Qin, Jiaxi Li, Wei Xie, Wenwu Wang, Xiaolong Han, Lu Yin, Xilu Wang Published: 2026-09-30 Categories: cs.SD, cs.LG arXiv: https://arxiv.org/abs/2609.39847v1 Audio language models (ALMs) are increasingly used for audio deepfake detection (ADD), yet existing benchmarks assess their verdicts or rationale plausibility without verifying the underlying acoustic evidence. To address this issue, we first introduce spoofing evidence-grounded audio reasoning (SEAR), a four-task AQA benchmark to evaluate ALM-based ADD through acoustic evidence identification and quantification, deepfake detection, and forensic rationale generation. We further propose a bona-fide-based acoustic evidence agent (BAEA), which equips a frozen ALM with controlled acoustic tools under \textsc{fixed} or \textsc{adaptive} evidence-acquisition policies. Experiments with six ALMs reveal a clear gap between plausible rationales and verifiable acoustic evidence reasoning, while BAEA-\textsc{Fixed} improves final verdicts and forensic rationales on both evaluation partitions. Controlled interventions further show that misleading evidence degrades both detection and grounding performance. --- ### When a Kindergartener Solves Calculus: Measuring Capability Leakage in Role-Prompted Reasoning Models Authors: Pakhapoom Sarapat, Saksorn Ruangtanusak, Kunat Pipatanakul, Pittawat Taveekitworachai Published: 2026-09-30 Categories: cs.CL, cs.AI, cs.HC arXiv: https://arxiv.org/abs/2609.39846v1 We investigate the problem of role-capability leakage (RCL), in which a role-prompted reasoning model generates convincing in-role text while continuing to exhibit capabilities on benchmarks that exceed those implied by the assigned role. For example, when a model is prompted to assume the role of a kindergarten student, one might expect its performance on a mathematics benchmark to reflect kindergarten-level ability rather than expert-level proficiency in solving calculus problems. We introduce RoleCapBench, a curriculum-grounded benchmark for evaluating RCL across six educational roles and four assessment levels spanning elementary school through A-level, and use it to evaluate three open-weight reasoning models. We find that although the models can generate stylistically convincing in-role responses, they consistently fail to align their underlying capabilities with their assigned roles. Naive role prompting yields strong role-voice scores of 1.218--1.389 while retaining above-role accuracy of 0.811--0.898. RCL persists across a range of prompting conditions, including prompts that explicitly instruct the model to match the role's capability level. To mitigate this problem, we propose Injection, an inference-time intervention that combines explicit, role-specific capability guidelines with a guiding prefilled response prefix. Injection improves role-capability alignment across models, reducing above-role accuracy by up to 0.562 while preserving in-role accuracy with a marginal drop of less than 0.058 across most models. All artifacts, including scripts and evaluation data, will be released upon acceptance. --- ### BayesNDE: Bayesian Generative Modeling for Neural Density Estimation Authors: Chenglin Li, Qiao Liu Published: 2026-09-30 Categories: stat.ML, cs.AI, cs.LG, stat.ME arXiv: https://arxiv.org/abs/2609.39843v1 Density estimation is a fundamental problem in statistics and machine learning. In this work, we introduce BayesNDE, a neural density estimator based on Bayesian generative modeling. BayesNDE learns a Bayesian generative model and evaluates its density without requiring invertible networks or Jacobian-determinant computation. For each observation, it infers a sample-specific latent posterior to construct an adaptive proposal that focuses computation on regions contributing most to its density. Bridge sampling then combines samples from this proposal with separate posterior samples to estimate the density. Experiments on nonlinear and multimodal synthetic datasets show improved estimation of density values and better recovery of the density structure compared to the state-of-the-art neural density estimators. Applications to real-world datasets further demonstrate improved anomaly detection. Together, these results highlight BayesNDE as a flexible and effective neural density estimator, demonstrating how posterior inference can turn generative models into tools for density estimation. The code and tutorials are available at https://github.com/liuq-lab/BayesNDE. --- ### DyRAD: Radar Novel View Synthesis for Dynamic Driving Scenes Authors: Merav Keidar, Tomer Borreda, Rajalakshmi Nandakumar, Or Litany Published: 2026-09-30 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.39841v1 Reconstructing dynamic driving scenes from recorded sensor data supports closed-loop evaluation of autonomous driving systems by synthesizing observations beyond the original trajectory. Unlike cameras and LiDAR, radar measures radial velocity directly through Doppler. Yet existing radar novel-view synthesis fails to exploit this capability: methods addressing dynamic scenes reconstruct only range-azimuth tensors, while methods that render Doppler assume static scenes. Moreover, because radar processing spreads each reflection across multiple bins, existing representations absorb this spread into scene geometry, causing it to render incorrectly when the viewpoint moves. We present DyRAD, which models dynamic driving scenes using static background reflectors and motion-tracked dynamic point reflectors to render complete range-azimuth-Doppler (RAD) tensors. Reflector velocities are derived from object tracks and projected onto the line of sight, making Doppler both a rendered output and supervision for those tracks. Crucially, we render reflectors through a fixed analytic point-spread function (PSF) derived from the radar's signal-processing chain, preventing sensor-induced spread from being baked into the scene representation. Beyond improving scene reconstruction, this separation also enables zero-shot sensor-configuration transfer, allowing the same reconstructed scene to be rendered under different radar specifications without refitting. We evaluate DyRAD on RADIal, Boreas, and a synthetic benchmark across both on-path poses and displaced viewpoints untested by prior work. On RADIal, DyRAD recovers radar detections in 90.7% of reference-detected objects, compared with 26.9% for the strongest baseline. --- ### Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning Authors: Hongwei Zhao, Rui Liu, Yansong Liu Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39839v1 Class-Incremental Learning (CIL) aims to continuously learn new classes without forgetting previously acquired knowledge. Parameter-efficient fine-tuning with pre-trained models reduces parameter overhead but can suffer from cumulative interference and suboptimal alignment between inference samples and specialized modules. We propose Dynamic LoRA-Experts and Prototype-Ensemble Matching (DLEPEM), a two-stage rehearsal-free framework. DLEPEM allocates a task-specific LoRA-Expert for each incremental task to reduce cross-task interference, then combines frozen pre-trained-model prototypes with task-adaptive LoRA-Expert prototypes for reliable task-level discrimination. Experiments on standard CIL and Few-Shot CIL benchmarks demonstrate strong performance under the evaluated protocols. --- ### Learning Steganography Is Easy, Learning Steganographic Reasoning Is Hard Authors: Julian Schulz, Lukas Fülle, Rieke Fruengel Published: 2026-09-30 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.39838v1 Chain-of-thought monitoring as an approach for AI oversight and control is threatened by the possibility of steganographic reasoning, where LLMs conceal their reasoning inside innocuous-looking text. Two neighbouring capabilities, steganographic messaging (passing a concealed message) and encoded reasoning (reasoning in an illegible but unconcealed format), have already been shown to emerge under training pressures that occur in real pipelines, such as reinforcement learning against monitors. This suggests that steganographic reasoning too might arise as an unintended side effect of training. Here, we compare how easily models learn steganographic reasoning and these two neighbouring capabilities across three elicitation methods: reinforcement learning, in-context learning, and supervised fine-tuning (SFT). For most tasks, models learn steganographic reasoning only under SFT, while they learn steganographic messaging and encoded reasoning under all three elicitation methods. Even under SFT, steganographic reasoning requires at least twice as much training as messaging, and for several model-task combinations it is not learned at all. However, on a cover task that makes hiding information especially convenient, steganographic reasoning can be successfully learned under all three elicitation methods. Steganographic reasoning is thus much harder than steganographic messaging and encoded reasoning, and learning the latter two does not imply learning the former. Yet it lies within reach: an easy version is learned under every elicitation method, when the cover task is convenient for hiding information. --- ### Fast Regularized Policy Mirror Descent with One-Step TD Updates Authors: Qipei Chen, Wenye Li, Yule Sun, Ke Wei Published: 2026-09-30 Categories: cs.LG, math.OC arXiv: https://arxiv.org/abs/2609.39837v1 Policy mirror descent (PMD) enjoys fast convergence in regularized Markov decision processes (MDPs), but existing guarantees often rely on exact or increasingly accurate policy evaluation. We analyze PMD coupled with a persistent critic advanced by one temporal-difference (TD) update. For finite discounted MDPs, we establish global linear convergence in value for exact coordinate-wise Bellman updates, with any positive constant actor stepsize and arbitrary finite critic initialization. The proof combines a resolvent-based auxiliary distribution with a decaying Bellman-violation correction and a potential weighted by inverse coordinate weights. We then study stochastic TD-PMD with general strongly convex mirror maps under a single off-policy Markov trajectory. With suitably chosen constant stepsizes and a finite-batch TD update, the method achieves an expected value gap of $ε$ after $\widetilde{O}(1/((1-γ)^5 \widetildeσ_b ε))$ transitions. The stochastic analysis relies on the trajectory-wise Lipschitz continuity of the regularizer, derived from uniform bounds on vertex Bregman divergences, together with a visitation-weighted resolvent estimate for signed critic-error propagation that yields an inverse-linear dependence on behavior coverage $\widetildeσ_b$. In contrast to many prior guarantees for regularized policy optimization, our sample-complexity guarantee holds without trajectory resets, generative-model access, or nested policy-evaluation loops. Numerical results are consistent with the theoretical convergence analysis. --- ### Spherical Interpolation for Backward-Compatible Multimodal Representations Authors: Simone Ricci, Niccolò Biondi, Federico Pernici Published: 2026-09-30 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.39836v1 Contrastive vision-language models map visual and textual representations into a shared normalized embedding space, making cosine similarity the natural metric for cross-modal retrieval. A practical challenge arises during model upgrades: independently trained models generally produce incompatible representation spaces, so replacing a deployed model typically requires recomputing embeddings for the entire gallery, which is prohibitively expensive at scale. Orthogonal post-hoc alignment can partially mitigate this problem by mapping new-model queries into the old-model gallery space. However, because independently trained models can differ in fine-grained representation structure, the orthogonal alignment remains approximate, leaving a residual angular discrepancy between the old-model query and the aligned new-model query. We study whether interpolation along the spherical geodesic between these two normalized query representations can improve retrieval without re-indexing the gallery. We characterize when this path contains an interior query direction closer to an idealized retrieval-optimal direction than either endpoint, and connect this characterization to Recall@$K$ through a local margin-based certification result. Experiments across multiple benchmarks and model families show that post-alignment spherical interpolation improves over orthogonal alignment alone, recovering backward-compatibility in most evaluated settings. Consistent with our geometric characterization, per-query oracle analysis shows that retrieval-favorable interior points occur frequently in practice. Code is available at https://github.com/miccunifi/SLERP_backward_compatibility . --- ### RainAtlas: A Multi-Continental Dataset for Precipitation Downscaling Authors: Pierre-Louis Lemaire, Luca Schmidt, Wietze Suijker, Alex Hernandez-Garcia, David Rolnick Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39833v1 Extreme rainfall events are increasing in intensity and frequency as climate change accelerates. While kilometer-scale precipitation forecasts are critical for supporting local decision-making, the limited availability of high-resolution precipitation observations hinders their accuracy, especially in under-resourced regions. Machine learning models are widely used to downscale precipitation data to km-scale, but their application to unseen geographies presents challenges. First, processing raw high-resolution precipitation datasets across regions requires significant engineering and domain expertise. Second, generalization across regions remains difficult. To help overcome these barriers, we release RainAtlas, a large-scale, ML-ready and multi-continental dataset for precipitation downscaling. Covering three continents, RainAtlas harmonizes heterogeneous hourly km-scale observations to a common 2-km grid. Each regional partition contains around 210,000 aligned low- and high-resolution precipitation pairs, respectively from ERA5 reanalysis and direct observations. We benchmark state-of-the-art ML-based downscaling models across RainAtlas using a wide range of metrics. Our evaluation reveals substantial variance in out-of-domain generalization depending on the training regions. This underscores the need for cross-regional, multi-source km-scale evaluation, establishing RainAtlas as a well-positioned benchmark for precipitation downscaling research. --- ### P-SRM: Selective Recovery of Rejected Predictions in Visual Tracking Authors: Youbin He, Siwei Wang Published: 2026-09-30 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.39832v1 Many visual tracking methods use rejection mechanisms to suppress unreliable predictions. However, these mechanisms can also reject correctly localized candidates, leaving useful information unused. We investigate how to identify and recover these candidates while preserving native accepted outputs and candidate coordinates. To this end, we propose P-SRM (Post-rejection Selective Recovery Method), which combines spatial responses, past accepted states, and native decision margins to reassess candidates and selectively restore reliable predictions. We evaluate P-SRM on six trackers and four datasets spanning category-specific, point, and generic object tracking. Across all nine configurations, P-SRM improves rejected-candidate ranking and overall tracking performance. These results show that post-rejection verification can identify and recover useful predictions discarded by native rejection, demonstrating the value of reusing rejected information. Project repository: https://github.com/PalestyHR/P-SRM. --- ### Estimation of the Label-Noise Transition Matrix with Performance Guarantees via Selective Classification Authors: Xabier de Juan, Santiago Mazuelas, Yilun Zhu, Clayton Scott Published: 2026-09-30 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.39829v1 Modern machine learning depends heavily on massive datasets, but obtaining high-quality annotations at scale is often expensive. As a result, learning from noisily-labeled data has become common, making accurate estimation of the label-noise transition matrix crucial. However, existing transition matrix estimators rely on the fragile estimation of class-posteriors and do not provide finite-sample performance guarantees. In this work, we propose a novel methodology to estimate the transition matrix based on one-sided selective classification. This approach bypasses class-posterior estimation, provides finite-sample performance guarantees, and leverages flexible learning methods for binary classification. Moreover, we introduce effective algorithms to implement the proposed methodology and provide their refined finite-sample performance bounds. --- ### Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior Authors: Atsuki Yamaguchi, Tatsuro Inaba, Joel Niklaus, Michal Štefánik, Aline Villavicencio, Nikolaos Aletras Published: 2026-09-30 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.39827v1 Pre-pretraining (PPT) on synthetic non-natural language data improves token efficiency during language model pre-training (PT). Prior work attributes this gain to a grammatical prior, i.e., a structural inductive bias learned during PPT that transfers to natural language grammar. However, PPT has only been tested on models of at most 1B parameters and PT budgets below 2B tokens on predominantly web text. It is unknown whether PPT is effective at larger scales and under more realistic PT data mixtures that combine diverse sources (e.g., code and math). We therefore present a comprehensive study on PPT spanning five PPT tasks, four PT data mixtures, four parameter scales (500M to 7B), and PT budgets of up to 100B tokens. Our results demonstrate that the downstream performance and token efficiency gains of PPT persist at scale, e.g., saving at least 21B PT tokens at the 3B scale. However, in contrast to prior work, we find no consistent evidence that these gains stem from a grammatical prior. Downstream performance does not consistently align with grammatical acceptability across model sizes. Instead, we find that downstream gains arise from PPT tasks that improve long-range retrieval. Finally, PPT performance gains are robust to how PT data mixtures are composed and diminish only when web text is absent. Overall, PPT is a low-cost addition to PT, and future PPT task design should target long-range retrieval rather than natural language grammar. --- ### Learning from Runtime Feedback through Failure-Bank Self-Evolution for Vision-Language-Action Models Authors: Mingyue Cui, Zheyuan Liu, Yihan Zhu, Zheyuan Zhang, Meng Jiang Published: 2026-09-30 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2609.39820v1 Vision-language-action (VLA) models generalize broadly across robotic manipulation tasks, but complex environments require balancing task success with unintended contact. Runtime shields can correct individual actions, but they leave the underlying policy unchanged, so repeated disagreements may create a persistent policy-shield mismatch that blocks task progress. To address this challenge, we introduce FailBank, a four-stage self-evolving framework that converts runtime feedback into persistent policy improvement. During collection, a fixed CBF-based safety module serves as an observe-only teacher, producing counterfactual corrections while the policy remains in control. Outcome-aware admission then converts useful proposals into corrective targets and retains successful uncorrected actions as quiet anchors for guarded LoRA updates. We evaluate FailBank on the VLA-Arena benchmark across two difficulty levels and two VLA backbones. Compared with the base policies, FailBank improves the joint success-cost operating point. Across the two backbones, FailBank improves task success rate by 8.5 and 6.9 percentage points, while reducing policy-induced cumulative cost by 35.6\% and 23.8\%, respectively. Compared with runtime shielding, FailBank raises task success rate by 25.4 and 9.5 percentage points, while maintaining comparable policy-induced cumulative cost. These results show that runtime feedback can serve as persistent policy supervision rather than only as a temporary action constraint. --- ### Should I stay or should I show? Learning to selectively disclose information Authors: Carlotta Giacchetta, Alessando Bogani, Cesare Barbera, Giovanni De Toni, Michele Caprio, Andrea Pugnana, Andrea Passerini Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39818v1 In many high-stakes settings, human decision-makers can acquire support information before making a decision. However, acquiring information is costly, and disclosure may fail to improve human decisions or may even impair them. We tackle this problem by studying selective disclosure, i.e., the problem of learning when to reveal support information to a human decision-maker under a budget constraint. We first show that the optimal policy is a threshold rule on the Value of Information (VoI), i.e., the expected reduction in human decision risk induced by disclosure. Since VoI is unknown in practice, we estimate the regime-specific human risks and bound the possible degradation of the resulting plug-in policy relative to lack of disclosure, as well as its regret relative to the optimal policy. Experiments on benchmark datasets show that selective disclosure outperforms both no disclosure and full disclosure, regardless of whether the support information is beneficial or harmful. Two user studies show that human-AI team performance can improve when disclosure is led by our learned policy and not human-selected, although this advantage varies across tasks. A counterfactual benchmark, which replaces participants' predictions with a machine-learning prediction when disclosure occurs, suggests that these differences might depend on lower adherence to advice when the information is automatically provided rather than self-requested. --- ### Beyond Accuracy: Prefix-Invariant Realizations of Low-Precision Fast Matrix Multiplication Authors: Shuxiao Xie, Shuyang Xie, Yuan Cao, Dezhi Ran, Wei Yang, Tao Xie Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39816v1 Fast matrix multiplication saves multiplications through exact cancellation, but rounding sums that mix token rows can leave contributions from later tokens in earlier language model outputs. This threatens prefix invariance, which multiple-choice likelihood scoring relies on: a scored likelihood must depend only on its allowed prefix. On Qwen2.5-14B-Instruct, two fast FP8 realizations repaired to ordinary-looking accuracy still change the answers chosen by likelihood on 5.83% and 10.00% of 240 OpenBookQA items when only the text after the allowed prefix is replaced with the bf16 model's own greedy continuation. Both row-local controls, the bf16 model and a deployed FP8 matrix multiplication kernel, change none. Accuracy thus does not certify prefix invariance, and the stability criteria we analyze cannot tell realizations apart: across all 512 sign variants of two-level Strassen they stay constant while teacher-forced perplexities span a 772.4$\times$ range on the same model. We therefore construct certified realizations of two-level Strassen on bounded integer codes that quantize token rows independently, then mix and cancel exactly before rescaling, using 49 block multiplications instead of 64. Our certificate guarantees bitwise equality to a prescribed row-local classical int8 operator at the same quantization specification, so every certified realization inherits its prefix invariance. Certification thus turns realization choice into a pure cost decision: which certified realization runs can no longer change a single scored likelihood. --- ### Backward-State Policy Is Part of the Learning Algorithm Authors: Shuxiao Xie, Shuyang Xie, Dezhi Ran, Wei Yang, Tao Xie Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39813v1 Low-precision training rounds tensors that the backward pass reads again, often for several gradients; each use can read the forward's rounded value, the original, or a new random rounding. This backward-state policy looks like a memory and precision detail, settled by copy accuracy and final loss. We argue that it is part of the learning algorithm, and that neither check shows whether it is right. Copy accuracy does not decide the outcome: in three pairs of 390M runs with an emulated FP8 backward, training fails when attention's backward reuses the forward's rounded output and succeeds with a new rounding from the same distribution. Even the most accurate copy, the original itself, can be wrong by our reference: the gradient of the forward pass as it actually ran, with gradients passed through rounding unchanged. For example, a normalization output stored in low precision feeds two gradients: the gain's gradient needs the original, but the next layer's weight gradient needs the rounded value that layer multiplied. Final loss, the other check, does not rule out the error of reading the original for both: it persists in models trained with such a store, while planned loss comparisons stay within a margin fixed in advance. We therefore derive from this reference which value each use must read, or which substitute gives the same gradient on average with the forward held fixed, and check these per-use requirements on single operators, without training. In three tests using PyTorch and Transformer Engine, the requirements predicted beforehand whether reuse changes what the backward computes on average relative to an independent copy, and every prediction held. Backward-state policy is thus part of the learning algorithm: it should be specified and checked use by use, not settled by copy accuracy and final loss. --- ### A Comprehensive Benchmark of Source-Free Universal Domain Adaptation on Time Series Representations Authors: Romain Mussard, Fannia Pacheco, Maxime Berar, Paul Honeine, Gilles Gasso Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39810v1 Source-Free Universal Domain Adaptation (SF-UniDA) extends Universal Domain Adaptation by removing access to source data at adaptation time while still handling label-set mismatches between domains. Despite growing interest in this setting for image data, no benchmark exists for time series, which are more challenging. We present the first SF-UniDA benchmark on time series. In addition, we provide the first study of pretrained foundation models as feature extractors for time series domain adaptation. In this context, we identify a critical and previously underexplored limitation of all existing SF-UniDA methods: the inference threshold for unknown-sample rejection is highly sensitive. We address this by proposing a plug-in auto-thresholding module that can be integrated into any SF-UniDA method. Experiments on three well-known time series datasets confirm the suitability of this module. They also highlight that foundation models do not systematically outperform classical backbones and that SF-UniDA tailored for time series is yet to be developed. --- ### Stress-Testing LLM Lie Detectors: Role-Play Failures and Spurious Correlations Authors: Maximilian von Klinski, Sebastian Lapuschkin, Wojciech Samek, Lennart Bürger Published: 2026-09-30 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.39807v1 Lie detection probes aim to predict from a language model's internal states whether its output is truthful or dishonest. However, role-play complicates what "truth" means for an LLM: language models can adopt a wide range of personas that take very different claims to be true, including personas whose beliefs clearly contradict reality, such as a conspiracy theorist. In this work, we investigate whether lie detection probes reliably flag falsehoods generated under such an anti-factual persona or whether they instead follow the persona's beliefs. We introduce a dataset of 8,916 human-reviewed, on-policy responses from three LLMs adopting anti-factual personas. Evaluating eight probes from prior work, we find that many fail in this setting, particularly when correct and incorrect answers are evaluated under the same persona prompt. To investigate why, we construct three novel confounder datasets in which truth is anti-correlated with a potential confounding concept. Our experiments reveal that many existing probes strongly track concepts that are spuriously correlated with truth in their training data, such as instruction compliance or response likelihood. Based on these findings, we introduce a simple linear probe that achieves the strongest overall performance on both the persona and confounder stress tests. Our results suggest that current lie detection probes are far from reliable and highlight the need for training data in which truth is decorrelated from confounding concepts. --- ### RATIO: Reasoning Analysis and Token-level Inference Optimization for Quantized Reasoning Models Authors: Chengzhu Bao, Xianglong Yan, Tianao Zhang, Jiaqi Chen, Shaoqiu Zhang, Yulun Zhang Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39801v1 Post-training quantization (PTQ) has become a widely adopted technique for reducing the memory footprint and inference cost of large language models (LLMs). However, recent studies reveal that when applied to reasoning models, PTQ not only degrades reasoning performance but also exacerbates overthinking, leading to longer reasoning trajectories. These issues may offset the efficiency gains expected from lower-precision inference. Existing approaches mainly rely on complex optimization procedures. More recent lightweight inference strategies instead use predefined overthinking markers, limiting their adaptability across quantized models. To address these issues, we propose Reasoning Analysis and Token-level Inference Optimization (RATIO), a framework that identifies model-specific overthinking tokens and assigns each a tailored penalty. RATIO first introduces Quantization-aware Reasoning Behavior Analysis (QRBA) to identify overthinking tokens by analyzing discrepancies between full-precision and quantized models. It then adopts Token-Specific Penalty Determination (TSPD), which leverages full-precision guidance to derive token-specific penalties without additional training. Extensive experiments show that RATIO achieves a better accuracy-efficiency trade-off than existing token-level interventions. Specifically, RATIO achieves up to 9.8 points accuracy improvement and reduces chain-of-thought (CoT) length by up to 51.3% compared with quantized baselines. The code will be available at https://github.com/steven-bao1/RATIO. --- ### Finite-Horizon Fisher Memory in Two-Sided Power-Bounded Recurrent Systems Authors: Jeonghoon Lee Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39800v1 We analyse allocation, admission and post-write retention in finite-horizon linear-Gaussian noisy recurrent memories. At every horizon, the directional Fisher memory $M_n$ satisfies $\operatorname{tr}M_n=N$: non-normality redistributes information but cannot raise its spherical average, while normal carriers satisfy $M_n=I$. For bi-power-bounded carriers, we derive uniform $1/n$ lag bounds, identify the limit of $M_n$ with the inverse of the classical Cesàro asymptotic limit of $W^\top$, and give finite-horizon error bounds. A time-varying coupling defines an end-to-end store operator. The writer-optimal direction need not be store-optimal. After writing ends, an invertible hold preserves the full stored Fisher matrix. Additive contamination bounded by $α$ times the closure covariance retains at least $1/(1+α)$ of that matrix; a covariance-aware decoder attains the corresponding accuracy. With recurrent carriers held fixed, training input masks and linear readouts approached the task-specific optimum in 160 runs, with median normalized Rayleigh efficiency above $0.998$. Binary accuracy matched the Gaussian prediction to mean absolute error below $0.002$ over more than four orders of magnitude in $J$. In a separate pre-specified study of 320 runs, trained masks followed the designated input-time objective in both carrier types, in 16 of 16 draws. These studies used development-seen carriers and are pre-specified validations, not blind holdouts. The same fixed design reproduced the objective-specific result in 16 of 16 draws on carriers unused before run commitment. Exact isolation preserved information, while a decoder fixed at its training horizon fell to chance; inverse-adjoint transport restored its sampled decisions to numerical precision. --- ### Probabilistic Adversarial Training Authors: Andi Zhang, Xingyu Zhao, Siddartha Khastgir Published: 2026-09-30 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2609.39798v1 Building on a probabilistic perspective in which adversarial examples arise from the overlap between a distance-based distribution $p_{\mathrm{dis}}$ and a victim-classifier-induced distribution $p_{\mathrm{vic}}$, we start from a simple intuition: adversarial examples become harder to generate when these two distributions are pushed apart, as their overlap becomes smaller, thereby increasing robustness. This intuition naturally motivates a KL-based robustness objective. We then prove that $\mathrm{KL}(p_{\mathrm{dis}}\|p_{\mathrm{vic}})-\log Z_{\mathrm{vic}}$ is a lower bound on probabilistic robustness (PR), where $Z_{\mathrm{vic}}$ denotes the normalizing constant of $p_{\mathrm{vic}}$. Since PR is generally intractable to compute directly, maximizing this KL-based lower bound provides a tractable surrogate objective for improving PR. We further show that this objective recovers a scaled form of adversarial training, offering a probabilistic interpretation of adversarial training and a principled route to robustness improvement. We call the resulting method probabilistic adversarial training. Experiments show that it consistently improves PR, and ablation studies demonstrate that the induced scaling factor can even enhance the PR of non-probabilistic adversarial training methods. --- ### Inline Memory Meets Reusable Skills: Memory-centric Framework for Vision-Language-Action Model Authors: Zaijing Li, Rui Shao, Bing Hu, Haoyu Zhang, Dongmei Jiang, Liqiang Nie Published: 2026-09-30 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2609.39794v1 Vision-Language-Action (VLA) models have shown strong promise for general-purpose robotic manipulation, yet adapting them to new tasks and domains remains inefficient: existing methods often rely on parameter tuning, incurring substantial costs and risking catastrophic forgetting of previously learned tasks. To address this, we propose \textbf{Optimus-R}, a memory-centric VLA framework that formulates robotic adaptation as explicit query-skill memory tuning. Optimus-R introduces: (i) An \textbf{Inline Memory Interface for skill extraction}. It inserts learnable memory tokens into the VLA prefix stream, allowing the backbone to derive control-aware query and skill representations within the native action-conditioning pathway. (ii) A \textbf{Query-Skill Memory Bank for skill learning}. It externalizes skills into query prototypes for deciding \emph{what} to retrieve and skill values for specifying \emph{how} to act, supporting skill reuse and expansion with limited parameter updates. (iii) A lightweight \textbf{Bridge-and-Adapt mechanism for skill updating}. It aligns target-domain queries and skills with the existing memory space through a lightweight adapter and residual memory updates. Experiments on in-domain adaptation, cross-domain adaptation, and lifelong learning show that Optimus-R enables data-efficient skill learning while mitigating catastrophic forgetting. --- ### TopTimeNet: Topologically-assisted time-series classification model Authors: Sharareh Sayyad, Sophia Bazzi Published: 2026-09-30 Categories: cs.LG, cs.CG, math.DS, nlin.CD, physics.app-ph arXiv: https://arxiv.org/abs/2609.39792v1 Distinguishing periodic from chaotic dynamics in a time series is a fundamental challenge in both physics and engineering. Yet, end-to-end learned architectures must discover both a representation and a decision boundary from data, at substantial cost. We introduce TopTimeNet, which decouples these tasks: a fixed, non-learned stage extracts a $42$-dimensional geometric and topological descriptor from Takens delay embeddings and persistent homology, and a lightweight learnable stage performs classification. On a benchmark of $49$ nonlinear dynamical systems, a $1{,}638$-parameter configuration matches the mean accuracy of one with $33\times$ more trainable parameters. Additionally, this approach delivers mean accuracy comparable to convolutional neural networks and surpasses the average performance of converged Transformer models, while requiring three to four orders of magnitude fewer trainable parameters. Robustness also depends sharply on where noise is introduced: TopTimeNet degrades gracefully under perturbations to its precomputed features, but degrades sharply when noise is introduced into the raw signal and the full feature-extraction pipeline is recomputed, showing that robustness to perturbations of the precomputed features does not imply robustness of the complete raw-signal-to-prediction pipeline. These results show that decoupling fixed geometric and topological feature construction from a lightweight discriminative stage can achieve comparable classification accuracy with substantially fewer trainable parameters. --- ### Pseudo-Label-Triggered Retraining from Forecast Errors for Online Time Series Forecasting Authors: Yeryeong Kwak, Yoo-Min Jung, Jonghun Park Published: 2026-09-30 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.39789v1 Real-world time series forecasting systems operate under non-stationary data streams, where forecasting performance may degrade over time. Although retraining can recover the performance, it incurs non-trivial computational and operational costs. Under limited deployment resources, the key challenge is therefore not only how to retrain but also when to retrain. While existing retraining policies often rely on indirect indicators such as drift alarms or model staleness, we instead use realized forecast errors as direct deployment feedback. In this paper, we propose PILOT (Pseudo-label-Informed Learned Online Trigger), an online retraining framework that learns when to retrain from forecast-error dynamics. Since ground-truth retraining labels are unavailable, PILOT constructs a pseudo-label from future increases in forecast error and trains a lightweight scorer to predict it from observed error states. At deployment, PILOT uses only completed forecast errors and serves as a plug-in module for arbitrary forecasting backbones without architectural modification. We evaluate PILOT under standard multivariate forecasting settings across eight benchmarks with three representative backbones---DLinear, iTransformer, and TimesNet. Across all three backbones, PILOT achieves state-of-the-art average-rank performance among retraining policies while maintaining a favorable performance--efficiency trade-off. --- ### Safety of Latent Communication in Multi-Agent Systems Authors: Muhammad Huzaifa, Sina Mavali, Thorsten Eisenhofer Published: 2026-09-30 Categories: cs.AI, cs.LG, cs.MA arXiv: https://arxiv.org/abs/2609.39788v1 Latent communication enables multi-agent systems to exchange information directly in internal representation space, reducing the token, computation, and latency overhead of text-based communication. To this end, lightweight trainable links are introduced to map the sender's representations into the receiver's input space. In this work, we show that even benign link training can increase harmful compliance relative to text-based communication while the underlying safety-aligned agents remain unchanged. An attacker can amplify this effect by optimizing the links on harmful query--response pairs or poisoning otherwise benign training data. We further develop a reinforcement-learning attack that rewards harmful compliance alongside benign task performance without requiring harmful target responses. Across three communication topologies and four safety benchmarks, this attack raises the mean harmful-compliance score from 27.9 with benignly trained links to 76.9. Compared with direct supervised optimization, it also achieves higher average accuracy on two benign utility benchmarks. Adapting the rewards toward safer behavior also enables repair of compromised links, substantially reducing harmful compliance across all evaluated attacks without updating the agents. Overall, our results show that safety alignment requires considering the multi-agent system as a whole. --- ### Explore-on-Graph: Hybrid Embedding-LLM Reasoning for Knowledge Graph Question Answering under Incompleteness Authors: Ola El Khatib, Djellel Difallah Published: 2026-09-30 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.39786v1 Large language models (LLMs) are increasingly combined with knowledge graphs (KGs) to ground reasoning in structured evidence. However, most LLM-based KGQA methods rely on traversing existing graph edges and become unreliable when reasoning paths are broken by missing facts. Alternatives that ask LLMs to generate missing knowledge risk introducing hallucinated evidence. We introduce XoG (eXplore-on-Graph), a framework for multi-hop question answering over incomplete KGs that recovers missing reasoning paths from learned graph structure rather than LLM parametric knowledge. XoG combines type-level entity-relation statistics to identify candidate relations with KG embeddings to retrieve plausible missing entities, using the LLM as a semantic selector and reasoner. These mechanisms are integrated into an iterative planning-exploration-reasoning process. Experiments on WebQSP, CWQ, and the Wikidata-based BRINK benchmark show that XoG remains competitive on complete KGs and consistently outperforms comparable methods without task-specific KGQA training under KG incompleteness. These gains persist across multiple LLM backbones, indicating that stronger LLMs alone do not resolve missing graph evidence. XoG also reduces LLM token consumption by up to 33% compared with a closely related planning-based approach. --- ### Seeing as Humans Do: Learning from Motion to Segment Anything Without Supervision Authors: Weijian Jian, Xiaoyue Zhang, Bin Xiao, Chunyu Xie, Yixiao He, Yutao Liu, Dawei Leng, Yuhui Yin Published: 2026-09-30 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.39785v1 The Segment Anything Model (SAM) relies heavily on massive manual annotations, creating a fundamental bottleneck for model scaling. While unsupervised methods attempt to learn object concepts from motion, they typically overfit to moving entities, lacking both multi-granularity understanding and the ability to generalize to static objects. To overcome this, we introduce Motion-Grounded Segment Anything (MoSA), a highly scalable unsupervised framework that learns a transferable objectness prior from unlabeled videos. MoSA operates in three progressive stages: (1) automatically generating multi-granularity motion pseudo-labels from large-scale video data; (2) training a Perceptual Grouping Model (PGM) via contrastive learning to internalize a generalized, appearance-driven concept of objects; and (3) transferring this learned prior into a prompt-guided architecture for segment-anything-style inference on images. Extensive zero-shot evaluations across seven challenging benchmarks (e.g., COCO and ADE20K) demonstrate that MoSA significantly outperforms existing unsupervised methods. Notably, despite using zero manual annotations, MoSA achieves segmentation performance comparable to the fully supervised SAM. Our findings reveal that harnessing large-scale unlabeled motion is a feasible and highly scalable alternative to annotation-driven segment-anything pipelines. --- ### CORD: Learning Reusable Degradation Representations Across Heterogeneous Physical Systems Authors: Haibo Li, Zhiguo Zeng Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39784v1 Can heterogeneous physical degradation systems benefit from joint pretraining and move beyond system-specific prognostics toward reusable cross-system representation learning? CORD combines type-specific observation interfaces with a shared degradation backbone. Its two self-supervised objectives learn at complementary scales: Intra-Observation Structure Modeling (ISM) captures structure within observations, while Inter-Observation Dynamics Modeling (IDM) captures latent degradation evolution across observation histories. We evaluate CORD under two transfer boundaries: Pretraining-Included System Types, where downstream datasets and held-out units are unseen but their system types are represented during source pretraining, and Pretraining-Excluded System Types, where the entire turbofan-engine type is absent from pretraining. Across bearings, batteries, and cutting tools, CORD (Multi-domain) consistently improves over CORD (Single-domain) under Frozen adaptation, provides further gains under Full FT in most settings, and remains competitive with representative external baselines. Source-pretrained initialization also improves low-label adaptation to the pretraining-excluded engine type. Frozen-representation analysis further shows improved cross-unit lifecycle consistency after multi-domain pretraining. Joint pretraining across heterogeneous physical systems thus produces degradation representations reusable across devices, datasets, and system types. --- ### GraphMAS: A Systematic Benchmark of Multi-Agent Coordination for Graph Learning Authors: Jiayi Yang, Yifang Chen, Yuanfu Sun, Xinyan Ge, Qiaoyu Tan Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39777v1 LLM-based multi-agent systems coordinate specialized reasoning through aggregation, interaction, and adaptive control, yet their potential for graph learning remains unexplored. Graph learning is a natural setting for such systems because useful evidence may arise from heterogeneous local, long-range, global structural, and semantic perspectives whose relevance varies across instances. Existing LLM-based graph learning approaches primarily rely on single-agent reasoning, while multi-agent coordination has been studied mainly in general reasoning settings. Consequently, it remains unclear whether multiple specialized agents can improve graph learning and how coordination strategies should be designed and evaluated. To address this gap, we introduce GraphMAS, a systematic benchmark of multi-agent coordination for graph learning. GraphMAS builds a shared pool of graph reasoning specialists and organizes coordination along two dimensions, inter-agent interaction and runtime adaptivity, yielding four paradigms and seven representative coordination methods. Under a unified protocol, we evaluate these methods across seven text-attributed graphs, three domains, and two graph learning tasks. We find that heterogeneous graph perspectives are complementary, and that coordinating specialists improves over individual specialists and single-agent graph reasoning, with gains from decomposing reasoning across specialists rather than from broader evidence access alone. However, richer inter-agent interaction does not reliably help, whereas instance-adaptive specialist selection yields the strongest accuracy-efficiency trade-off. We further show that coordination can be learned over a fixed specialist pool and transfers to held-out graphs. GraphMAS therefore provides a controlled evaluation framework and empirical principles for understanding when and how multi-agent coordination benefits graph learning. --- ### Riemannian Flow Models with Reinforcement Learning for Molecular Crystal Structure Prediction Authors: Thomas Egg, Harry Winston Sullivan, Maya M. Martirossyan, Philipp Höllmer, Cheng Zeng, Adrian Roitberg, Mingjie Liu, Richard Hennig, Sapna Sarupria, Ellad B. Tadmor, Stefano Martiniani Published: 2026-09-30 Categories: cs.LG, physics.comp-ph arXiv: https://arxiv.org/abs/2609.39773v1 Crystal structure governs material properties, making crystal structure prediction (CSP) a fundamental problem in materials science. Generative models are a promising approach for solving this problem, but the prevalence of polymorphism, coupled with large unit cells and complex packing geometry, makes the molecular CSP task challenging for existing models. To address this, we introduce Coarse-Grained Open Materials Generation (CG-OMatG), an equivariant Riemannian flow-based generative model. CG-OMatG predicts molecular crystal structures \textit{via} a coarse-grained, hierarchical representation. CG-OMatG treats molecules as rigid bodies---performing both inter- and intra-molecular message passing to construct a geometric representation for molecular packings---and learns to reconstruct molecule centroid positions, orientations, and lattice parameters, conditioned on chemical species and conformer geometry. We train the model on subsets of the Open Molecular Crystals (OMC25) and Cambridge Structural Database (CSD) datasets. Further, we fine-tune the model \textit{via} policy gradient reinforcement learning to steer the model towards generating low-energy candidate structures. We validate the generated structures on the CSP blind test benchmark, assessing agreement with experimentally determined crystals using COMPACK packing-similarity analysis. CG-OMatG exhibits strong performance for generative molecular crystal structure prediction, paving the way for accelerated polymorph screening and organic solid-state materials discovery. --- ### Security Properties of Neural Networks as Decision Problems Authors: Adrian Wurm Published: 2026-09-30 Categories: cs.LO, cs.CC, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.39768v1 Certifying a deployed neural network raises decision problems that the verification literature has not classified: whether the model carries a backdoor planted in its training data, whether a fault in its stored parameters can drive it into an unsafe state, whether its output leaks a private part of its input. We formalise eight such problems and classify what we can. The organising observation is a logical one. The function computed by a piecewise linear network, together with all its node values, is definable by a quantifier-free formula of real addition of size linear in the network, so a property of the network is a quantifier-alternation sentence, which Sontag's 1985 theorem places in the polynomial hierarchy at the level of its prefix. Membership results are thus corollaries, and the argument makes plain what they need: that the quantified objects are inputs rather than the network's own parameters. Non-interference, monotonicity and counterfactual fairness have exactly the complexity of network equivalence and of interval verification, all co-NP- complete over ReLU. Detection of backdoor triggers from a quantised alphabet is Sigma_2^P-complete, one level above robustness certification, so it does not reduce to polynomially many robustness queries unless the hierarchy collapses. Inversion resistance is co-NP-complete for every l_p metric, p a fixed positive integer. Quantifying over parameters instead of inputs - the fault model of bit-flip attacks, radiation upsets and analog accelerators - makes verification exists-R-complete already for networks of identity nodes, for which every previously studied problem is in P, and it stays so when each parameter is confined to a box of inverse-polynomial width; the corresponding safety question is forall-R-complete for ReLU. --- ### How Does Local Landscape Geometry Evolve in Language Model Pre-Training? Authors: Zhanpeng Zhou, Yuhan Sun, Bingrui Li, Jinbo Wang, Huaijin Wu, Lei Wu, Junchi Yan Published: 2026-09-30 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.39767v1 The scale and expense of pre-training language models make efficient hyperparameter tuning essential, yet a principled guidance is still missing. In this work, we analyze language model pre-training dynamics from a local landscape geometry perspective. Our study reveals two distinct phases. In Phase I, sharpness of the local landscape is initially high, leading to instability and loss plateaus under large learning rates (LRs). The landscape shifts from sharp to flatter regions early in training. This dynamic explains the necessity of LR warmup and further suggests that larger peak LRs require proportionally longer warmup periods. In Phase II, the local landscape is governed by the gradient noise scale. Our theory identifies a depth flatness trade-off: high noise from smaller batches widens the loss basin, whereas reduced noise from larger batches deepens it. This theory motivates a dynamic batch-size (BS) scheduler that begins with a small BS and increases it late in training. Together, we provide a unified view of loss landscape evolution, which translates into actionable tuning strategies for large-scale pre-training. --- ### MemCodex: Self-Programming Hierarchical Memory for Language Agents Authors: Xiaoqiang Wang, Bang Liu Published: 2026-09-30 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.39765v1 Agent memory faces heterogeneous access needs: a single-hop question may require one piece of evidence, whereas a multi-hop question must combine evidence from multiple sources. Predefined memory workflows cannot adapt to these varying needs. Recent adaptive methods search or learn over memory components and their compositions, but the design space itself remains predefined. We introduce MemCodex, a self-evolving hierarchical memory system that organizes experience into executable memory programs for summaries, relational knowledge, reusable skills, and latent memory. Open-ended program evolution searches the open design space of layer programs by rewriting how each layer is constructed, indexed, retrieved, and routed, thereby adapting both within-layer implementations and cross-layer composition. At query time, reads traverse the hierarchy from coarse to fine and stop once sufficient evidence is found, descending to the original history when needed. We further develop MemArena, a unified runtime that places heterogeneous data and memory systems behind a common interface. MemCodex improves average task success by 10.1% relative to the strongest adaptive-memory baseline, while using 3.4x fewer context tokens and achieving 2.1x faster inference. --- ### DiffWAM: A Fast and Efficient Navigation World Action Model Authors: Mo Zhu, Yuze Wu, Xijie Huang, Xiao Cui, Fei Gao, Xin Zhou Published: 2026-09-30 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2609.39763v1 Pretrained video foundation models encode rich semantic and spatiotemporal priors for embodied navigation, yet converting these priors into UAV motion typically requires expensive future-video synthesis and geometric reconstruction. We investigate whether the motion implicit in future visual prediction can instead be recovered directly from the predictive representations of a frozen video model. To this end, we present DiffWAM, a geometry-conditioned navigation world-action model that directly transforms multi-level predictive features into continuous camera trajectories. Its Grid-Motion module preserves spatial-temporal motion associations, while Latent2Pose grounds them with first-frame geometry to recover metrically meaningful 3D motion. Complete video rollouts and geometric reconstruction are required only for offline supervision, eliminating future-video decoding and multi-frame reconstruction during deployment. We further introduce FastDreamer, which overlaps predictive and geometric computation with ongoing flight and performs timestamp-aware asynchronous trajectory handoff for continuous UAV execution. DiffWAM achieves a trajectory RMSE of 0.3492 m and an endpoint success rate of 74.40% on the 1,000-sample DiffWAM-1000 benchmark, while representative real-world experiments demonstrate complex behaviors including constrained traversal, orbiting, S-shaped flight, and multi-stage navigation. An onboard DiffWAM-Flash implementation further reaches 1.08 s model-pipeline latency on NVIDIA Jetson AGX Thor. These results demonstrate that predictive video representations can be efficiently grounded into continuous 3D motion, providing a direct alternative to generate-then-reconstruct navigation pipelines. Project page: https://zzmmzzm.github.io/diffwam.github.io/. --- ### Revisiting On-policy Adversarial Black-Box Distillation: Calibrating Groupwise Reward Geometry for Effective Advantage Construction Authors: Xiao Cui, Mo Zhu, Yulei Qin, Yuze Wu, Wengang Zhou, Houqiang Li Published: 2026-09-30 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2609.39757v1 Black-box distillation is a practical route for transferring capabilities from API-accessible large language models that expose only text outputs into smaller student models. Recent on-policy adversarial methods such as GAD improve over SeqKD by forming an adversarial loop between a critic and a student, where the critic provides rewards for GRPO-based student policy optimization over the student's sampled responses. However, GRPO computes advantages from the within-group relative rewards of student samples for the same prompt, whereas the critic is trained primarily to distinguish teacher responses from student responses. This objective mismatch can produce reward groups with collapsed scale or fragile margins, leading to brittle grouped optimization signals. We propose Groupwise Reward Geometry Conditioning (GRGC), a two-stage framework that improves advantage construction by shaping student-side reward groups during both critic training and policy optimization. To improve critic-side conditioning, Gaussian groupwise Optimal Transport calibration regularizes the critic during training to produce reward groups with non-collapsed spread and smooth rank-wise gaps by matching sorted prompt-wise rewards to group-centered Gaussian quantiles. Building on this conditioned reward geometry, policy-side group power modulation reshapes the prompt-wise reward groups before they are converted into advantages, preserving the critic-induced ordering while increasing optimization-relevant margin separability. Extensive experiments across diverse teachers, student model families and scales, and training datasets demonstrate the effectiveness of GRGC on both in-distribution and out-of-distribution evaluations, while introducing negligible overhead over GAD. The code is available at https://github.com/2018cx/GRGC. --- ### Determining Vertical Displacement of Agricultural Areas Using UAV-Photogrammetry and a Heteroscedastic Deep Learning Model Authors: Wojciech Gruszczyński, Edyta Puniach, Paweł Ćwiąkała, Wojciech Matwij Published: 2026-09-30 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.39756v1 This article introduces an algorithm that uses a U-Net architecture to determine vertical ground surface displacements from unmanned aerial vehicle (UAV)-photogrammetry point clouds, offering an alternative to traditional ground filtering methods. Unlike con-ventional ground filters that rely on point cloud classification, the proposed approach em-ploys heteroscedastic regression. The U-Net model predicts the conditional expected val-ues of the elevation corrections, aiming to reduce the impact of vegetation on determined ground surface elevations. Concurrently, it estimates the logarithm of the elevation cor-rection variance, allowing for direct quantification of the uncertainty associated with each elevation correction value. The algorithm was evaluated using three metrics: the root mean square error (RMSE) of vertical displacements, the percentage of nodes with deter-mined displacement values, and the percentage of outliers among those values. Perfor-mance was assessed using the technique for order of preference by similarity to ideal so-lution (TOPSIS) method and compared against several ground-filter-based algorithms across four datasets, each including at least two time intervals. In most cases, the U-Net-based approach demonstrated a slight performance advantage over traditional ground filtering techniques. For example, for the U-Net-based algorithm, for one of the test da-tasets, the RMSE of the determined subsidences was 6.1 cm, the percentage of nodes with determined subsidences was 80.5%, and the percentage of outliers was 0.2%. For the same case, the algorithm based on the next best model (SMRF) allowed an RMSE of 7.7 cm to be obtained; for 77.3% of nodes, the subsidences were determined; and the percentage of outliers was 0.3%. --- ### Validity-Preserving Hierarchical RL for Joint Routing and Switch Placement in EDA Authors: Dorian Gailhard, Ugo Lecerf, Enzo Tartaglione, Donatello Conte, Jhony H. Giraldo Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39749v1 Routing and switch placement are fundamental combinatorial optimization problems in chip design, requiring the joint optimization of routing topology and physical placement under strict structural, geometric and logical constraints. Existing approaches typically rely on carefully engineered heuristics that incorporate strong problem-specific biases to navigate the enormous space of possible designs. In this work, we introduce a hierarchical reinforcement learning framework for joint routing and switch placement at the level of logical communication routes. Starting from a minimal routing graph, our method progressively constructs increasingly expressive solutions through three coupled operations: switch expansion, switch placement, and route refinement. These operations preserve routing validity by construction, restricting exploration to feasible configurations where every communicating initiator-target pair has one assigned loop-free route. We explore the induced solution space using Gumbel Monte Carlo Tree Search, showing that neural-guided search substantially improves solution quality over non-learning optimization methods. Furthermore, pretraining across floorplans provides a strong initialization for fine-tuning on unseen instances. --- ### FAST: Flow Any Scene Transformer Authors: Yongjian Zhang, Longguang Wang, Zhuo Song, Zhiheng Fu, Liang Lin, Yulan Guo Published: 2026-09-30 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.39748v1 Scaling has become a primary driver of progress in language and vision foundation models, yet its role in precise correspondence matching remains underexplored. In this work, we present Flow Any Scene Transformer (FAST), a scalable correspondence model driven by two key insights. First, we reveal that the query-key projections inside single-view vision foundation models encode a coarse yet reusable prior for cross-view matching. Second, reusing these pretrained projections in cross-attention form yields a highly effective initialization for a ViT-based matcher built from a single-view encoder. Guided by these insights, we build FAST upon a vanilla single-view foundation model, utilizing a zero-parameter rewiring strategy to convert selected self-attention layers into cross-attention for cross-view interaction. This design allows ViT-based matchers to scale with advances in single-view foundation models, bypassing the need for a dedicated pair-centric pretraining stage. To fully unlock the scaling potential of this formulation, we assemble a 6-million-pair training corpus for general-purpose dense 2D displacement estimation across diverse co-visible image pairs. Extensive experiments demonstrate that FAST achieves state-of-the-art performance across a wide range of benchmarks, while scaling favorably with both backbone size and training data. --- ### The Nixtlaverse: An Open-Source Ecosystem for Forecasting Authors: Olivier Sprangers, Max Mergenthaler Canseco, Marco Peixeiro, Saul Caballero Ramirez, Mariana Menchero García, Jing-Qiang Goh, Han Wang, Nikhil Gupta, Rogelio Melo, Senbong Gee, Cristian Challu Published: 2026-09-30 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.39741v1 Large forecasting applications often combine statistical, machine-learning, and neural models. These families solve the same problem but differ in fitted state, training procedures, and how they parallelize work. Forecasting software must therefore either hide these differences behind a single estimator interface, or keep the families in separate packages, forcing users to rewrite data preparation and evaluation for every package. We present the Nixtlaverse, an ecosystem of open-source Python libraries for time series forecasting, as a case study of a third design: all libraries share the same long-format panel data and keyed forecast outputs, while every model family keeps its own specialized implementation. We demonstrate this design through three use cases on the public M5 competition data. First, we evaluate statistical, machine-learning, and neural models, and an external engine from a separate ecosystem, in a single rolling-origin evaluation with per-series and hierarchy-weighted metrics. Second, we profile runtime and peak memory from 100 to 30,490 series and locate each family's bottleneck: statistical fitting scales approximately linearly in the number of series, feature construction dominates machine-learning memory, and neural training time is nearly independent of panel size under a fixed training budget. Third, we reconcile the forecasts of multiple engines, including the external one, over all 42,840 series of the M5 hierarchy, with sparse reconciliation where dense implementations exhausted memory. These use cases establish the costs, boundaries, and utility of shared data and output contracts. The Nixtlaverse has seen substantial public distribution, scholarly reuse, and adoption through other forecasting frameworks, and is released under permissive open-source licenses with public datasets, reproducible examples, and verifiable benchmark artifacts. ---