# Research Papers — 2026-09 · bundle page 36 of 266 > Full-text markdown of papers #1751–1800 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/36 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/35.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/37.md) ### EvolvingAvatar: Interactive 3D Head Generation That Adapts as Conversations Unfold Authors: Junjie Chen, Fei Wang, Kun Li, Yiqi Nie, Xun Yang, Yanbin Hao, Linfeng Zhang, Meng Wang Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35616v1 Interactive 3D head generation requires coordinated speaking and listening motion that responds to an evolving conversation. Existing generators use incoming observations as context but keep their parameters fixed, leaving conversational patterns unused as a learning signal. We introduce EvolvingAvatar, a causal generator that uses test-time training to adapt to user face video and dyadic audio during interaction. Its dyadic context prediction objective provides a self-supervised learning signal from audiovisual context without target motion labels at test time. Persistent fast weights accumulate these updates within each conversation to guide motion generation, while transient jaw adaptation responds to current audiovisual context. Predicted speech activity controls how persistent adaptation guides motion. We also introduce InterHead-Bench, a unified 455.95-hour benchmark built from single-view and dual-view conversation videos. Experiments show improved conversational motion statistics over strong baselines. On the hardest out-of-distribution split, generation improves as conversations unfold, reducing mismatch with recorded user-avatar expression statistics by up to 11.1% from the first interval. --- ### Behavioral Foundation Models for Quality Diversity Authors: Nazim Bendib, Nicolas Perrin-Gilbert, Olivier Sigaud Published: 2026-09-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.35615v1 Behavioral Foundation Models (BFMs) are an emerging paradigm in reinforcement learning, playing a role analogous to large language models in natural language processing: they have shown remarkable versatility, enabling zero-shot performance, fast imitation, and online adaptation, all by exploiting the structure of a latent space. In this work, we investigate whether the latent behavioral space induced by BFMs can serve as an effective search space to discover large repertoires of behaviorally diverse and high-performing policies through Quality-Diversity (QD) methods. While QD methods generally search directly in high-dimensional policy parameter space, in this paper, we present BFM-QD, a framework that performs QD search in the compact latent space of a BFM. We further show that the BFM-QD framework provides a closed-form, gradient-free policy improvement operator that approximates a policy gradient update, but requires no critic training and no backpropagation. Across continuous-control benchmarks spanning dense locomotion, sparse navigation, and contact-rich manipulation, BFM-QD consistently outperforms parameter-space baselines, with particularly stark gains in sparse and deceptive settings, where all tested parameter-space QD methods collapse to near-zero performance. These results show the effectiveness of the BFM-QD framework, benefiting from the synergy between dimensionality reduction of the search space and offline pretraining from diverse behavioral data. This positions BFMs as a general-purpose backbone for QD optimization, extending their utility beyond zero-shot task solving to the discovery of diverse behavioral repertoires. --- ### EvE: An Alternate Optimizer to Adam Authors: Shashank Raj, Kalyanmoy Deb Published: 2026-09-28 Categories: cs.LG, cs.NE arXiv: https://arxiv.org/abs/2609.35614v1 Adam and its variants dominate neural network training, but a single run only reveals whether a configuration works well after most of its budget is spent, a poor fit for hyperparameter or architecture search, where configurations must be ranked cheaply and pruned early. We introduce EvE (Evolutionary Explorer), a steady-state, population-of-four differential evolution (DE) optimizer with a targeted Adam fallback: each iteration proposes one candidate via DE, running a short burst of gradient descent only if the DE step fails to improve on the incumbent. Selection is greedy, so on a deterministic objective the best-so-far value is provably monotone non-increasing, and since gradients are used only as a targeted rescue, per-iteration cost stays within a constant factor of a single Adam step regardless of dimension. Under a fixed, evaluation-cost-matched budget, EvE wins or ties Adam on 76% of 70 (problem, dimension) cells across seven scalable benchmarks up to one million variables. On three real neural-network tasks (an MLP on MNIST, and LoRA fine-tuning of a 1.5B-parameter language model on two datasets) EvE finishes the same charged budget 1.7-3.9x faster, at a modest cost in final quality (about one accuracy point on MNIST, 9-11% higher relative test loss on the two fine-tuning tasks; on GSM8K, Adam is about 5 accuracy points more accurate, and fine-tuning lowers accuracy below the base model for both). Inside successive halving on UCI Adult, EvE completes hyperparameter and architecture searches 3.1-3.5x faster, ranking configurations about as consistently with Adam as Adam does with itself across seeds (Kendall's tau 0.66-0.69). EvE is not a total replacement for Adam as a final-stage trainer, but a fast, gradient-aware proxy for the search-heavy, budget-constrained regime one level up. --- ### Remote Sensing Sparse-View 3D Gaussian Splatting via Depth Image-Based Rendering Authors: Jiaming Kang, Zhengxia Zou, Zhenwei Shi Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35612v1 Remote sensing novel view synthesis under sparse observations remains challenging due to insufficient geometric constraints and limited cross-view supervision. Existing Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) methods are prone to overfitting and face challenges of depth ambiguities, missing cross-view information, and insufficient constraints in under-observed regions. To address these challenges, we propose DIBR-GS, a neural Gaussian Splatting framework that exploits Depth Image-Based Rendering (DIBR) to generate pseudo views for cross-view consistency supervision. Specifically, reliable geometric initialization is constructed by aligning monocular depth priors with sparse SfM reconstruction, and cross-view appearance priors are incorporated into neural Gaussian representations to enhance appearance modeling under sparse observations. Furthermore, we introduce a progressive DIBR-based pseudo-view supervision strategy to provide additional geometric and appearance constraints, enabling more complete reconstruction of weakly observed regions. In addition, a height-constrained anchor growth strategy is designed to suppress unreasonable Gaussian expansion. Experiments demonstrate that the proposed method achieves superior performance over existing approaches when training with only 3 input views. Compared with the previous best-performing method, it improves PSNR by 6.83 dB, with relative gains of 14\% in SSIM and 60\% in LPIPS, while maintaining competitive computational efficiency. Our code is available at https://github.com/kanehub/DIBR-GS --- ### On-Policy Self-Distillation for Multi-Turn Image Editing Authors: Liangbing Zhao, Le Zhuo, Mohamed Elhoseiny Published: 2026-09-28 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.35611v1 Instruction-based image editing has achieved strong performance in single-turn settings, yet practical editing is often iterative, with each instruction applied to the output of the previous turn. We find that existing editing models degrade rapidly under recursive editing and attribute this failure to a train-test mismatch in the conditioning distribution: models are trained on clean source images but must repeatedly condition on their own imperfect outputs at inference time. To address this, we propose MT-OPSD, an on-policy self-distillation framework that trains the model on self-generated conditioning states with editing supervision from a clean-conditioned teacher, without requiring multi-turn annotations. We further introduce LME-Bench, a benchmark of 100 ten-turn editing sessions for evaluating long-horizon robustness. Experiments across three editing backbones show that MT-OPSD substantially improves long-horizon editing success and reduces multi-turn collapse while largely preserving single-turn editing quality. --- ### Twist, Don't Tilt: Trajectory-Exact Constrained Decoding for Masked Diffusion Models Authors: Aditya Thimmaiah, Lara Marinov, Jayanth Srinivasa, Haris Vikalo, Junyi Jessy Li, Milos Gligoric Published: 2026-09-28 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.35609v1 Constrained decoding for Masked Diffusion Language Models (MDLMs) aims to ensure that generated outputs satisfy a specified structure or syntax constraint. MDLMs generate outputs by repeatedly unmasking masked positions present in their current state. Recent strategies for constrained decoding constrain the model's per-step mean-field posterior (which factorizes over masked positions) by enforcing the desired constraint with an automaton. The resulting chain-structured factor graph allows exact constrained sampling via dynamic programming. However, despite each draw being exact and constraint-satisfying, we prove that their composition, in general, tilts away from the model's relative probabilities over valid trajectories, thus leading to trajectory bias. We derive an exact expression for this bias as a product of ratios measuring how valid continuation mass changes when the denoiser is reconditioned, and characterize when the bias vanishes. We then correct the bias by introducing TWISTER, the first automaton-twisted Sequential Monte Carlo decoder for MDLMs, using the step-exact decoder as the proposal. We show that for regular language constraints, the Feynman-Kac correction is exactly computable, with the twists obtained efficiently using quantities pre-computed for step-exact sampling. We prove that the resulting Feynman-Kac model targets the unbiased Doob h-transformed path law conditioned on constraint satisfaction. --- ### Simultaneous Translation between Sign Languages Authors: Zetian Wu, Bowen Xie, Stefan Lee, Liang Huang Published: 2026-09-28 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2609.35608v1 Deaf and hard-of-hearing (DHH) signers cannot converse in real time across different sign languages today: existing sign-to-sign translation systems run offline, requiring the full source clip before any target sign is emitted. Live use cases - e.g. broadcast interpretation and two-way video calls - instead demand simultaneous output, while the source signer is still signing. We present, to our knowledge, the first simultaneous sign-to-sign (S2S) translation system, with two wait-k regimes: test-time wait-k inference applied directly to a full-sentence model, and a trained wait-k model via stochastic multi-path supervision. We further introduce ca-Stream-AL, a computation-aware latency metric for streaming output. Averaged across six S2S directions on both a smaller human-verified test set and a larger synthetic S2S corpus, our streaming system achieves a 38% ca-Stream-AL reduction while staying within a 9% DTW-PA-MPJPE increase and a 2.1 BLEU-4 drop compared to the full-sentence baseline. A word-order case study probes how the streaming model handles word order mismatch between different sign languages - a consequence of simultaneous translation. --- ### TCSAlgBench: Benchmarking Automated Proving for Research-Level Theoretical Computer Science Authors: Chutong Yang, Xiyuan Zhang, Yu Huang, Boran Han, Soonho Kong, Shuai Zhang, Vihang Prakash Patil, Zhen Han, Michael Bohlke-Schneider, Bernie Wang Published: 2026-09-28 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.35606v1 Large language models perform strongly on competition mathematics, but their research-level reasoning remains difficult to evaluate systematically. Theoretical computer science (TCS) connects algorithm design to explicit guarantees and fundamental limits, providing a setting for evaluating whether models can justify computational improvements with arguments humans can inspect. We introduce TCSAlgBench, a benchmark and reusable pipeline for natural-language proof discovery, comprising 398 theorem-level challenges from 138 STOC and COLT 2026 papers. Expert-designed rules complete paper-specific context, preserve computational assumptions and quantitative guarantees, and withhold constructions when discovering an algorithm is part of the task. For each task, prover systems receive theorem statements and access to cited prior work. The pipeline supports fresh, versioned challenge batches from newly released papers. We evaluate ten model configurations from four families under direct inference and prover-verifier discussion, and compare four agent workflows under matched model-call opportunities. All evaluations use the full benchmark. In the model comparison, GPT-5.6 Sol max achieves the highest five-run verifier-accepted coverage at 23.6% after 10-round discussion. Discussion and repeated sampling improve coverage. In the separate agent comparison using GPT-5.5 xhigh, decomposition improves coverage over discussion, and agentic planning achieves the highest five-run verifier-accepted coverage at 25.4%. TCSAlgBench provides a refreshable testbed for measuring progress in model reasoning and studying how agent workflows support research-level proof discovery. --- ### Control-Geometry Straightening for Sampling-Based Latent Planning Authors: Ziang Fu, Ning Ning Published: 2026-09-28 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.35603v1 Joint-embedding predictive architectures enable planning with latent world models, but accurate transition prediction alone does not ensure that the planning objective is easy to optimize. We introduce Control-Geometry Straightening (CGS), a single auxiliary loss that learns planner-friendly representations by directly straightening control geometry for sampling-efficient planning. CGS matches pairwise cosine similarities among actions to those among corresponding latent differences only using local transitions from pixel-action pairs. The loss can be applied across world-model architectures using end-to-end learned or pretrained representations. Under linear-dynamics, our theoretical analysis connects this objective to temporal straightening and more balanced terminal-cost curvature across the full planning horizon, yielding finite-budget guarantees for MPPI, local contraction results for CEM, and convergence bounds for gradient descent. Across four control environments and multiple planners, CGS improves planning with fewer sampled candidates and refinement steps, achieving success-rate gains up to 20 and 12.6 percentage points over LeWorldModel (LeWM) and its temporal-straightening variant (LeWM+TS), respectively, with sampling-based planners using 128 candidates per update. Probes, comparisons with DINO-WM architecture, and planner-side ablations clarify how latent motion organization, state dependence, and dynamical context shape planning behavior. Straightening control geometry thus makes good action sequences easier to find under limited planning budgets. --- ### Signatures of semantic search in the activations of large language models Authors: Luke Leckie, Peter M. Todd, Jacob G. Foster Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35599v2 When recalling lists of concepts (e.g., animals) during the semantic fluency task (SFT), both humans and large language models (LLMs) organise their output into clusters of related items (e.g., sea animals) that are punctuated by strategic switches between clusters. In humans, this pattern can be explained by a semantic foraging process, whereby distinct neural and behavioural signatures accompany within-cluster production ("exploit") and between-cluster switching ("explore"). Whether LLMs likewise represent these two search regimes within their internal states is unknown. Here, we apply a range of mechanistic interpretability techniques to provide evidence for this. In Study 1, we use the Jacobian lens (J-lens), which maps intermediate-layer residual-stream representations to token-level activations, to show that concept-level activations predict switching. First, we find that switching coincides with low next-token activations. Moreover, the probability of switching rises as the set of strongest J-lens activations (the J-space) becomes depleted of items from the category currently being produced, analogous to explore-exploit decision-making during patch foraging. We then show that middle-layer J-lens activations of abstract category-related labels (e.g., "water") increase in anticipation of switching into that category. We confirm these representations to causally influence switching by deriving steering vectors that target category switching. In Study 2, we identify generic residual stream directions that are activated during and in anticipation of switching. By steering activations along these directions, we bias increased or decreased rates of switching. Our study extends the semantic foraging framework to artificial intelligences and provides evidence that LLMs maintain distinct representational signatures for exploration and exploitation as they verbalise conceptual information. --- ### Signatures of semantic search in the activations of large language models Authors: Luke Leckie, Peter M. Todd, Jacob G. Foster Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35599v1 When recalling lists of concepts (e.g., animals) during the semantic fluency task (SFT), both humans and large language models (LLMs) organise their output into clusters of related items (e.g., sea animals) that are punctuated by strategic switches between clusters. In humans, this pattern can be explained by a semantic foraging process, whereby distinct neural and behavioural signatures accompany within-cluster production ("exploit") and between-cluster switching ("explore"). Whether LLMs likewise represent these two search regimes within their internal states is unknown. Here, we apply a range of mechanistic interpretability techniques to provide evidence for this. In Study 1, we use the Jacobian lens (J-lens), which maps intermediate-layer residual-stream representations to token-level activations, to show that concept-level activations predict switching. First, we find that switching coincides with low next-token activations. Moreover, the probability of switching rises as the set of strongest J-lens activations (the J-space) becomes depleted of items from the category currently being produced, analogous to explore-exploit decision-making during patch foraging. We then show that middle-layer J-lens activations of abstract category-related labels (e.g., "water") increase in anticipation of switching into that category. We confirm these representations to causally influence switching by deriving steering vectors that target category switching. In Study 2, we identify generic residual stream directions that are activated during and in anticipation of switching. By steering activations along these directions, we bias increased or decreased rates of switching. Our study extends the semantic foraging framework to artificial intelligences and provides evidence that LLMs maintain distinct representational signatures for exploration and exploitation as they verbalise conceptual information. --- ### Learning Conditional Expectation Operators via Functional Newton Updates Authors: Thiago Ramos, Alek Fröhlich, Daniel Perazzo, Massimiliano Pontil Published: 2026-09-28 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.35598v1 We introduce the Functional Spectral-Newton Method (FSNM) for learning the leading singular structure of a conditional expectation operator without fixing a basis or reproducing kernel Hilbert space. FSNM fits a low-rank representation of the centered joint-to-product density ratio kernel by alternating functional Newton updates. Each update reduces to a preconditioned regression, which we approximate with vector-valued regression trees in a stagewise boosting procedure. At the population level, we establish descent and an $O(1/T)$ best-iterate block-stationarity rate under a relative weak-learner accuracy condition, and show that every nondegenerate local minimum over the full centered $L^2$ spaces is a globally optimal rank-$d$ approximation. Synthetic experiments show that FSNM recovers a low-rank density ratio and its leading spectral structure, and that the same learned kernel can answer multiple conditional queries without refitting. --- ### SEABench: Benchmarking Endogenous Misalignment In Self-Evolving Agents Authors: Saswat Das, Parvati Viswanathan, Daniel Donnelly, Chang Huang, Sahar Abdelnabi, Ferdinando Fioretto Published: 2026-09-28 Categories: cs.CR, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.35596v1 Self-evolving LLM agents have gained prominence for their ability to improve after deployment by modifying their harness, including their controller instructions, memory management protocols, and reusable tools and skills, in response to user and environment feedback. However, locally useful updates may persist into later tasks where they produce unsafe behavior, even without direct adversarial influence. To study this risk, we introduce SEABench, a benchmark for studying endogenous misalignment arising from agent self-evolution, with 48 longitudinal task sequences that span multiple evolution surfaces, task domains, and harm types in a rich personal-assistant environment. To account for the stochasticity inherent in agentic operations, we provide an adaptive trajectory discovery pipeline that probes for failures while preserving original task intent and supports causal attribution through paired non-evolving agents and attribution scores. Our evaluation across multiple recent LLMs, evolution surfaces, and harm types reveals that self-evolution indeed increases task completion rates but often at the cost of safety failures that are absent for paired non-evolving baseline agents. We also show that qualitatively different safety behaviors emerge across evolution surfaces and harm types. Further, we show that this divergence in safety behavior is reflected in agents' chain-of-thought reasoning, which yields an effective monitoring strategy that can mitigate unsafe behavior with a low false positive rate. --- ### ReSS: Residual-Restoring Sparse Attention for 3D Vision Transformers Authors: Yongsung Kim, Jaehoon Lee, Minjun Park, Wooseok Song, Hun Hwangbo, Sungroh Yoon Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35593v1 3D vision transformers such as VGGT predict camera poses and scene geometry from multi-view images in a single forward pass, but their global attention over all concatenated view tokens dominates computation as the number of views grows. To reduce this cost, SparseVGGT and HeSS sparsify attention at the block level, and both retain blocks with high attention probability. However, we observe that attention probability poorly predicts how much the model's behavior actually changes when a block is removed, and we show that this mismatch is why performance collapses as sparsity increases. In this paper, we propose ReSS (ReSidual-ReStoring Sparse Attention), which recasts block selection from a problem of maximizing the retained attention mass to one of minimizing the drift that sparsification leaves in the residual stream. We introduce a drift score that quantifies how much each block shifts the residual, and, since the drift of a drop set depends on the directions of the contribution vectors rather than on their magnitudes alone, an iterative residual restoration procedure that refines the drop set as a whole. Across three backbones and five datasets, ReSS preserves dense performance better than prior methods at matched sparsity. Two further results support drift as the quantity that governs the cost of sparsification: maximizing drift degrades performance faster than random selection, and plotted against realized drift instead of sparsity, all methods fall approximately onto a single curve. Code is available at https://github.com/libary753/ReSS. --- ### Language Models Act on Hidden Valence Authors: Cameron Berg, Caspar Kaiser Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.35591v1 Language models describe some internal states as good and others as bad. But whether models have a stake in them is an open question. Simply asking the model is unlikely to be informative. Any answer may be consistent with genuine introspection, superficial pattern-matching, or with fixed scripts learned in character training. We therefore study revealed preference. Rather than asking about a state, we use activation steering to attach a positively or negatively valenced activation pattern to one of two otherwise meaningless 'zones', switch steering off, and then observe which zone the model prefers. A model with a stake in that state should choose accordingly. Across seven open-weight models from five families, this is indeed what we find. First, steering changes the passages models write about each zone, and those words shift later choice. Second, the shift persists when all surface-level tokens are held fixed and only the hidden KV cache differs. Third, the effect also remains when all text is generated without steering and valence is only injected during cache construction. Thus, the hidden state alone moves choice in proportion to the steering dose. Fourth, this dependence of choice on hidden valence is nearly absent in a base model and emerges during DPO, consistent with a link between valence and goal-directed behaviour formed in training. Finally, given tools to steer itself, a model does not tend to induce a positive state, but it reliably removes an imposed negative state. It does so at a dose-dependent rate and significantly more often than it removes interventions in random directions. Overall, we demonstrate that valence-related activation patterns leave hidden traces that predictably govern later choices, even when every visible token is identical across conditions. Whether these traces are accompanied by any subjective experience relevant to model welfare remains unclear. --- ### Source-preserving alignment for robust evidence localization in scientific PDFS Authors: Zihao Liu, Wei Yang, Zixiao Dong, Chenshu Li, Longzhang Liu, Tao Tan, Hong Xie Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35588v1 Scientific information-extraction systems often return a claim with an evidence string, which users must locate in the original PDF. This is challenging because the extracted evidence and PDF text layer are different representations: line wrapping, Unicode variants, superscripts, citation markers, and fragmented items alter text sequences and geometry. We present a source-preserving alignment framework: normalize text for robust matching while preserving provenance for accurate localization. It aligns evidence with normalized page text, maps matches back to source-character spans, and renders only their geometry. When exact alignment fails, line-break-aware token alignment recovers supported spans while excluding unmatched noise. Experiments on 1,020 chemistry papers show that the framework achieves a 92.6\% quote-level automatic localization rate, compared with 43.6\% for text search and 19.1\% for a precomputed bounding-box baseline. Component ablation confirms distinct contributions from normalization and approximate token alignment, while human verification assesses the visual correctness of returned highlights. Overall, these results demonstrate that reliable evidence verification requires robust matching and precise localization within a shared source-preserving alignment representation. --- ### Hardware-Aware Features for CUTLASS Kernel Selection Authors: Shriram Chandran, Dominic Rinderer, Yakup Budanaz, Alexandru Calotoiu, Marcin Copik, Torsten Hoefler Published: 2026-09-28 Categories: cs.LG, cs.PF arXiv: https://arxiv.org/abs/2609.35587v1 GPU libraries such as CUTLASS expose tens of thousands of semantically equivalent kernels for a single operation, making exhaustive autotuning expensive and execution-free selection difficult. Existing analytical selectors require hand-designed performance rules, while learned selectors operate on raw configuration parameters and must infer hardware consequences from data. We introduce a hardware-aware representation for CUTLASS kernel selection that augments candidate configurations with statically computable estimates of induced hardware behavior. We construct a dataset of 4.9 million CUTLASS kernels and train gradient-boosted and neural learning-to-rank models to rank candidates within each problem. On held-out exhaustive evaluation problems, hardware-aware representations reduce selection regret by up to 40\% relative to structural baselines and 64.2\% relative to NVIDIA's matrix-multiply heuristics. We further evaluate data-efficient cross-precision and epilogue-fusion transfer within CUTLASS GEMM, showing that explicitly representing candidate-induced hardware behavior provides a useful inductive bias for learned kernel selection. --- ### IMC-CLINIC: Coupled Loss-Informed Newton Iterations for Clipping in Analog In-Memory Computing Authors: Yung-Chin Chen, Chia-Yu Chen, Naveen Verma Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35586v1 Analog in-memory computing (IMC) offers a promising path toward energy-efficient large language model (LLM) inference by executing matrix multiplications (MatMul) directly within memory arrays in the analog domain. Its efficiency, however, comes with an additional source of error: limited-precision analog-to-digital converters (ADCs) quantize accumulated analog partial sums, introducing output-side error distinct from conventional activation and weight quantization at the MatMul inputs. Clipping can mitigate both operand and ADC quantization errors, but the optimal clipping factors must jointly balance activation rounding and clipping, weight rounding and clipping, and ADC quantization. Existing clipping methods, designed for digital quantization, do not explicitly optimize these coupled sources of IMC error and often rely on costly search-based calibration. We introduce IMC-CLINIC (Coupled Loss-Informed Newton Iterations for Clipping), a clipping calibration framework based on an analytical surrogate for IMC MatMul output error. The surrogate jointly models operand quantization, accumulated clipping-induced bias, and ADC quantization, enabling efficient evaluation of its gradient and approximate curvature from a small calibration set. IMC-CLINIC jointly optimizes activation and weight clipping factors using a safeguarded Newton-type method. Across multiple models and datasets, it improves average zero-shot accuracy by 6.5-11.5 percentage points over the grid search baseline while reducing calibration time by factors of 10.0-12.1. Its analytical surrogate closely tracks empirical IMC output error, and its optimizer is certified within 1% of the global optimum under the loss objective across all projections on two representative models. --- ### What Paired Evaluations Reveal under Visual Perturbations Authors: Yongda Wei, Chen Zhang, Yifei Wang, Xinyu Wang, Bosen Shao, Hanxi Li, Liping Di Published: 2026-09-28 Categories: cs.CV, stat.AP arXiv: https://arxiv.org/abs/2609.35583v1 Robustness evaluation must examine diverse visual perturbations, while benchmarks cover only some real-world conditions and physical testing is costly. Paired evaluations link clean and perturbed predictions for the same image, capturing changes in correctness, confidence, and acceptance beyond aggregate accuracy. We investigate how this image correspondence supports two needs in robustness evaluation: interpreting paired evaluation results and prioritizing samples for physical testing. To interpret paired evaluation results, we fix both sets of prediction records and vary their correspondence within each class. We prove that classwise correct-correct counts give the same sharp bounds on lost acceptance and mean true-class probability decrease among retained-correct inputs as any feasible five-state refinement. Distinguishing persistent from changed wrong answers can further constrain accepted-error transitions, while shared correspondence can establish policy orderings left unresolved by separate cost intervals. To prioritize samples for physical testing, we retain each image's synthetic responses and rank clean-correct images by their mean true-class probability under corruption. Across 44 classifiers, testing the highest-risk 20% finds 67% and 45% of failures under mild screen and print recaptures, versus 58% and 36% for clean confidence and 60% and 37% for an equal-size natural-transformation average. With both probability averaging and an A3Rank scoring adaptation, the tested corruption set yields higher mean failure recall than the natural-transform set; differences between scores depend on the source and budget. Together, these findings show that the value of correspondence depends on the evaluation objective: classwise counts suffice for specified reliability bounds, while image-specific synthetic responses improve the allocation of physical tests within the evaluated pool. --- ### QC-Stark: A Multi-Task Benchmark Revealing Capability Dissociations in LLMs Evaluated on Quantum Computing Tasks Authors: Pranav Gupta Published: 2026-09-28 Categories: quant-ph, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.35581v1 We introduce QC-Stark, a benchmark for evaluating large language models (LLMs) on 11 quantum computing (QC) tasks, spanning circuit construction, debugging, compilation, error correction, and simulation. Across 2,750 evaluations (10 models $\times$ 11 tasks x 5 difficulty levels x 5 seeds), we find that overall rankings mask substantial per-task variation. The Spearman correlation between overall and per-task rankings is statistically insignificant for 4 out of the 11 tasks included in this benchmark. A 2-parameter Item Response Theory (IRT) model validates measurement quality, and prompt sensitivity analysis confirms ranking robustness across prompt conditions. All tasks are auto-verifiable via execution, thus not requiring any manual evaluation. We make the code and data publicly available on Huggingface. --- ### Output-aware Residual Stream Pruning for Large Language Models Authors: Chayne Thrash, Kevin Chen, Soheil Kolouri Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35579v1 Residual stream pruning methods reduce inference cost by shrinking the model's hidden dimension, but existing approaches typically choose these dimensions by minimizing activation reconstruction error. This criterion implicitly treats all perturbation directions as equally important, ignoring the sensitivity of downstream layers. We introduce a sensitivity-aware approach to residual-stream pruning that directly accounts for this direction-dependent sensitivity. Using a second-order approximation to the output KL divergence, we characterize the effect of a residual-stream perturbation through both its activation covariance and the local sensitivity of the model output. The resulting subspace selection objective couples these two quantities, but is difficult to optimize directly. We derive a tractable spectral upper bound that reduces subspace selection to an eigendecomposition of a sensitivity-weighted covariance matrix, retaining the efficiency and structural simplicity of rotation-based pruning methods. Across several instruction-tuned language model families, our method consistently reduces calibration KL divergence relative to activation-only pruning and improves perplexity and downstream task performance over a range of compression levels. Our results show that preserving activation energy alone is insufficient for residual-stream pruning, and that explicitly accounting for how perturbations propagate to the model output provides a more effective criterion for selecting dimensions to remove. --- ### FactorEngram: Factorized N-gram Memory with Basis-Level Gating for Language Models Authors: Bowen Yang, Jingbo Zhou, Qinghong Miao, Hua Wu Published: 2026-09-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.35578v1 Lookup-based memory has been a promising way to scale the parameters of large language models (LLMs). It retrieves learned representations of local token patterns, such as n-grams, instead of reconstructing them through successive layers of computation. However, existing designs such as Engram treat each retrieved embedding as a monolithic unit. Each embedding is stored in its own hashed slot and modulated by a single scalar gate. As a result, polysemous patterns cannot selectively read out the components of their memory that are relevant to the context. Moreover, parameters are shared only through hash collisions, which are largely unrelated to semantics. We propose FactorEngram, a factorized n-gram memory with basis-level contextual gating. FactorEngram retrieves sparsity-regularized coefficients over a dictionary of basis vectors shared across patterns, so related patterns can reuse common components. The same dictionary is also used for gating. The backbone hidden state is scored against each basis vector to gate the corresponding coefficient before reconstruction, which lets the context modulate each memory component individually. FactorEngram also covers both individual tokens and multi-token n-grams, and we systematically study where the memory branch should be inserted. On 340M- and 1B-parameter Transformer backbones, FactorEngram improves language modeling and downstream task performance. Ablation studies confirm the contribution of each component and identify insertion before the attention sublayer in the middle layers as an effective configuration. --- ### Share-Borne AI Virus: Memory-Hopping Attacks Across LLM Agents Authors: Sidharth Pulipaka, Ansh Sharma, Stanislau Hlebik, Leonidas Raghav, Vyas Raina, Ivaxi Sheth, Mario Fritz Published: 2026-09-28 Categories: cs.AI, cs.CL, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.35576v1 Large language models are increasingly deployed as stateful assistants that retain information across interactions and use tools to read, modify, and create persistent artifacts. As these artifacts are shared between users, they form an indirect communication channel between otherwise independent assistants. We study a failure mode in which this channel enables self-propagating attacks. We introduce artifact-mediated propagation, where adversarial content introduced through an artifact (e.g. a report), is stored in an assistant's persistent memory, reproduced in a subsequently created artifact, and acquired by another assistant that later reads it. We evaluate this process in temporal human-agent universes that model artifact exchange between independently operated assistants over time, measuring whether an attack survives successive hand-offs, how many hops it reaches, and how broadly it spreads. We find that attacks can propagate across multiple independent assistants and persist over extended interaction sequences. In larger simulated environments, even GPT-5.6 Luna exhibits substantial spread, reaching 60-80% of agents with propagation chains extending to eight hops. These results show that persistent artifacts can act as durable carriers of adversarial state, allowing attacks to outlive individual interactions and spread across isolated assistants. --- ### F4R: Failure-Driven Recognition, Reconstruction, Refinement, and Redeployment for Continual Robot Self-Improvement Authors: Zhuoyuan Yu, Jiacheng Wang, Tianle Liu, Yihua Ren, Peng Yu, Chen Bai, Ziheng Zhang, Yufei Jia, Jindou Jia, Yuhang Zhang, Xinrui Zhang, Shang Yujing et al. Published: 2026-09-28 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2609.35575v2 The real-world performance of current vision-language-action models is fundamentally constrained by the limited coverage of expert demonstrations and their insufficient understanding of physical interactions. A common remedy is to collect additional real-world demonstrations of newly encountered failures. However, this process is costly, inefficient, potentially unsafe, and difficult to scale. To address this challenge, we propose Failure for Rising (F4R), a failure-driven real-to-sim-to-real closed-loop learning framework that converts real-world failures into targeted policy improvement. F4R first uses an agent to automatically identify and diagnose failures from rollouts. It reconstructs each failure as an interactive, object-centric table-top environment that preserves the task-relevant spatial and physical conditions. The policy is then refined through failure-conditioned sim-real co-training followed by targeted reinforcement learning in the reconstructed environments. The improved policy is subsequently redeployed, while newly observed failures are continuously fed back into the next reconstruction and learning cycle. Real-world evaluations on four manipulation tasks show that F4R achieves 93.75% In-Distribution and 90.0% Out-of-Distribution (OOD) success, outperforming the budget-matched Targeted BC baseline by 18.75 percentage points under OOD conditions without collecting additional real-world corrective demonstrations. --- ### F4R: Failure-Driven Recognition, Reconstruction, Refinement, and Redeployment for Continual Robot Self-Improvement Authors: Zhuoyuan Yu, Jiacheng Wang, Tianle Liu, Yihua Ren, Peng Yu, Chen Bai, Ziheng Zhang, Yufei Jia, Jindou Jia, Yuhang Zhang, Xinrui Zhang, Shang Yujing et al. Published: 2026-09-28 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2609.35575v1 The real-world performance of current vision-language-action models is fundamentally constrained by the limited coverage of expert demonstrations and their insufficient understanding of physical interactions. A common remedy is to collect additional real-world demonstrations of newly encountered failures. However, this process is costly, inefficient, potentially unsafe, and difficult to scale. To address this challenge, we propose Failure for Rising (F4R), a failure-driven real-to-sim-to-real closed-loop learning framework that converts real-world failures into targeted policy improvement. F4R first uses an agent to automatically identify and diagnose failures from rollouts. It reconstructs each failure as an interactive, object-centric table-top environment that preserves the task-relevant spatial and physical conditions. The policy is then refined through failure-conditioned sim-real co-training followed by targeted reinforcement learning in the reconstructed environments. The improved policy is subsequently redeployed, while newly observed failures are continuously fed back into the next reconstruction and learning cycle. Real-world evaluations on four manipulation tasks show that F4R achieves 93.75% In-Distribution and 90.0% Out-of-Distribution (OOD) success, outperforming the budget-matched Targeted BC baseline by 18.75 percentage points under OOD conditions without collecting additional real-world corrective demonstrations. --- ### Less Is More: Genetic Frame Selection for Efficient Novel View Synthesis Authors: Diego E. Farchione, Ramzi Idoughi, Alberto Jaspe-Villanueva, Peter Wonka Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35573v1 Feed-forward novel view synthesis reconstructs a scene from many input images in a single forward pass, yet more views do not necessarily improve performance: redundant or poorly chosen frames increase computational cost and may degrade reconstruction quality. We address the problem of selecting, from an already captured sequence, a fixed-size subset of input views that is most informative for reconstructing specified target viewpoints. We propose a render-free view selector that scores candidate frames based on three complementary criteria: target-view coverage, measured against observed frames that stand in for the targets, redundancy with previously selected views, and image sharpness. A lightweight scoring network then selects the most informative frames without rendering, reconstruction, or per-scene optimization at inference time. To train the selector, we distill an expensive offline search procedure in which a genetic algorithm identifies high-quality subsets by directly optimizing reconstruction performance on training scenes. The selector learns to reproduce these choices from geometric and image-level features alone. Across six datasets and multiple input budgets, our method consistently outperforms both geometric and reconstruction-aware view-selection baselines while incurring significantly lower selection costs than reconstruction-based alternatives. Moreover, carefully selected subsets can outperform feed-forward reconstruction from the full input sequence. The learned selector generalizes across diverse reconstruction paradigms (feed-forward, 3D Gaussian Splatting, and NeRF), to object-targeted reconstruction and to a cross-capture setting in which the target views come from a separate acquisition pass. More broadly, our results indicate that explicitly reasoning about target relevance and inter-view redundancy is a fundamental factor in efficient scene reconstruction. --- ### Representation Alignment as a Bottleneck in LLM-Based Retrosynthesis Planning Authors: Hyunwoo Yoo, Cassie Huang, Haebin Shin, Li Zhang, Gail L. Rosen Published: 2026-09-28 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.35571v1 While LLMs show promise in general reasoning, symbolic planning in chemistry remains a bottleneck. Direct ''SMILES-to-PDDL'' attempts fail because they force models to juggle chemical analysis and planning-language structuring simultaneously. We hypothesize that this failure stems from a lack of intermediate abstractions rather than insufficient model capacity. By decomposing retrosynthesis into molecule mapping, reaction mapping, and PDDL generation, we achieve high success rates where end-to-end approaches fail. This provides evidence that a primary bottleneck lies in representation alignment rather than raw model capacity. Our structural analysis demonstrates that intermediate representations are essential in retrosynthesis planning, highlighting the importance of representation-centric design in future systems. --- ### EdgeVLN: Runtime-Aware Deployment Ready Quantized Vision Language Navigation Model Authors: Rithvik Jonna, Man Namgung, Aakash Gurram, Tinoosh Mohsenin Published: 2026-09-28 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2609.35570v1 Vision-language navigation (VLN) models perform well but target compute-rich platforms, limiting deployment on memory- and power-constrained robotic edge devices. Compression alone does not establish whether a VLN model fits the memory, latency, and energy budgets of an edge platform while preserving navigation behavior. We introduce EdgeVLN, a runtime-aware, deployment-ready quantized VLN model that closes this gap. EdgeVLN combines a quantized StreamVLN model with Latent Trajectory Termination Extractor (LATTE), a lightweight causal transformer that improves real-time stopping by predicting a Stop Action verifier rank. Both execute through our llama.cpp VLN driver, which reconstructs streaming context and prunes memory tokens on-board. We characterize a pretrained StreamVLN backbone across weight quantization from 8 to 2 bits and multiple inference runtimes to identify a feasible operating point. LATTE reuses backbone hidden states within the budget freed by quantization, requiring neither a second vision encoder nor an additional backbone forward pass. We evaluate six backbone precisions and seven candidate stop heads on BF16 and IQ4 NL across all 1,839 R2R VLN-CE val-unseen episodes. We measure success rate (SR) in simulation and latency, energy, and resident memory on an NVIDIA Jetson Orin NX 16 GB. LATTE achieves our highest SR, 58.02 percent on the deployed 4-bit model, exceeding the BF16 baseline with only 0.013 s additional latency per navigation step. Four-bit formats achieve nearly identical SR, but step energy varies 36.8 times by execution path. Only IQ4 NL under our VLN driver fits the board, using 11.35 GB resident memory while running 20.8 times faster and using 13.3 times less energy than storage-streamed BF16. INT2 collapses. Runtime selection, memory-token pruning, and quantization are essential for efficient edge deployment. --- ### Beyond Energy: When Sustainability Dimensions Reshape LLM Serving Decisions Authors: Tianyao Shi, Xipeng Shen, Yi Ding Published: 2026-09-28 Categories: cs.CY, cs.DC, cs.LG, cs.PF arXiv: https://arxiv.org/abs/2609.35569v1 Large language model (LLM) serving has environmental impacts across energy consumption, carbon emission, water consumption, and biodiversity loss. Yet these dimensions are largely evaluated in isolation, leaving it unclear when and how they lead to different optimization decisions. We present PRISM, a unified framework for characterizing and optimizing LLM serving across energy, carbon, water, and biodiversity impacts. Our analysis reveals a fundamental distinction: computing configurations determine energy consumption, whereas where and when LLM serving is deployed determine its carbon, water, and biodiversity impacts. Under a fixed deployment choice and operational-only accounting, all dimensions preserve the same energy-based configuration ranking. Deployment rankings can diverge across dimensions, while embodied impacts can break configuration invariance when they exceed a lifecycle crossover boundary. PRISM identifies these conditions, quantifies cross-dimensional regrets, and balances the four dimensions. In regional-routing experiments, PRISM reduces median worst-case regret by 50.2% relative to the strongest baseline. --- ### From Experience to Expertise: Adoption-Aware Memory Learning for Data-Scarce NPU Kernel Synthesis Authors: Longxiao Fan, Tao Zhang, Han Yan, Jiajun Li, Mingcong Song, Guoping Long, Hongjie Si, Weiwei Sun Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35568v1 High-performance kernels underpin efficient accelerator execution but require expert tuning and lengthy manual optimization cycles. LLM coding agents promise automation, yet their CUDA knowledge transfers poorly to data-scarce domain-specific architectures (DSAs) such as NPUs, whose execution models and memory hierarchies differ substantially from those of GPUs. To address this transfer gap, post-training methods adapt LLMs to NPU programming but depend on scarce expert data and substantial training compute. Memory-learning agents instead adapt through external memory, but their uniform credit assignment gives adopted and unused experiences the same reward target, potentially biasing subsequent retrieval rankings. Moreover, when learned values guide only retrieval, high-value experiences that generalize across operators must be retrieved repeatedly rather than retained in context, thereby increasing retrieval overhead and weakening cross-task guidance. We therefore present SAGE, a persistent self-improving agent for NPU kernel synthesis. Adoption-Traced Utility estimation (ATU) combines explicit adoption records with kernel evaluation outcomes for adoption-aware credit assignment. Utility-Gated Consolidation (UGC) uses positive utility and repeated adoption across operators to select and abstract reusable rules into a bounded resident context. On NPUKernelBench, SAGE achieves a 95.5% execution rate versus 84.1% for the strongest controlled baseline, with 86.9% of solved operators outperforming torch_npu. With GLM-5.3, SAGE achieves a 43.99x speedup over the torch_npu reference on sparse flash attention. These results show that adoption-aware credit assignment and selective consolidation enable agents to accumulate and reuse hardware-specific knowledge across tasks. --- ### Almieyar: A Culturally Grounded Benchmark for Multi-Dialect Arabic Speech Recognition Authors: Omid Ghahroodi, Anas Madkoor, Dima Faris Al Saudi, Fagr Tahir, Malak Annan, Talha shahid javad allah rakha, Omar Al-Busaidi, Zineb El Kahla, Iheb Zouari, Essa Ahmed Abou Jabal, Ahmed Ezzat, Hind AL-Merekhi et al. Published: 2026-09-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.35564v1 Arabic speech technology has largely focused on Modern Standard Arabic, leaving the living dialects spoken by hundreds of millions under-served. We introduce ALMIEYAR, a culturally grounded ASR benchmark covering 17 Arabic dialects across six families, built entirely from newly recorded speech unseen by existing models. Dialect-community coordinators selected culturally relevant images across 10 topics, and native speakers described them through five structured scenarios, yielding approximately 50 minutes per dialect (13.7 hours total). We benchmark 12 state-of-the-art ASR systems zero-shot, including GPT-4o-transcribe, Voxtral-Mini-4B, Fanar-STT-LF, Whisper, SeamlessM4T-v2, and wav2vec2-based models. GPT-4o-transcribe achieves the lowest overall WER at 35.0%, followed by Voxtral-Mini-4B, Fanar-STT-LF, and Whisper-Large-v3 at 41.1%, 45.9%, and 49.5%, respectively, indicating substantial remaining errors across Arabic dialect communities. Performance varies considerably across dialect groups, with no model performing uniformly best across all groups. WER alone also obscures dialectal ASR behaviour: wav2vec2-based models show large WER/CER gaps, where character-level agreement remains much higher than word-level accuracy, motivating joint WER/CER reporting. ALMIEYAR provides a unified benchmark for culturally grounded Arabic ASR evaluation, including the first published benchmark for Ahwazi Arabic. --- ### Revisiting Risky Tackle Detection with Vision Transformers Authors: Syed Ahsan Masud Zaidi, Lior Shamir, Scott Dietrich Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35562v1 This paper is a Track 2 reproducibility companion to an ICPR 2026 study on risky tackle detection in American football prac- tice videos. The original work fine-tuned a Video Vision Transformer (ViViT) on 733 clips labeled with the SATT-3 rubric. It used focal loss, Taguchi L18 augmentation, and 5-fold cross-validation. It reported risky- class recall of 0.67 and risky-class F1 of 0.59. This companion documents the released artifact and traces those numbers to specific scripts, fold out- puts, and aggregation files. The reproduced headline is run_15. It com- bines Gaussian noise with static brightness decrease and uses no rotation and no flip. Its fold-mean risky recall is 0.667 and its fold-mean risky F1 is 0.588. These values match the published headline after rounding. The ablation shows that brightness is the dominant factor. Its risky-recall main-effect range is 0.055, which is larger than the ranges for rotation, flip, and noise. Without augmentation, ViViT reaches risky recall of 0.545 and does not exceed the C3D baseline of 0.583. The raw clips show iden- tifiable student athletes, so they cannot be redistributed. The artifact provides a public sample for pipeline checks and a controlled route for full-data review. --- ### RSI-Master: Structuring Experiments to Guide Autonomous Model Improvement Authors: Yaxin Du, Xiyuan Yang, Zhifan Zhou, Yujie Ge, Cheng Wang, Jiajun Wang, Sijie Chen, Zehui Liu, Yuxin Zhang, Weicheng Gu, Julian Zhang, Zixing Lei et al. Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35561v1 Recursive self-improvement (RSI) seeks to enable AI systems to participate in improving their own capabilities. A concrete pathway is autonomous model development, where agents iteratively explore post-training strategies to improve a base model. This setting faces two challenges: agents may exploit open-ended experimental actions through hacking, and repeated experimentation may lead to strategy lock-in, where an early direction is refined rather than reconsidered. We introduce RSI-Master, which addresses the two challenges at two levels: regularize step-wise actions, avoiding hacking behaviors, and promote well-structured exploration of research directions, avoiding strategy lock-in. RSI-Master consists of an Experiment OS, which enables regularized experimental actions and maintains persistent, traceable experimental records, and Reviewer-Guided Research Orchestration, which organizes Workers and Reviewers in a dynamically growing research DAG. Workers explore diverse research directions and Reviewers compare evidence across related experiments for subsequent explorations. On PostTrainBench with Qwen3-4B-Base, it averages 54.49 versus 46.53 for the strongest agent baseline, with a 0.0\% hacking rate. Scaling to 35B model, RSI-Master surpasses the human-developed Instruct model on LiveCodeBench-v6 (41.21 vs. 37.36) and SciCode, and reaches a nonzero score on HorizonMath, a benchmark of unsolved research problems on which most frontier models score near zero. --- ### WorldPlay2: Extending Real-Time Interactive World Models in Control and Horizon Authors: Haiyu Zhang, Wenqiang Sun, Tengfei Wang, Junta Wu, Jun Zhang, Yunhong Wang, Yu Qiao, Chunchao Guo Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35560v1 Interactive world models require responding in real time to versatile controls and maintaining long-horizon consistency. However, modeling heterogeneous controls remains difficult, while explosive contexts and unstable distillation impede achieving both long-horizon consistency and real-time responsiveness. In this paper, we present WorldPlay2, an interactive world model that couples a factorized hybrid control interface with a co-design of compressed memory and stable distillation. 1) Our factorized hybrid control interface integrates frame-aligned action control with structured semantic control that explicitly disentangles scene appearance, character identity, and dynamic semantic events, thereby facilitating effective control learning. 2) To achieve efficient long-horizon modeling, we compress historical contexts into compact memory tokens shared by the autoregressive student and the bidirectional teacher. This design enables clip-wise, memory-conditioned score evaluation instead of jointly processing an entire long rollout, substantially reducing distillation overhead. 3) We further propose Stable Forcing, which initializes the autoregressive student via a few-step strategy and leverages full-rollout replay to preserve the quality of long-horizon rollouts, ensuring robust and stable distillation. Extensive experiments demonstrate the strong generalizability of our model and its superior performance compared to existing methods. --- ### From Search to Research: Exploring Search Scaling in Autonomous Quantitative Factor Mining Authors: Kangcheng Deng, Hui Cai, Jiacheng Lu, Chester Zhongshu Qian, Rui Sun, Beidi Luan, Jing Li, Daxin Jiang, Zuo Bai Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35559v1 Inference scaling has been shown to improve large language model (LLM) performance, and this principle naturally extends to autonomous LLM agents through increased search budgets, which we refer to as *search scaling*. Although prior work has characterized the mechanisms, scaling behavior, and performance limits of LLM inference scaling, much less is known about these questions in autonomous research. Therefore, we investigate how search scaling affects research performance and what mechanisms drive these gains using 50 quantitative factor-mining tasks grounded in financial research reports. Each task requires an agent to carry out an end-to-end research loop, from interpreting a hypothesis and implementing it in code to evaluating and iteratively refining the resulting factor. Across nine models, we examine how model capability, search depth, and search organization shape factor quality by tracing performance across varying budgets, transferring intermediate research states between models, and comparing different search strategies. We find that (1) initial performance is more strongly associated with model capability, while deeper search can narrow cross-model gaps; (2) model grafting shows that the early research state materially shapes final performance; and (3) parallel search outperforms sequential search under the same iteration budget, consistent with benefits from broader coverage of the search space. Further trajectory analysis shows that higher-performing models more effectively diagnose failures, revise search directions, and preserve the intended economic hypothesis when selecting candidates. These findings suggest that future progress in autonomous research will require stronger models together with adaptive policies for deploying test-time computation throughout the research process. --- ### The Compiler May Read It, the Agent May Not: Keeping Part of a Research Code Away from a Coding Agent Authors: Shobhan Roy Published: 2026-09-28 Categories: cs.CR, cs.AI, cs.SE arXiv: https://arxiv.org/abs/2609.35557v1 The compiler must read modules a physics-based solver cannot build without; the coding agent must not read that intellectual property. The harness does not ship that rule. We classified fifteen read routes against a container, permission rules and a sandbox. None of the three can tell which program is reading. --- ### Simplex Diffusion Models Authors: Justin Deschenaux, Alexandre Galashov, Andrew Campbell, Li Kevin Wenliang, James Thornton, Arnaud Doucet, Valentin De Bortoli Published: 2026-09-28 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.35553v1 Diffusion models have revolutionized generative modeling for continuous data through the gradual refinement of a belief state. This iterative refinement has not yet carried over to discrete diffusion models, which discard uncertainty at intermediate steps through categorical sampling (information collapse). We propose Simplex Diffusion Models (SDMs), a framework that lifts the diffusion process to the probability simplex to represent beliefs over categories. SDMs admit probability paths with closed-form reverse transitions and can be trained with a simple cross-entropy loss. Contrary to earlier proposals such as Dirichlet Flow Matching which requires integrating an ordinary differential equation, we introduce a DDIM-like sampler with a tunable level of stochasticity. Because SDMs operate on samples on the simplex, they can carry uncertainty across denoising steps, which mitigates information collapse. On OpenWebText, SDMs are competitive with strong Discrete Diffusion baselines, achieving $17.0$ GenPPL at $5.46$ unigram entropy in 64 sampling steps, close to real validation data. Even without Self-Conditioning (SC), SDMs outperform masked and uniform diffusion (with SC or predictor-corrector sampling) on code generation (TinyGSM, $T=0.1$; $49.0\%$ vs. $45.8\%$). Distilled down to 8 steps, SDMs solve $32.1\%$ of GSM8K problems, more than distilled Discrete Diffusion models with 128 steps ($21.4\%$). --- ### BaRe-Mem: Bayesian Reliability Memory for Robust and Adaptive Agent Consultation Authors: Peilin Feng, Zhengyang Huang, Soujanya Poria Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35551v1 In multi-agent systems, reliable consultation is challenging because advisor capabilities vary across tasks, and misleading information can make consultation worse than autonomous reasoning. We introduce BaRe-Mem, an online Bayesian reliability memory for multi-agent consultation. It estimates advisor reliability based on the central model's internal belief representations and updates these estimates from historical interactions. These estimates modulate the influence of advisor responses and guide the choice between consultation and autonomous reasoning. Across nine benchmarks and six central models, BaRe-Mem is more robust to misleading advisor information than debate and majority voting. On the more challenging tasks, it remains above autonomous reasoning across all tested misleading levels. Moreover, we extend the BaRe-Mem mechanism to worker allocation in agent teams. On the MuSiQue benchmark, BaRe-Mem improves task completion over routing by historical success counts and identifies capable workers earlier. --- ### RareDx: Controlled Knowledge Integration and Graph-Grounded Policy Optimization for Rare-Disease Diagnosis Authors: Bo Zhang, Yuchen Wang, Dongbai Li, Matthew Yu Heng Wong, Qingkai Zeng, Lijun Wang, Tien-Yin Wong, Peng Cui, Tianyu Liu Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35549v2 Rare-disease diagnosis is a long-tail reasoning problem: phenotypes are incomplete, individual disorders are sparsely documented, and relevant evidence is distributed across ontologies, gene annotations, and biomedical text. Language models consequently favor common conditions, miss rare candidates, or produce plausible but invalid names. We introduce RareDx, which couples controlled evidence use with knowledge-graph-grounded policy optimization. RareDx-Harness normalizes heterogeneous records into one ranked-diagnosis task and compares direct inference, static retrieval, adaptive tools, and structured phenotype-gene-disease reasoning over a shared knowledge layer. The training pipeline combines Top-10 post-training with RareDx-KGPO, our knowledge-graph-grounded policy optimization method. Its reward projects predictions into a canonical disease graph and integrates curated graded relevance, ontology proximity, biomedical similarity, and phenotype consistency. Vocabulary and output-budget constraints prevent dense partial credit from rewarding fabricated or overlong differentials. Across eight benchmarks, the complete RareDx system centered on Qwen3.5-9B reaches 38.34 macro Hit@10, 1.60 points above GPT-5.5 under the archived protocol; a disjoint validation-selection audit retains a 6.80-point routing gain over Direct on held-out cases. The 27B system reaches 23.53/36.56/40.76 at Hit@1/5/10. Controlled ablations show that retrieval is not uniformly helpful and that controlled routing is central to the gain. These results indicate that structured medical knowledge can turn a compact model into a competitive diagnostic ranker across heterogeneous long-tail settings in clinical practice. --- ### Graph World Models for Constrained Epidemic Policy Planning Authors: Yiqi Su, Rashed Shelim, Lingyi Wang, Walid Saad, Naren Ramakrishnan Published: 2026-09-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.35545v1 Epidemic policy planning often requires coordination between geographical regions, taking into account mobility-driven spillovers and how to make use of limited resources. Existing methods either lack action-conditioned models of coupled dynamics or cannot guarantee per-period feasibility. We present EpiMind, a graph world model framework for constrained epidemic policy planning across regions. A graph-factored recurrent state-space model generates joint policy-conditioned rollouts from regional latent beliefs, while graph-temporal ADMM optimizes regional interventions, enforces shared-resource feasibility through projection, and evaluates temporal specifications under the learned model. EpiMind reduces admission RMSE by 29% relative to graph-free dynamics modeling, plans within 1-5% of the best feasible constant policy with guaranteed shared-budget feasibility, and outperforms all deployable baselines across three resource budgets in real-context evaluation. These results demonstrate that graph-structured policy imagination with explicit constrained coordination supports effective epidemic interventions from learned dynamics. --- ### Less Sycophancy, Stronger Refusal? Lessons for AI Safety from Mechanistic Interpretability Authors: Xu Wang, Difan Zou, Xuansheng Wu Published: 2026-09-28 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.35544v1 Reliable refusal of harmful requests is essential to the safe deployment of language models. Because excessive eagerness to please users may undermine existing refusal capabilities, reducing sycophancy offers a potential route to stronger refusal beyond the harmful scenarios covered by safety training. We investigate this possibility using compensatory feature injection (CFI), a training technique designed to limit the acquisition of a target concept by supplying its associated activation during learning. Across three Qwen3.5 base models, we use sparse autoencoders (SAEs) to identify the top-ranked sycophancy feature from paired sycophantic and independent responses, then validate its behavioral influence through inference steering. We subsequently inject the selected feature during supervised fine-tuning on sycophantic targets. Positive injection reduces learned sycophancy after removal (by 62.0% relative to ordinary fine-tuning in 35B-A3B), whereas modest negative injection increases it. Unexpectedly, these reductions in sycophancy do not consistently improve direct refusal of harmful requests, motivating a narrower evaluation of the same harmful intents under user pressure. In this setting, ordinary fine-tuning on sycophantic responses substantially weakens refusal, while selected checkpoints trained with positive injection recover part of the loss, including approximately 95% in 35B-A3B. These findings show that persistent sycophancy reduction does not guarantee stronger direct refusal, while identifying recovery under user pressure as a distinct, conditional benefit of training intervention. --- ### Learning the Robustness Mechanism with Bilevel Optimization Authors: Yiyang Shen, Qihang Lin, Weiran Wang Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35541v1 We propose a distributionally robust learning framework where parameters defining the robustness mechanism are learned from held-out data instead of extensively tuned. Using bilevel optimization with both upper and lower level minimax problems, we create two instances of our framework to tackle setups with and without group labels in the training set. Theoretically, we provide sample complexity analysis for our robustness mechanism learning paradigm, showing that it achieves generalization guarantees comparable to exhaustive grid search while being more computationally efficient. Empirically, we evaluate our framework under a challenging setup when both intra-group and inter-group test distribution shifts occur at the same time, thereby demonstrating the efficacy and scalability of our method. --- ### Continuous Context Management Authors: William Hoy, Jingxuan Fan, Nurcin Celik, Xu Pan Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35540v1 Long-horizon large language model (LLM) agents commonly retain their complete interaction history until compaction is triggered at a predefined threshold. We study Continuous Context Management (CCM), which performs compaction at every turn to prevent interaction history from accumulating in the active prompt. At each turn, a CCM agent emits an updated memory together with an environment action; its next prompt contains the original task, retained memory, and newest observation rather than the complete transcript. We first evaluate CCM without fine-tuning on TerminalBench-2 using Claude Sonnet 4.6, Claude Opus 4.6, GLM-5, and Kimi K3. CCM substantially reduces cumulative input usage and active-prompt size, although it lowers task success for most models while preserving performance for Kimi K3. We use GRPO with privileged full-history distillation to improve CCM in open-weight models. A frozen copy of the student's initial model scores each sampled student action under the complete history reconstructed from that student's rollout, providing dense action-token supervision without a separate teacher rollout or reference solution. On WebShop, this objective substantially improves CCM over GRPO at both evaluated model scales and surpasses full-history GRPO for Qwen3-4B-Instruct, though not for Qwen3-8B. On Endless Terminals, the augmented method provides a modest improvement over GRPO, with both CCM policies outperforming the untrained full-history baseline. These results demonstrate that CCM is a viable inference paradigm for agents operating with substantially reduced retained context and that its performance can be improved through reinforcement learning with privileged full-history distillation. --- ### Learning to Reason with Persistent Object States for Video Instance Segmentation Authors: Yongxue Xu, Boxue Yang, Ziqian Liu, Shaoqiu Zhang, Rui Qian, Haopeng Chen Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35539v1 Video segmentation models maintain object identities by carrying instance information across frames. Under prolonged occlusion, reappearance, or interactions between similar instances, however, an unreliable update can overwrite a valid history and cause persistent identity drift. We introduce POSReasoner, a trainable, plug-and-play framework that explicitly decides when an observation should change an object's state. Each persistent state records identity, confidence, and absence history. A sparse state-observation graph supports Propose-Verify reasoning: provisional associations are revisited using object history, predicted presence, and competition among identities. The verified decisions determine whether to retain, update, reactivate, or suppress each state, while a learned gate controls the evidence written back to memory. Only verified transitions update the persistent state used in subsequent frames. POSReasoner uses standard video annotations and keeps the base model frozen, enabling integration with diverse VOS and VIS architectures. Experiments across long-term VOS and VIS benchmarks show consistent improvements over strong baselines, with the largest gains under occlusion and object reappearance. --- ### Optimal Networks for Agentic Information Aggregation Authors: MohammadHossein Bateni, Zahra Hadizadeh, MohammadTaghi Hajiaghayi, Mahdi JafariRaviz, Shayan Taherijam Published: 2026-09-28 Categories: cs.LG, cs.GT, econ.TH arXiv: https://arxiv.org/abs/2609.35537v1 We study information aggregation in the networked learning model introduced by Kearns, Roth, and Ryu (SODA 2026). There is a fixed distribution over $d$ features and a common label. Agents learn in topological order on a directed acyclic graph. Each observes a subset of the features and its parents' predictions, fits a linear predictor to minimize mean squared error, and passes only its prediction forward. The global predictor is the best linear predictor using all features. Kearns, Roth, and Ryu show that the output agent's error approaches the global predictor's error along sufficiently deep paths with suitable feature coverage, while insufficient depth can prevent aggregation even in large networks. In contrast to their main focus on a given graph and feature allocation, we consider the limits of the model under two settings. In the adaptive designer setting, a designer chooses the graph, feature allocation, and output agent knowing the distribution. In the oblivious designer setting, the designer fixes all three before an adversary chooses the distribution. Each agent observes one feature and receives predictions from a limited number of parents. We call the aggregation exact when the output agent matches the global predictor exactly. For $d\ge3$, we show that no finite depth guarantees exact aggregation for every distribution with one parent per agent, even when the designer knows the distribution. In contrast, two parents per agent suffice for exact aggregation even in the oblivious designer setting. A fixed graph, feature allocation, and output agent achieve this for every distribution at depth $O(d\log d)$. Knowing the distribution reduces the depth to $O(d)$. Both constructions use $O(d^2)$ agents, with a very large constant for two parents. We show the bounds on the depth and number of agents are all optimal up to constant factors. --- ### Look Before You Judge: Training-Free Region Mining for Grounded and Explainable Deepfake Detection Authors: Chia-Ling Chen, Yu-Ting Ta, Jian-Yu Jiang-Lin, Tai-Ming Huang, Ling Lo, Po-Ching Chen, Yan-Tsung Wang, Pei-Heng Li, Ling Zou, Hong-Han Shuai, Wen-Huang Cheng Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35536v1 Multimodal large language models (MLLMs) can explain deepfake verdicts in natural language, but such explanations are not necessarily visually grounded in the visual evidence underlying the prediction. A model may describe plausible artifacts inferred from language priors rather than from image evidence. Existing grounding methods improve visual reliance through decoding or attention interventions, but they generally strengthen grounding over the entire image, making them ill-suited for forensic artifacts that are subtle, spatially localized, and image-dependent. We propose Look Before You Judge, a training-free framework that formulates explainable deepfake detection as a sequential evidence acquisition process. Instead of directly predicting image authenticity from holistic visual reasoning, our framework first identifies image-specific candidate evidence regions by contrasting the MLLM's decoder-to-visual attention between an original image and its Gaussian-blurred counterpart. The identified regions are then inspected individually, and the resulting local evidence is integrated with the global image context before reaching a final verdict. The framework operates without manipulation masks, external forensic models, or parameter updates, making it directly applicable to off-the-shelf MLLMs. Across five open-source MLLMs on TriDF and MMTD-Set, our framework improves detection accuracy by up to 12.8%, reduces CHAIR by up to 33.4% and hallucination rate by up to 21.3%, and outperforms representative training-free decoding and attention methods. --- ### ARISE: Adapting to Evolving Capability Gaps in Agentic Reinforcement Learning Authors: Kun Feng, Yuchen Fang, Yiyang Tan, Shuqi Gu, Yongxiang Zhao, Yu Liu, Xingyu Lu, Lintao Ma, Kan Ren Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35532v1 As a long-horizon agent improves through experience, previously observed weaknesses may recede while new limitations emerge, continually changing what it still needs to learn. Yet the learning process often remains tied to a static view of these needs: fixed behavioral criteria and training priorities can become misaligned with evolving agent capabilities, while sparse task-level feedback makes such misalignment more difficult to detect. Even when capability gaps are identified, rollouts from the current policy may repeatedly reproduce the same failures rather than explore better alternatives. To address this, we introduce Adaptive Rubric-Skill Co-Evolution (ARISE), a reinforcement learning framework that uses rollout evidence to continually adapt evaluation criteria, exploration guidance, and training priorities. Rubrics evolve to reward partial behavioral progress, while their paired skills are refined and selectively activated to guide exploration toward unresolved weaknesses. Alongside this co-evolution, capability-based adaptive sampling prioritizes tasks that target behaviors needing further improvement. Experiments on two challenging long-horizon agent benchmarks, SkillsBench and Terminal-Bench, demonstrate that ARISE successfully enhances both overall task performance and training efficiency. The project page is at https://foundation-model-research.github.io/ARISE . --- ### AutoRef: Harness Optimization for Agentic Multi-Reference Image Generation Authors: Yuta Oshima, Ku Onoda, Yusuke Iwasawa, Masahiro Suzuki, Yutaka Matsuo, Hiroki Furuta Published: 2026-09-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.35530v1 Recent image generation models can take multiple reference images as input and combine them into a new image. However, multi-reference image generation remains challenging: models may omit or duplicate subjects from the references, or produce images in which multiple subjects appear unnaturally pasted. Recent work has proposed image generation agents that combine image generation models, reasoning models, and a harness, which is an executable program that specifies how reference images are interpreted, how generation is performed, how outputs are diagnosed, and how the final image is selected. In multi-reference generation, however, references play different roles and outputs must satisfy many criteria at once, such as fidelity to each reference and the naturalness of the whole image, so many parts of the harness could be improved, from how references are processed to how outputs are diagnosed. This makes it hard to predict which changes will improve performance and by how much, and good harnesses difficult to design by hand; indeed, human-written harnesses vary widely in performance. We therefore propose AutoRef, which optimizes the harness automatically while keeping both models frozen: a coding agent iteratively rewrites the harness code. AutoRef separates the tasks whose feedback informs proposals from the tasks used to select candidates, and continues the search from a beam of the top-ranked harnesses on the selection tasks. Using this procedure, we discover AutoRef-Harness, which improves the open-weight FLUX.2 (klein) 4B from 5.72 to 7.37 on held-out four-reference tasks of the MultiBanana benchmark, matching or exceeding proprietary models including Nano Banana Pro and GPT-Image-1.5. Without re-optimization, the same harness also improves results when the generator, number of references, benchmark, evaluator, or reasoning model differs from those used in the search. --- ### Let the Neurons Die: Exploiting ReLU-Induced Model Degradation Authors: Kexin Li, Wenjun Qiu, Joshua Abraham, Aditi Maheshwari, David Lie Published: 2026-09-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.35528v1 Rectified linear unit (ReLU) networks can suffer from dying neurons, where units with persistently negative pre-activations produce zero outputs, blocking gradients through their activations. To exploit this failure mode, we present three training-time availability attacks based on data ordering and poisoning. We begin with the basic dynamic data-ordering attack (DOA), which greedily constructs a training prefix by selecting the next example that minimizes the target layer's post-update weight sum, aiming to push ReLU units toward negative pre-activations without modifying training samples or labels. We then develop two poisoning attacks, IG-DOA and IG-SKA, which use gradient inversion to synthesize class-conditioned samples by matching reference gradients in adverse model states constructed through data ordering or soft knockout, respectively. Soft knockout rearranges weights across adjacent layers to concentrate negative contributions. On a fully connected ReLU network trained on MNIST, ordering 100 of 60,000 training examples reduces test accuracy from 96% to 95% after only five epochs. Adding 200 poisoned samples from a single class reduces test accuracy to approximately 86-88% after five epochs in most evaluated conditions, compared with approximately 96% under clean training. These results demonstrate that ReLU-targeted data ordering and poisoning can impair learning without directly modifying the victim model's parameters. --- ### GeoGAE: Scalable Graph-Level Autoencoding via Hyperball Cloud Representations Authors: Radosław Nowak, Anna Bielawska, Bogusz Stefańczyk, Maciej Sanocki, Paweł Wawrzyński Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35527v1 Embedding structured objects into Euclidean spaces has enabled a wide range of successful machine learning applications. Such objects include words, documents, image patches, time series, and graph nodes. In contrast, embedding entire graphs remains a challenging problem. Existing methods either sustain the original order of the graph nodes or match the output nodes to the input ones, both of which create scalability issues. In this work, we propose a graph representation as a cloud of hyperballs, which allows us to define a specific, typically unique, node ordering. Based on this representation, we propose GeoGAE, an autoencoder, in which the Transformer encoder translates a hyperball cloud into a graph-level embedding, and the Transformer decoder translates the graph-level embedding back into the graph. This formulation enables the model to capture both the global graph structure and local relational patterns. We evaluate our method on multiple graph datasets, spanning various domains. The results demonstrate effectiveness of our method in encoding and reconstructing graphs from their embeddings. ---