# Research Papers — 2026-10 · bundle page 143 of 156

> Full-text markdown of papers #7101–7150 of 7799 published in 2026-10, newest first. AI-readable mirrors of the full 7799-paper 2026-10 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-10/page/143
[← Previous bundle](https://aipod.dev/papers/archive/2026-10/page/142.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-10/page/144.md)

### Grounded in Time: A Multi-Source Dataset and Benchmark for Temporal Grounding in Robotic Manipulation

Authors: Yi Wang, Yang Yang, Guangqi Xu, Sumin Lin, Ning Kang, Pengxiang Lu, Xiaotong Chen, Zeyu Xue, Ping Deng, Xing Liu, Chenguang Yang, Zhenyu Lu
Published: 2026-10-03
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2610.04255v1

Robotic manipulation often requires inferring task-relevant states from past interactions when the current observation alone is insufficient to determine the appropriate action. Despite progress in benchmarking memory-augmented vision-language-action (VLA) models, application-oriented tasks requiring history-dependent semantic inference remain underrepresented. We introduce GiT (Grounded in Time), a dataset and benchmark for grounding manipulation decisions in past events across biolaboratory, household, and industrial scenarios. It includes real-robot and Universal Manipulation Interface (UMI) style demonstrations covering 18 bimanual tasks, together with simulation data and a ManiSkill-based evaluation suite covering nine tasks. Fine-grained subtask annotations and annotated counterfactual task pairs, in which similar current observations require different actions depending on prior events, support policy learning and targeted evaluation of history use. Evaluations of representative end-to-end VLA models in simulation and on selected real-world tasks reveal substantial room for improvement in history-dependent manipulation. The dataset and benchmark are available at the project page.

---

### Spec2Game: Can LLMs Generate Complete Playable Games from Detailed Specifications?

Authors: Yixue Cai, Yuzhe Zhao, Hanxiang Chao, Qingsen Ma, Ziheng Xiong, Jinhu Qi, Irwin King
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04253v1

Generating an executable program does not necessarily mean that it correctly implements the behavioral requirements specified in natural language. To evaluate large language models' ability to realize detailed specifications as complete interactive programs, we introduce Spec2Game, a benchmark that requires models to generate complete Pygame projects from detailed natural-language game specifications. Spec2Game comprises 15 game families and 150 task instances, with one canonical task and nine controlled rule variants per family, spanning three levels of implementation complexity. Using source-code, runtime, and visual evidence, we evaluate generated projects along four dimensions---Executability, Specification Realization, Code Quality, and User-Facing Quality. Across 14 LLMs and 3,330 generated projects, we find that high executability does not imply faithful specification realization. Component-level analysis further shows that models perform substantially better on Game Element Modeling than on Rule and Mechanism Modeling or Goal and Termination Modeling, indicating that faithfully implementing game rules and termination logic remains a major challenge.

---

### Optimizer Geometry Sets the Pace: Spectral Learning Dynamics in Matrix Factorization

Authors: Mahalakshmi Sabanayagam, Simon Lucey
Published: 2026-10-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.04249v1

Recent successes of matrix- and curvature-based optimizers have renewed interest in how update geometry shapes learning. These methods normalize or precondition updates, changing how different components progress during training. In deep matrix factorization, the geometry that slows gradient descent (GD) favors low-rank solutions by delaying the emergence of small singular modes. This raises the question of what remains of that spectral bias when normalization or curvature correction weakens or removes the slowdown. We study how optimizer geometry and depth jointly govern this behavior through a common framework for singular mode dynamics. Under explicit balance and alignment assumptions, we derive birth, saturation, and decay laws for Euclidean GD, coordinate-wise updates of SignGD and an instantaneous Adam approximation, spectral updates of Muon and cumulative Shampoo, and a block curvature model of K-FAC. The resulting picture is not a simple ordering from stronger to weaker low-rank bias. To highlight, SignGD and ideal Muon eliminate the divergent birth barrier and drive unsupported modes to zero in finite time. Cumulative Shampoo initially retains GD's depth-dependent barrier, then accumulated gradients produce a catch-up phase while making previously active modes increasingly persistent. Undamped K-FAC cancels the factorization-induced slowdown while preserving the ordering of the target singular values, whereas positive damping introduces a spectral threshold below which the slow GD phase laws reappear. These results give normalization, accumulated state, damping, and depth a direct interpretation as controls determining when modes emerge, persist, and disappear during training.

---

### Benchmarking Psychological Dynamics in Generative Agents

Authors: Sumer S. Vaid, Ashley V. Whillans
Published: 2026-10-03
Categories: cs.HC, cs.AI
arXiv: https://arxiv.org/abs/2610.04246v1

Large language models (LLMs) are increasingly deployed to simulate human behavior, acting as computational replicas of human subjects. Yet the lived psychological experience of humans is difficult to benchmark, particularly as it unfolds over time. We introduce a psychometric benchmark for computational replicas: personas that carry a fixed identity through an evolving sequence of events. Built entirely from published norms and meta-analytic effects, the benchmark scores two dimensions of psychological realism. The first, internal validity, quantifies whether generated trajectories reproduce the internal structure of repeated human measurement: distributions, the between- versus within-person variance partition, temporal dependence, and range. The second, external validity, quantifies whether replicas recover established trait, state, and indicator relations. Across 36 open-weight and proprietary LLMs from nine developers (1B-671B parameters), most recover the direction of established relations (84.3% mean agreement) and the variance partition (26 of 34), yet the within-person correlation and distributional structure elude recovery. Internal validity is independent of scale and capability: a mid-size open LLM (Gemma-3-27B) strikes the best trade-off between the two dimensions. The benchmark is a precondition for using computational replicas in causal inference across domains (e.g., marketing, healthcare), and identifies within-person grounding as the central challenge ahead.

---

### On the Steering Dimensionality of Refusal in Language Models

Authors: Han Wang, Erik Miehling, Dennis Wei, Karthikeyan Natesan Ramamurthy, Huan Zhang
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04245v1

Existing activation steering methods often assume that a high-level concept can be mediated by a single steering direction. To support this, two complementary interventions should be achieved: additive steering should induce the target behavior, while directional ablation should suppress it. Yet behaviors may occupy richer activation geometries beyond a single direction, and semantically similar behaviors may be represented by distinct directions. In this work, we study how many directions can reliably control two different types of refusal behaviors: refusal triggered by the safety alignment and refusal in general contexts. Given the limited expressive capability of a single steering vector, we study the general setting of steering subspaces and introduce the notion of steering dimensionality as the minimum subspace dimensionality required to reliably control a behavior. We characterize sufficient steering subspaces that cover the full extent of the target behavior through both the (monotonic) improvement before the sufficient dimensionality, and the saturation beyond it. Empirically, we find that refusal triggered by safety alignment is 1-dim steerable, while multiple distinct steering directions can achieve comparable control. In contrast, refusal in general contexts exhibits substantially richer activation geometry where even 5-dim steering subspaces fail to reliably capture its full steerable variation. Our results reveal that the activation geometry underlying refusal is highly context-dependent and can be substantially more complex than a single linear steering direction.

---

### Stochastic Adaptive Fourier Decomposition for Operator Learning

Authors: Pengqing Shi, Liming Zhang, Tao Qian, Stephen Tierney, Jie Yin, Junbin Gao
Published: 2026-10-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.04241v1

Fourier Neural Operators (FNOs) offer an efficient paradigm for solving partial differential equations (PDEs). However, FNOs rely on a fixed Fourier basis and hard frequency truncation, which inherently limit their ability to model non-periodic, localized, and fine-scale solution structures. We propose the Stochastic Adaptive Fourier Decomposition Neural Operator (SAFDNO), a spectral neural operator that replaces predefined Fourier modes with an adaptive Takenaka-Malmquist (TM) orthonormal system derived from the theory of Stochastic Adaptive Fourier Decomposition (SAFD). Instead of performing an expensive greedy pole search in classical SAFD, SAFDNO amortizes stochastic pole selection through a neural pole predictor and constructs the adaptive TM system directly from latent features. The resulting operator performs learned filtering on coefficients of analytic branches in the Hardy space with respect to the input-adaptive TM system, preserving the global receptive field of spectral operators while providing a more flexible representation than using fixed spectral bases. Across nine PDE benchmark problems, SAFDNO achieves the best performance on all six regular-grid problems among strong neural operator baselines, with especially notable gains on Darcy, Burgers, and Navier-Stokes, where fixed Fourier modes are often less effective at modeling localized oscillations, sharp transitions, and multiscale structures. SAFDNO also exhibits stronger zero-shot super-resolution performance and shows less performance degradation when deployed on finer discretizations. These results suggest that the input-adaptive TM system provides a promising alternative to fixed spectral representations for neural operator learning.

---

### Attention-Based Surface Representation Learning for Robot State Prediction and Open-Ended Surface Classification

Authors: Oleg Kushnarev, Alexander Belyaev
Published: 2026-10-03
Categories: cs.RO, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2610.04240v1

For ground robots operating in outdoor environments, understanding the properties of the underlying terrain is essential for ensuring reliable operation. In most perception-based studies, this problem is formulated as categorical classification with a fixed number of classes defined during training. We propose an approach that enables new surface classes to be added as trainable vectors, which can subsequently be used to address higher-level tasks. By employing a learning paradigm based on predicting the robot's next state in time and using attention blocks, we improved classification accuracy to 98.56% on the Belyaev-Kushnarev dataset and 94.8% on BorealTC.

---

### Conformal Prediction with Paraphrase-Aware Scoring for LLM Uncertainty Quantification

Authors: Jiayi Xin, Evan Qiang, Zihan Zhu, Xiang Li, Weijie J. Su, Qi Long
Published: 2026-10-03
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2610.04239v1

Uncertainty quantification (UQ) for large language models (LLMs) aims to provide reliable measures of predictive confidence, yet current methods are often unstable under meaning-preserving perturbations. Semantically equivalent paraphrases can induce substantial variability in predictive confidence, even for methods with formal guarantees, such as conformal prediction. To address this issue, we propose a paraphrase-aware UQ framework robust to semantic rewordings. Our approach trains a lightweight proxy model on LLM hidden states and aggregates its predictions across paraphrases to construct label-wise nonconformity scores. Under score exchangeability, conformal calibration retains marginal coverage. This guarantee can also hold under test-only rewording, provided that the paraphrase pipeline satisfies an additional distributional alignment condition. We evaluate three settings (normal, fully reworded, and semi-reworded) which apply rewording to neither dataset, both calibration and test datasets, or only the test dataset, respectively. Across seven multiple-choice QA benchmarks and multiple model families, our method produces compact prediction sets with empirical coverage generally near the nominal target, even in the semi-reworded setting. Ablation studies show that the learned proxy accounts for most of the reduction in set size, while paraphrase-augmented training and inference-time aggregation improve stability under rewording. Code is available at https://github.com/Raina-Xin/PA_Score.

---

### Humanoid Rickshaw Pulling: Whole-Body Locomotion under Coupled Wheeled Loads

Authors: Yangzhi Yang, Xiansheng Lin, Zhaoming Xie, Xiaobin Xiong
Published: 2026-10-03
Categories: cs.RO, cs.LG
arXiv: https://arxiv.org/abs/2610.04238v2

Humanoid robots could transport payloads substantially heavier than themselves by pulling passive wheeled vehicles instead of carrying the load. This capability, however, creates a coupled locomotion problem: the robot must maintain persistent upper-body contact while adapting to unknown, configuration-dependent forces arising from the payload, vehicle, and terrain. We present a whole-body control framework for humanoid rickshaw pulling that tracks commanded vehicle motion while preserving balance and stable grasps under uncertain load dynamics. During training, a privileged teacher exploits vehicle states, interaction forces, and load properties. Its actions and latent are distilled into a history-conditioned student that implicitly infers coupled dynamics from proprioceptive responses, followed by reinforcement-learning fine-tuning. Comparisons with \emph{No History} and \emph{Only History} baselines show that the resulting policy achieves accurate vehicle tracking while reducing vehicle oscillation, torso tilt, and actuation cost. Behavioral analysis shows that Unitree G1 propels the rickshaw and generates gait-synchronized whole-body reactions that stabilize its lateral and roll motions. Moreover, pulling redistributes joint effort and yields a lower robot-normalized cost-of-transport proxy than unloaded walking over most tested load--speed conditions. On hardware, a single policy performs starting, sustained pulling, turning, and stopping with both rigid payloads and human passengers, handling a loaded rickshaw mass of up to 115~kg without load-specific retuning. These results demonstrate robust heavy-load transportation through coordinated and persistent humanoid--vehicle interaction.

---

### Neyman-Pearson Classification with Indecisions: Simultaneous Control of Selective Type I and Type II Errors

Authors: Bradley Rava, Bowen Gang
Published: 2026-10-03
Categories: stat.ME, stat.ML
arXiv: https://arxiv.org/abs/2610.04233v1

Binary classification often requires controlling both type I and type II error rates at prespecified levels. When the class distributions overlap substantially, the two error limits may be incompatible if every observation must be classified. We study Neyman-Pearson classification with indecisions (NPI), which permits uncertain observations to remain unclassified and assesses selective type I and type II errors among definitive classifications within each class. For an independently trained score, we develop a model-free calibration procedure that combines all candidate threshold pairs passing calibration into a rule whose decision region contains that of every passing candidate. We couple the entire threshold search to a common martingale within each class and compute its finite-horizon crossing probability exactly, avoiding a multiplicity adjustment for the number of candidate pairs. The resulting classifier satisfies both population selective error limits simultaneously with user-specified high probability in finite samples, without assumptions on the accuracy of the fitted score. We bound the additional indecision due to calibration and evaluate the procedure in simulations and applications to recidivism and credit-default prediction.

---

### Chance-Constrained Bi-Objective Evolutionary Optimization for the Open-Pit Mining Operational Planning Problem

Authors: Ishara Hewa Pathiranage, Aneta Neumann
Published: 2026-10-03
Categories: cs.NE
arXiv: https://arxiv.org/abs/2610.04227v1

Open-pit mining operational planning involves allocating limited resources while satisfying production, equipment, and ore-quality requirements. Existing approaches often assume deterministic ore grades or rely on simulation-based evaluation under uncertainty. In this paper, we propose a chance-constrained bi-objective formulation of the open-pit mining operational planning problem under uncertain ore grades. We aim to maximize ore production and minimize fleet cost while satisfying stochastic quality requirements through chance constraints. We model ore grades as independent normally distributed random variables and derive a deterministic reformulation of the two-sided chance constraints, avoiding sampling or simulation during fitness evaluation. We evaluate four multi-objective evolutionary algorithms on benchmark instances under different levels and structures of uncertainty. Our results show that NSGA-II and NSGA-III generally obtain the best feasible ore-production values, with NSGA-II requiring less computational time. GSEMO has the lowest computational cost but obtains feasible solutions less consistently and generally provides lower solution quality. We also observe that higher confidence and uncertainty levels reduce feasibility, with the effect depending on the uncertainty structure. The results highlight the importance of considering both the magnitude and structure of uncertainty in short-term open-pit operational planning.

---

### PaLoRA: Paced Low-Rank Adaptation for Continual Learning

Authors: Yuxuan Li, Fanhu Zeng, Hao Tang
Published: 2026-10-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.04226v1

LoRA-based continual learning methods mitigate catastrophic forgetting through various mechanisms, yet nearly all complement these with small learning rates as a heuristic to restrict gradient scaling magnitude. Such fixed heuristics lack theoretical guidance on how the strength of this restriction should evolve as tasks accumulate. We reveal that even under directional constraints such as nullspace projection, finite-precision updates inevitably leak into the subspace of accumulated prior knowledge along multiple directions. While small learning rates attenuate such leakage, they cannot prevent the accumulated forgetting from intensifying as the effective rank of historical knowledge grows. We show that the optimal magnitude restriction should adaptively increase with this effective rank to balance stability and plasticity, i.e., preservation of previous knowledge and acquisition of new task information. Under an anisotropic leakage model, we derive a pacing law $s^*=\sqrt{R/c}$ that characterizes the optimal scaling of gradient steps, i.e., the magnitude restriction itself, where $R$ is the effective rank of past updates. Based on this insight, we propose PaLoRA, which compresses historical knowledge via adaptive SVD truncation, projects gradients onto the nullspace of prior tasks, and applies rank-aware adaptive pacing. Experiments demonstrate consistent improvements over prior methods, with particularly strong performance in long-horizon settings, achieving substantial gains of 4% accuracy on challenging 50-task ImageNet-A and ImageNet-R benchmarks.

---

### FlashGaze: Training-Free Multi-Scale Patch Pruning For Efficient Video Understanding

Authors: Ziye Zhu, Yanghao Zhou, Lixing Tan, Jialiang Kang, Shuxuan Li, Xiao Yang
Published: 2026-10-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.04225v1

Multimodal Large Language Models (MLLMs) have demonstrated strong performance in video understanding, yet efficiently processing long, high-resolution videos remains challenging. Such videos often contain substantial spatiotemporal redundancy, and processing redundant visual tokens can incur avoidable computational overhead. Many existing methods prune visual tokens during or after vision transformer (ViT) encoding, leaving much of the encoding cost unaddressed. Some approaches prune patches before encoding but rely on learned auxiliary networks for patch selection, incurring additional training and inference overhead. To address these limitations, we propose FlashGaze, a training-free method that reduces spatiotemporal redundancy before ViT encoding without introducing auxiliary networks. FlashGaze uses pixel-space differences as a proxy for information loss and employs Quadtree Dynamic Programming to jointly optimize patch dropping, merging, and keeping under a fixed budget. Experiments on two MLLM backbones across multiple benchmarks demonstrate substantial efficiency gains while largely preserving accuracy. On Qwen3-VL-8B, FlashGaze retains 98% of the full-input baseline accuracy on LongVideoBench while achieving up to 5.4x and 17x speedups in ViT encoding and MLLM prefill, respectively, and reducing peak GPU memory usage by a factor of 1.8. These efficiency gains enable the model to process videos with more frames and higher resolutions on the same GPU hardware, unlocking video understanding at scales previously out of reach.

---

### Adaptive Operator Selection in Bilevel Large Neighborhood Search for Electric Autonomous Dial-a-Ride Problem under Uncertainty

Authors: Ishara Hewa Pathiranange, Aneta Neumann
Published: 2026-10-03
Categories: cs.NE, cs.AI
arXiv: https://arxiv.org/abs/2610.04219v1

The electric autonomous dial-a-ride problem (EADARP) extends the classical dial-a-ride problem by incorporating battery and charging constraints for electric vehicles. In practice, travel-time uncertainty can cause violations of time-window constraints. Large neighborhood search is effective for solving the EADARP, but its performance can depend on the choice of insertion operator during the repair phase. This paper investigates insertion-operator selection within a bilevel large neighborhood search framework for deterministic and chance-constrained variants of the EADARP. In the chance-constrained variant, arc travel times are modeled as independent normally distributed random variables, and upper time-window constraints are enforced probabilistically. We consider six selection methods, namely fixed greedy insertion, fixed regret-based insertion, random selection, a deterministic state-based rule, performance-adaptive ALNS selection, and LLM-based state-aware selection. Experimental results show comparable performance on smaller instances, while differences become more evident on larger and more constrained instances. There is no single strategy that performs best across all instances, and the relative performance of the LLM-based, rule-based, and ALNS strategies varies with the problem instance and experimental setting.

---

### TCMClinicalReason-Bench: Can Language Models Reason from Pathogenesis to Prescription over Real-World Clinical Cases?

Authors: Jirui Dai, Chenkai Zhang, Yan Jia, Yukai Wang, Ruiyang He, Changyong Luo, Zhi Liu
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04215v1

Large language models (LLMs) can generate clinical narratives that are insufficiently grounded in patient-specific evidence. In traditional Chinese medicine (TCM), errors can propagate from etiology and pathogenesis through syndrome diagnosis and treatment principles to prescription generation. We developed TCMClinicalReason-Bench using 2,000 multicenter electronic health record cases to distinguish case-grounded responses from fluent but unsupported diagnostic and therapeutic conclusions. Five general-purpose and two TCM-specific LLMs were evaluated in zero-shot settings. An evidence-constrained rubric assessed seven diagnostic and therapeutic components and three cross-block relations, allowing case-supported alternatives. Qwen3.7-Plus with TCM retrieval served as the automated judge, alongside parallel blinded ratings by five senior TCM clinicians on a 600-case subset. Structural completeness was nearly saturated (99.3-100.0%), but normalized content scores ranged from 40.7% to 54.1%. The five general-purpose models averaged 50.0%, versus 41.3% for the two smaller TCM-specific models. Cross-block logic consistency ranged from 60.8% to 66.8% and correlated moderately with content across cases (Pearson's r = 0.515-0.656). Deficits were greatest in prescription generation, prescription analysis, and symptom-guided modification. In judge stress testing on 100 independent cases, perturbation detection rates across the three relations were 57%, 56%, and 31%, with contradictions detected more reliably than omissions. Separating component quality from cross-block consistency localizes failures missed by endpoint and completeness metrics and identifies where clinician oversight remains necessary.

---

### FTD-GNO: Memory-Efficient Graph Neural Operators through Functional Tensor Decomposition of the Kernel

Authors: Xiaomin Zhang, Boyue Wang, Junbin Gao, Yongli Hu anbd Baocai Yin
Published: 2026-10-03
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2610.04212v1

Graph Neural Operators (GNOs) provide flexible surrogate models for learning solution operators of partial differential equations (PDEs). However, standard GNOs typically parameterize the integral kernel with a monolithic neural network and evaluate kernel interactions over graph edges, leading to substantial computational and memory overhead at high resolutions or with large neighborhoods. To address these limitations, we propose Functional Tensor Decomposition Graph Neural Operator (FTD-GNO), a memory-efficient GNO framework that decouples the high-dimensional continuous integral kernel into low-dimensional mode-wise functions. By instantiating the kernel with classical tensor decomposition formats, including CP, Tensor-Train, and Tucker decompositions, FTD-GNO enables algebraic reconstruction of the integral operator without explicitly materializing full edge-wise kernel tensors. This factorized formulation reduces the memory footprint of kernel evaluation and aggregation while retaining the continuous operator-learning structure of GNOs. Theoretical complexity analysis shows that FTD-GNO substantially lowers parameter and activation-memory costs associated with high-dimensional kernel construction. Experiments show lower peak memory than the corresponding unfactorized graph-integral baselines, with shorter recorded training times. Fourier-graph experiments further demonstrate that FTD can improve the efficiency of a graph-integral layer within a hybrid operator and has good scalability.

---

### CurveCodec 2: Skeleton-agnostic animation compression with a learned entropy model

Authors: Mingyi Shi, Huancheng Lin, Xuelin Chen, Taku Komura
Published: 2026-10-03
Categories: cs.GR, cs.AI, cs.RO
arXiv: https://arxiv.org/abs/2610.04211v1

Skeletal motion is stored as every joint's transform at every frame, yet most of it is implied by the body rather than by what the motion is about. Compression is one way to ask what a motion must still say once the body is known, and a production codec must answer it for any skeleton with a stated error bound. Our earlier codec, CurveCodec, matched the mean error of ACL, the production library of modern game engines, with a learned prior over sparse anchors, but not ACL's worst case, and it counted its payload as floats rather than bits. Here we ask where the redundancy of skeletal motion lies and which part of a codec a learned model should take over. Measurements give three answers. At production precision the largest saving comes from predicting each quantized curve from its own past, the second from choosing per joint, in closed loop through the hierarchy, which samples not to code. On the gaps such an encoder leaves, a nearest-neighbour oracle over millions of training samples is no better than linear interpolation, and no learned in-betweener we tried paid for itself. What a network does learn is the distribution of the residuals the codec must send. CurveCodec 2 codes every sub-track as a curve in the log map, quantized in closed loop and thinned to rate-distortion-selected keys, with residuals entropy-coded under a small learned model whose integer inference is bit-exact across platforms. Two contracts are verified on every decoded clip: ACL's own worst case per joint within a stated tolerance, or ACL's mean error per clip. On a held-out test side of 4,472 clips from 33 datasets, CurveCodec 2 needs 0.37x ACL's bytes at ACL's default precision of 0.01 cm under the worst-case contract and 0.22x at 0.1 cm under the mean contract, decodes on one CPU core, and transfers without retraining to a species absent from training. Project page: https://rubbly.cn/publications/curvecodec/

---

### Can LLMs Separate Pasted Artifacts from User Speech? Absorption at Unmarked Prompt Seams

Authors: Sugam Panthi, Muhaiminul Yeamin, Rabab Abdelfattah
Published: 2026-10-03
Categories: cs.CL
arXiv: https://arxiv.org/abs/2610.04210v1

Large language models (LLMs) receive each user message as plain text, even when it combines text from different sources. For example, a user may paste text into a prompt and keep typing a comment directly below it. We study absorption: a phenomenon where the model treats a trailing user comment as part of the pasted text, returning it inside the edited text. This happens even though the user did not intend the comment to become part of that text. Existing instruction-data separation benchmarks tell the model which text is instruction and which is data, then test whether it obeys that separation. They do not test harmless user speech following an unmarked paste. We introduce SEAM, a controlled benchmark of 300 editing examples. Each example is tested under six matched conditions that vary how the boundary between pasted text and later user speech is expressed. Across 20 models, absorption at a bare newline ranges from 7.7% to 66.7%. Adding a blank line does not significantly reduce absorption in any model, while boundary markers reduce it in 19 of 20 models. Comments that fit the pasted text, such as a code comment typed after code, are absorbed significantly more often in 17 of 20 models. Models often fail to separate pasted material from later user speech, and explicit boundaries reduce but do not remove this failure.

---

### Fine-Tuning VLM for Enhancing AI's Spatial Intelligence: Understanding 3D and 2D Rotations

Authors: Uttamasha Monjoree, Wei Yan
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04206v2

Spatial intelligence is a fundamental skill in multiple domains, such as Science, Technology, Engineering, and Mathematics (STEM), Medicine, Architecture, and Construction. Recent studies indicate that Vision-Language Models (VLMs) still face limitations in spatial reasoning, which inhibits artificial intelligence (AI) from performing practical spatial tasks. Using multiple object-rotation datasets developed for training and evaluation, our experiments demonstrated promising improvements in both 2D and 3D rotation detection. Fine-tuned Google DeepMind-built Gemma-4 mixture-of-experts (MoE) models significantly outperformed fine-tuned Gemma-4 generalist models in predicting rotations defined by both their axes and angles. Fine-tuning also substantially improved angle estimation for 2D representation without requiring an explicit coordinate system. Furthermore, identifiable objects did not improve angle-detection accuracy; instead, objects with prominent linear features showed improved performance.

---

### Clean: Second-order LLM Training at Linear Memory Cost via Nystr{ö}m Sketching

Authors: Beheshteh T. Rakhshan, Sahar Rajabi, Maziar Sargordi, Shikai Fang, Guillaume Rabusseau, Sirisha Rambhatla
Published: 2026-10-03
Categories: cs.LG, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2610.04204v3

Training large language models (LLMs) entails a fundamental trade-off: memory-efficient optimizers such as Adam discard cross-parameter curvature, whereas full-curvature methods such as SOAP can accelerate convergence at prohibitive memory costs. We introduce Clean, a memory-efficient and full-curvature optimizer designed to resolve this bottleneck. Clean leverages the randomized Nystrom method to accurately approximate the left and right preconditioners in SOAP, and to reduce the optimizer's memory complexity from quadratic to linear in terms of model dimensions. We subsequently reintegrate the off-subspace components to capture curvature information beyond the low-rank approximation, preserving rich curvature at minimal memory cost. We further propose Q-Clean, a low-precision variant that aggressively compresses optimizer states. Q-Clean reduces optimizer memory consumption by over 50% compared to Muon when pre-training a LLaMA-1.3B architecture, all while maintaining strong and competitive predictive performance. Notably, Clean operates with a smaller optimizer-state footprint than standard AdamW while reaching AdamW's final performance 26% faster in wall-clock time. Furthermore, our methods uniquely enable the pre-training of a 13B-parameter model on a single 80GB GPU, providing a scalable, efficient, and accessible approach to large-scale model optimization.

---

### Sparse-GS2Mesh: 3D Gaussian Splatting Guided by Novel Stereo Views and 2DGS for Sparse View Surface Reconstruction

Authors: Younghyun Noh, Minje Kim, Tae-Kyun Kim
Published: 2026-10-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.04203v3

Surface reconstruction under sparse-view settings remains challenging due to limited geometric cues. Volume rendering methods based on signed distance functions often produce over-smoothed surfaces, while 3D Gaussian Splatting (3DGS), though time-efficient, suffers from incomplete geometry due to the lack of reliable depth supervision and the limitation of being optimized only from given input views. In this paper, we present Sparse-GS2Mesh, a stereo-aware framework for surface reconstruction from sparse views. While 3DGS and stereo matching have been leveraged for surface reconstruction under dense view settings, we extend them to operate effectively under sparse view conditions by first initializing 3DGS using epipolar depth priors to mitigate the 3DGS overfitting problem, followed by our three key components: (I) adaptive baseline selection, (II) fine-tuning with a stereo matching network, and (III) 2D/3D co-regularized fine-tuning. Given a warmed-up 3DGS initialized with epipolar depth, the adaptive baseline selection automatically determines a baseline to synthesize for each sparse view. We then fine-tune 3DGS by backpropagating depth-refining gradients from the stereo matching network, effectively specializing the 3DGS for stereo matching. The 2D/3D co-regularization further helps obtain stable reconstruction, addressing weak geometric cues in close stereo views. Sparse-GS2Mesh achieves a 15\% improvement over state-of-the-art methods in little-overlap settings and comparable results in large-overlap settings. Project page: https://yhnoh8623.github.io/projects/s-gs2mesh/.

---

### CellSplat4D: PSF-Aware 4D Gaussian Splatting for Sparse Robotic Live-Cell Imaging

Authors: Yingda Tao, Guoyu Lu
Published: 2026-10-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.04199v1

A robotic microscope watching living cells cannot afford to look as often as it would like. Every volume it acquires costs photons the specimen does not get back, and time owed to other wells. What such a platform exists to produce is a record of individual cells through time: which cell is which from one volume to the next, and which cell divided into which two. Sampling sparsely breaks that record exactly where it matters, and the fault lies in the acquisition schedule rather than in the analysis software. We fill the gaps by reconstructing them, fitting a 4D Gaussian model to whatever volumes the hardware could afford. The model is a cloud of light-emitting blobs, each carrying a position, a shape and a lifetime. Being continuous in time, it renders any missing volume on demand, decoupling how often the robot analyses from how often it can afford to look. The microscope's point-spread function is measured from the data rather than inherited from acquisition metadata or left to the optimizer, because metadata inflates it and the optimizer cannot recover it at all: a wider blur around a smaller blob fits the images equally well. Each blob's lifetime is stored in frames rather than as a fraction of the recording, so that it denotes a fixed duration on any sequence. Unmeasured timesteps are supervised at coarse scale by a 3D U-Net that predicts the intermediate volume directly and estimates no motion field, since a dividing cell becomes two and no motion describes that. On two Cell Tracking Challenge sequences, a C. elegans embryo and a Chinese Hamster Ovarian (CHO), with fidelity scored per cell nucleus, our reconstruction holds the highest nucleus fidelity at every distance from an acquired frame, has the flattest decay across the gap, and best recovers focal planes it was never shown with graceful degradation across the gap.

---

### ALoDLM: Adaptively Looped Diffusion Language Models

Authors: Liancheng Fang, Zhuowei Li, Youngeun Kim, Tianchen Zhao, Rajat Koner, Jiaye Wu, Linghan Xu, Xuanbai Chen, Xiang Xu, Zheng Zhang, Jakub Zablocki, Nishant Sankaran et al.
Published: 2026-10-03
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2610.04198v1

Diffusion language models (DLMs) enable fast generation by predicting multiple tokens in parallel, but their practical adoption remains limited by a persistent quality gap relative to comparably sized autoregressive (AR) models. We attribute this gap to a computation-difficulty mismatch: within a partially observed sequence, some unknown tokens are easy to predict, while others require substantially more computation. Existing DLMs nevertheless apply uniform computational depth to all unknown positions at each denoising step. We introduce ALoDLM, which replaces uniform computation with token-adaptive latent recurrence. At each denoising step, ALoDLM iteratively refines latent representations and allocates computation according to token difficulty. Tokens ready to commit are fed back as discrete context, while unresolved tokens retain and further refine their latent states through additional recurrent passes. To learn token prediction and computation allocation jointly, we formulate token-wise computation schedules as latent variables and derive a conditional negative evidence lower bound (NELBO). We train ALoDLM at 1.7B and 8B parameter scales. Across eleven benchmarks, ALoDLM outperforms all evaluated DLMs and the corresponding AR baselines in average benchmark score at both scales. ALoDLM also retains fast parallel decoding, yielding a strong quality-efficiency trade-off among evaluated autoregressive and diffusion models under optimized inference engines.

---

### Asynchrony Is Nearly Free for Evolution Strategies on Long-Horizon LLM Agentic Tasks

Authors: William Hoy, Jingxuan Fan, Nurcin Celik, Xu Pan
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04196v2

Synchronous policy updates can limit the efficiency of post-training for long-horizon LLM agents. When trajectories span multiple interaction turns, completion times can vary substantially, and synchronous update barriers leave faster workers idling, wasting compute resources. Asynchronous reinforcement learning addresses this inefficiency but introduces policy lag and off-policy optimization. Evolution strategies (ES) are a backpropagation-free alternative to gradient-based training, yet require many perturbed-policy evaluations and have remained largely synchronous in current LLM post-training. We introduce bounded-staleness asynchronous ES and investigate its training behavior under stale fitness measurements. On WebShop, we systematically vary the lag and fraction of stale fitness measurements and additionally introduce two stress tests to probe extreme staleness cases. Natural asynchrony closely tracks synchronous ES, and degradation in evaluation accuracy remains small even under extreme staleness, including when all fitness measurements are stale. These findings suggest that ES can tolerate substantial policy staleness without explicit off-policy correction, supporting its use in asynchronous post-training for long-horizon LLM agents.

---

### MemLeak: Cross-User Semantic Leakage in Multi-Tenant AI Agent Memory

Authors: Priyanka Mudgal, Kai Zhao, Guilin Zhang, Andy Olsen, Ezekiel Miller, Xu Chu, Aletta Johanna Blanken
Published: 2026-10-03
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2610.04195v1

Personal AI agents in enterprise multi-tenant deployments share a common vector store for long-term memory. Shared embedding spaces create a surface for cross-user memory leakage: a user's query can retrieve semantically adjacent memories belonging to another user through ordinary cosine-similarity retrieval, without any exploit. We formalize this as cross-user admissibility failure and evaluate it across six experiments, plus follow-up ablations, under both sparse (TF-IDF) and production-faithful (MiniLM-L6-v2) retrieval. Non-adversarial, incidental leakage reaches 70--100\% under pooled {same-team} retrieval; adversarially crafted memories achieve 90--100\% top-$k$ placement, exceeding weaker keyword-based attacker baselines, with score lifts of $+0.416$ to $+0.511$ under production-faithful dense retrieval (Config B); and end-to-end response contamination reaches 5.00/5 under a production retrieval path and 4.67/5 with Claude Sonnet~4.5, with contaminated responses often scoring as helpful or more helpful than clean ones, a gap validated against human judgment. Among three architectural mitigations, only hard post-retrieval ownership gating consistently restores the clean baseline (1.00/5) across {two generation models, at a measured latency overhead of roughly 1.4~ms per query.

---

### Agentic AI with Structured CoT for Enhancing AI's Spatial Intelligence: Visualization and Reasoning of Rotation

Authors: Uttamasha Monjoree, Wei Yan
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04188v3

Recent studies show that artificial intelligence (AI) with language and vision capabilities still experiences limitations in spatial reasoning. In this paper, we have studied the spatial capabilities of advanced generative AI to understand the rotations of objects in 3D space, utilizing AI's image processing and language processing features. We trained and examined the spatial intelligence of a generative Agentic AI model (GPT-5.6) to understand the spatial rotation process with rotation diagrams based on the revised Purdue Spatial Visualization Test: Visualization of Rotations (Revised PSVT:R). We improvised the Revised PSVT:R by superimposing additional graphical and contextual features to evaluate how different Chain-of-Thought (CoT) reasoning strategies influence model performance. The results indicate that structured CoT reasoning improves the spatial reasoning performance of the base GPT-5.6 model in both datasets (PSVT:R and PSVT:R with coordinate system). We used three CoT approaches - (1) Structured CoT, (2) few-shot Structured CoT, and Structured CoT with Self-optimized Prompt. The three CoT approaches evaluated in this study showed no significant performance difference. Results showed that combining structured CoT reasoning with relevant contextual information leads to considerable improvements in VLM performance on 3D rotation tasks, demonstrating the potential of agentic AI for more effective spatial reasoning. However, when contextual information is removed, structured CoT reasoning alone provides limited improvement, and the models continue to exhibit notable difficulties in understanding spatial transformations. These findings suggest that effective spatial reasoning in VLMs relies on the integration of visual, textual, and reasoning-based information in future agentic AI systems for spatial intelligence.

---

### Referring Multi-Object Tracking in Moving-Camera Videos via Global Motion Compensation

Authors: Hsin-Chen Pai, Jyun-Kai Wang, Yi-Cheng Peng, Wei-Ta Chu
Published: 2026-10-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.04185v1

Referring multi-object tracking (RMOT) takes a video and a language expression as input and tracks all referred objects. Many tracking requirements involve how an object moves rather than how it appears. However, in a video captured by a moving camera, a parked vehicle may appear to move, while a moving vehicle may show little displacement. Existing RMOT methods relate motion with text but do not explicitly remove camera-induced motion. In this paper, we propose extracting residual motion across frames by estimating camera motion in driver-view videos and compare motion characteristics with the query expression. We consider the motion-matching extent and integrate it with the RMOT method's prediction result through late fusion. In the evaluation, we verify the performance gain of taking the motion compensation module as a plug-in across different RMOT hosts.

---

### EvalResearchBench: Can AI Agents Design Their Own Evaluations?

Authors: Yaolun Zhang, Tianyi Xu, Yujie Zhao, Jishen Zhao, Qingyun Wu, Huazheng Wang
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04184v1

Recursive self-improvement (RSI) relies on evaluation feedback to assess progress and guide further research, yet repeatedly running complex benchmarks is costly and slows iteration. Human experts reduce this cost by selecting benchmark subsets or designing compact suites. We ask whether AI agents can automate this design process and introduce EvalResearchBench (ERB), a benchmark for autonomous evaluation research. Given target materials, development references, candidate APIs, and fixed time and API budgets, an agent called the researcher selects or synthesizes tasks, implements graders, and revises them in pilot tests before freezing an executable evaluator for coding, co-work, and reasoning. We study 9 researchers and 13 candidate models and compare each frozen evaluator with 14 target benchmarks on score concordance and pairwise agreement. The best evaluators order about 75\% of candidate pairs as the targets do, below the 91\% ceiling set by disagreements among the targets. No researcher leads on every metric, and the best evaluator on development targets is not the best on sealed targets hidden from the researcher. A human-designed sample of public tasks remains a strong baseline, and the evaluator with the lowest recorded execution cost attains the highest pairwise agreement. Agents repair tasks and graders through pilot feedback, yet their evaluators can still truncate answers, exhaust the evaluation budget, or let a few questions dominate a domain score.

---

### Language Model Activations Inhabit Privileged Error-Correcting Basins

Authors: Matthew Finlayson, Francisco Pernice, Eric Todd, Amir Zur, Daniel Wurgaft, Fenil R. Doshi, Vasudev Shyam, Matt Feiszli, Satchel Grant, Lucius Bushnaq, Tal Haklay, Usha Bhalla et al.
Published: 2026-10-03
Categories: cs.AI, cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2610.04183v1

Language models exhibit remarkable robustness, continuing to produce coherent text even when their activations are perturbed by interventions like linear steering. We hypothesize that this robustness is a result of passive dynamics, i.e., constraining mechanisms in the forward pass that funnel activations toward "good" regions that produce coherent outputs. To investigate these hypothesized error-correcting mechanisms, we probe the geometry of language model activation space by observing the action of model layers on low-dimensional curves. In doing so, we discover that model activations occur within a cluster of distinct attracting basins, which differentiate natural activations geometrically from distributionally similar synthetic activations. Applying this lens to language model steering, we observe feature-specific basins along semantic steering directions, and find that steering moves activations between these basins. To demonstrate the active role of this geometry in neural computation, we show that adaptively modulating steering strength to transport activations across basins improves inter-language steering, significantly increasing the probability of sampling tokens from the target language compared to fixed-strength steering. Our findings establish analysis of activation space geometry as a promising approach to interpreting and controlling language models.

---

### SHarP: Saliency-based Pruning of Agent Harnesses

Authors: Xinyi Gao, Qiucheng Wu, Kaizhi Qian, Handong Zhao, Shiyu Chang, Yang Zhang
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04178v1

Agent harnesses are systems that coordinate model calls, tool use, and task execution to help large language models complete complex tasks. To meet task requirements and address failures, these systems are often iteratively refined by amending and patching their instructions, tools, and workflows, continuously increasing harness complexity. It is therefore unclear whether some resulting harness modules are redundant, introducing substantial token overhead with little, if any, performance gain. Inspired by neural network pruning, in this paper, we study harness pruning as a means of striking a better balance between task performance and token cost. We propose SHarP (Saliency-based Harness Pruning), a simple yet effective pruning strategy based on the saliency of each harness module with respect to performance and efficiency. Specifically, we first identify tools, instructions, and supporting mechanisms as components that can be individually disabled. We then estimate the saliency of each module by ablating it and assessing its task performance and token cost relative to the full set of single-module ablations. Modules with the smallest contribution to performance or largest computational overhead are subsequently pruned. Our evaluation across various harnesses on held-out validation sets reveals a surprising finding: most harnesses that we studied are highly redundant and can maintain comparable performance and efficiency even after a substantial portion of their modules are pruned. Our pruning approach and empirical findings provide new perspectives on agent harness design and optimization.

---

### DimSteer: Steering LLM Authoring with Automatically Discovered Stylistic Controls

Authors: Ajit Mallavarapu, Ziwei Gu
Published: 2026-10-03
Categories: cs.HC, cs.CL
arXiv: https://arxiv.org/abs/2610.04174v1

Large language model writing interfaces often make users steer outputs by repeatedly articulating desired changes in natural language. Yet writers may recognize useful stylistic directions only after seeing alternatives, making revision recall-heavy. We present DimSteer, an authoring interface that samples prompt-local completions, discovers high-variance activation-space axes of variation, labels them, and exposes them as sliders with pole previews, diff comparison, and reset controls. Users can manipulate discovered dimensions, reducing the need to reformulate prompts for each stylistic adjustment. In a within-subjects study with 16 participants against a matched prompt-only baseline, DimSteer reduced mental demand, effort, and frustration while preserving comparable perceived success. Participants valued the surfaced dimensions, yet 15 of 16 disagreed that they would have thought to request the same changes in a prompt. Results suggest prompt-local controls can shift LLM authoring from recall-based prompting toward recognition-based exploration and direct manipulation, while preserving prompting for open-ended edits.

---

### Risk-Calibrated Proposal Transport for Finite-Particle Diffusion Steering

Authors: Ziseok Lee, Jaehyeon Kim, Seungwon Kim, Seunghyun Moon, Haneul Choi, Wooyeol Lee, Donghyun Koh, Minhyeong Lee, Kyungsu Kim
Published: 2026-10-03
Categories: stat.ML, cs.AI
arXiv: https://arxiv.org/abs/2610.04171v1

Inference-time steering combines pretrained diffusion experts or rewards without retraining by changing the dynamics that transport noise to data. Feynman-Kac correction compensates for proposal mismatch through importance-weighted sequential Monte Carlo (SMC), whose finite-particle behavior depends on the proposal. Variance-controlling guidance (VCG) improves that proposal by fitting a linear drift correction to minimize empirical log-weight-rate variance. Although its population optimum cannot worsen residual variance, finite-particle VCG can nearly eliminate its fitting residual while increasing residual risk on new states by orders of magnitude. The resulting update can degrade unweighted generation or accelerate particle collapse. We show that the centered Feynman-Kac rate is the normalized transport residual and that expected out-of-fit benefit is exactly population headroom minus coefficient-estimation penalty. Under regularity assumptions, a Wasserstein analysis bounds the unweighted proposal's terminal error using this residual. These results motivate Risk-Calibrated Proposal Transport (RCPT), which uses deletion leave-one-out residuals to calibrate the retained fraction of the VCG update, adding no model calls and only small linear-algebra overhead. Experiments on 2D checker distributions, scaffold decoration, molecular property optimization, and class-conditional CIFAR-10 generation demonstrate recovery from harmful fitted updates. Across molecular and image domains, RCPT mitigates harmful fitted updates and improves a broad range of terminal metrics relative to uncalibrated VCG.

---

### Language Model Fingerprinting Requires Rethinking Watermark Teachers

Authors: Jeongyeon Hwang, Anshul Nasery, Sewoong Oh, Jungseul Ok
Published: 2026-10-03
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2610.04169v1

LLM fingerprinting via watermark distillation embeds a statistical watermark signal into model weights, enabling model owners to identify their models behind black-box APIs. Revisiting a recent protocol, we find that its utility evaluation understates text quality degradation in open-ended generation, favoring overly strong watermark teachers. Weakening the watermark improves text quality but sacrifices detectability. To move beyond this trade-off, we rethink whether text watermarks designed for verifying generated text are suitable distillation teachers for model fingerprinting. Such watermarks are typically designed to remain detectable from an individual output, limiting how sparse the watermark signal can be. In contrast, fingerprint verification can aggregate signal across queries, making sparser watermark signals viable. This raises a key question: where should the sparse signal be placed? We analyze signal placement through token surprisal and show that, even at comparable watermark strength, different placements can target tokens with different plausibility under the base model. This motivates near-tie restriction, which uses top-1-relative logit gaps to restrict the watermark bias to tokens close to the base model's top prediction. Across multiple models, near-tie improves detection--quality frontiers under deployment changes, preserves higher text quality across query budgets, and further improves existing watermarking schemes when combined with them.

---

### Agentic Cognitive Depth: Operational Criteria for Evaluating LLM Agents

Authors: Nijesh Upreti, Chris Sypherd, Vaishak Belle
Published: 2026-10-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.04168v1

Agentic large language model (LLM) systems are commonly implemented as an LLM in a loop with Planning, Memory, Tools, and Control Flow. This application-focused view connects agentic LLM research with deployable systems and leaves open how such systems should be evaluated beyond end-to-end task success. Building on this view, we define agentic cognitive depth as a trajectory-level profile across five operational criteria. The profile contains context sensitivity ($C$), temporal continuity ($T$), multimodal coordination ($M$), adaptive interaction ($A$), and metacognitive monitoring ($Mc$). The first four criteria measure how well Control Flow, Memory, Tools, and Planning are used across a trajectory. The fifth measures whether the system monitors and regulates the full run. For each criterion, we give operational proxies and a perturbation procedure, then connect the profile to the agent's world model. We provide the structure needed to extend benchmarks such as GAIA, SWE-bench, WebArena, and TRIP-Bench with per-criterion diagnostics. Symbolic verifiers, structured memory, planner coupling, and tool constraints provide practical ways to build and test these capacities.

---

### Principled Top-$k$ Selection for Language Models with Hybrid Gradients

Authors: Xuchen Gong, Junfei Sun, Tian Li
Published: 2026-10-03
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2610.04162v1

Selecting the best $k$ items out of $m$ candidates is a critical component of modern large language model systems, such as document selection in Retrieval-Augmented Generation (RAG) and expert routing in Mixture-of-Experts (MoEs). However, training these selection modules remains challenging due to weak gradient signals and suboptimal exploration-exploitation tradeoffs. Furthermore, prior works often rely on heuristics, lacking principled objectives and approaches that explicitly model and solve the top-$k$ selection problem. In this work, we propose a principled objective for training selection modules, whose gradient naturally provides richer training signals in a hybrid form---containing a supervised-gradient component and a policy-gradient component. We show that the selection problem becomes harder as $m$ increases, and our algorithm converges at rate $O(1/\sqrt{T})$, with the optimal upper bound achieved by balancing between bias and variance. Practically, we apply our method to a set of tasks involving top-$k$ selection, including synthetic regression problems, RAG, and MoE systems, showing that our method outperforms the baselines in next-token prediction perplexity and QA accuracy.

---

### Diagnosis-Conditioned Spatial Gating and Decoder-Level Supervised Contrastive Learning for Radiology Report Generation

Authors: Md Mustafizur Rahman, Mylene C. Q. Farias
Published: 2026-10-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.04159v1

Radiology report generation models can produce fluent text while still containing finding-level inaccuracies. Diagnosis-driven methods improve generation by conditioning on predicted findings, but these predictions do not directly modify the visual patch features provided to the decoder, and global gating applies the same modulation across spatial locations. We introduce a Position-Aware Gate (PAG) that uses predicted finding representations to modulate visual patches spatially without region supervision. We also propose a Decoder-level Supervised Contrastive Loss (DSCL) that structures decoder representations using shared positive findings rather than instance identity. On MIMIC-CXR, PAG+DSCL improves clinical efficacy (CE) F1 from 0.484 to 0.502 over a matched global-gate reference, while PAG and DSCL individually reach 0.491 and 0.495. Without additional fine-tuning, the combined model achieves 0.226 CE F1 on IU X-Ray, compared with 0.211 reported by PromptMRG.

---

### How RL Reshapes LLM Reasoning: Transferability, Coverage, and Scaling Laws

Authors: Ziheng Cheng, Yixiao Huang, Hanlin Zhu, Somayeh Sojoudi
Published: 2026-10-03
Categories: cs.LG, cs.AI, stat.ML
arXiv: https://arxiv.org/abs/2610.04158v1

Recent studies on reinforcement learning (RL) report seemingly conflicting evidence about large language model (LLM) reasoning. Training on mathematics can improve performance in other domains, yet gains in Pass@1 can coincide with lower Pass@$N$ than the base model. This raises a fundamental question: does RL expand an LLM's reasoning boundary, or merely reweight its existing reasoning space? We revisit these phenomena across Qwen and Gemma model families, showing both cross-domain gains and forgetting, while coverage at large sampling budgets increases on some tasks and decreases on others. Detailed analysis of solution traces before and after RL indicates a shift in the reasoning strategies the model employs, motivating a two-stage autoregressive policy model that separates \emph{strategy selection} from problem-specific execution. Within this framework, we prove how RL's implicit bias reshapes strategy preferences, allowing gains on some tasks while suppressing strategies required by others. This mechanism can also broaden or narrow coverage at a given sampling budget even without expanding strategy support. We further provide theoretical justifications for log-sigmoid and log-linear scaling laws in RL compute, and evaluate their predictive power. Together, these results connect changes in strategy selection to cross-domain transfer, reasoning coverage, and compute scaling.

---

### Mitigating Over-squashing without Rewiring: A Sheaf Effective Resistance Perspective

Authors: André Ribeiro, Germano Barcelos, Amauri H. Souza, Diego Mesquita, Ana Luiza Tenório
Published: 2026-10-03
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2610.04157v1

Graph Neural Networks (GNNs) often struggle to capture long-range dependencies due to over-squashing -- a phenomenon in which the repeated compression of node embeddings into finite-size messages causes representations to collapse. Over-squashing is most often diagnosed as a property of the graph topology, with effective resistance serving as a principled measure of the bottleneck. We provide a complementary view on the matter: building on cellular sheaves, we introduce sheaf effective resistance, a generalization of effective resistance that depends on the sheaf attached to the graph, and we prove that for flat vector bundles, the over-squashing sensitivity in the Jacobian sense is upper bounded by a quantity related to the sheaf effective resistance between the nodes. The bottleneck thus need not lie in the graph itself: it can be relocated, and reduced, by adjusting the sheaf. We instantiate this idea in FlatNSD, a simple message-passing variant of Neural Sheaf Diffusion, and show that it implicitly learns to modulate total sheaf effective resistance, performing well on benchmarks designed to stress over-squashing without altering the original graph topology.

---

### Less Decoder is More Encoder: Geometric Representation Learning from Novel View Synthesis

Authors: Keerthi Kaashyap, Dennis Anthony, Akshay Krishnan, Nhi Ngoc Nguyen, Jeremy Collins, James Hays, Shreyas Kousik, Animesh Garg
Published: 2026-10-02
Categories: cs.CV, cs.AI, cs.RO
arXiv: https://arxiv.org/abs/2610.03717v1

This paper examines the role of Novel View Synthesis (NVS) in geometric representation learning. In principle, NVS should reason about 3D scene structure, thereby enabling transferable multi-view geometric representations. Yet, existing encoder-based NVS methods yield poor representations. This is not because of a lack of supervisory signal, but rather due to inconspicuous architectural choices: \textit{spatially expressive decoders} that dilute representational capabilities of the scene encoder, and \textit{low-level pixel-space targets} that hinder feature learning. We present SNAP, a self-supervised encoder-decoder transformer that addresses both through a pose-conditioned local decoder and a latent-space reconstruction objective. SNAP is task agnostic, and we show that it is competitive with special-purpose geometry-supervised methods. SNAP also performs competitively against self-supervised representations across five tasks: visual localization, pose estimation, point correspondence, depth estimation, and robot manipulation. Remarkably, SNAP's patch features exhibit emergent viewpoint invariance that approaches heavily supervised models despite lower compute and data budgets. Under camera shifts where standard 2D representations collapse, SNAP degrades more gracefully, revealing that restricting decoder expressivity actively prevents the suppression of transferable geometric structure. https://snap-nvs.github.io

---

### MoSE3: Learning World-Space SE(3) at Every Pixel

Authors: Jiahuan Cheng, Zhiyi Li, Tian Xia, Ruojin Cai, Yilun Du, Qianqian Wang
Published: 2026-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.03716v1

Dense 3D point tracking has been a prominent paradigm for modeling motion in dynamic scenes, but a point track is just a 3-DoF translation curve per pixel: it captures where pixels go, not the rotation of the underlying part, nor which pixels move together as one body. We propose MoSE3, the first feed-forward model that predicts dense SE(3) motion from monocular RGB video, producing full 6-DoF rigid transforms at every pixel in world space. Per-pixel SE(3) motion offers a richer view of how a scene moves: rotation, translation, and grouping all at once. Directly predicting SE(3) is challenging: rotations lie on a curved manifold that is ill-suited to Euclidean regression, and annotations for SE(3) are particularly difficult to acquire. To address these challenges, MoSE3 predicts per-pixel SE(3) through two jointly learned intermediates, 3D point tracks and rigidity embeddings, and recovers SE(3) by differentiably fitting transforms within each soft rigid cluster, enabling end-to-end prediction and supervision. To close the data gap, we introduce Art-Kubric, a large-scale synthetic dataset with dense SE(3) and rigidity labels for articulated objects with rich physical interactions. MoSE3 achieves state-of-the-art SE(3) estimation at pixel, part, and object levels on both rigid and articulated benchmarks, and state-of-the-art average 3D point tracking accuracy across three datasets, while showing strong generalization to real-world videos despite being trained solely on synthetic motion data.

---

### 4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes

Authors: Ruihong Shen, Žiga Kovačič, Peter Kulits, Xingrui Wang, Zizhang Li, Joshua B. Tenenbaum, Alan Yuille, Jieneng Chen, Jiajun Wu
Published: 2026-10-02
Categories: cs.CV, cs.AI, cs.GR
arXiv: https://arxiv.org/abs/2610.03715v1

We introduce 4DCodeBench, a benchmark for 4D inverse graphics through code generation, in which agents reconstruct dynamic scenes from video as executable graphics programs. To accomplish this, agents must translate visual observations into compact representations of scene structure and dynamics, by implementing abstractions such as physical simulations to reproduce complex behavior. To evaluate this capability, we curate a set of real-world videos and construct synthetic scenes spanning diverse physical phenomena, including deformation, fluid flow, and fracture. We perform extensive benchmarking of frontier models, finding that strong static reconstruction capabilities do not yet translate into reliable reconstruction of complex dynamics. 4DCodeBench provides a testbed for tracking progress toward agents that can interpret the dynamics of the world through code. Our benchmark is available at https://github.com/4DCodeBench/4DCodeBench

---

### What Should World Models Forget? Stratified Retention for Continual Adaptation

Authors: Nishit Anand, Ramani Duraiswami, Dinesh Manocha
Published: 2026-10-02
Categories: cs.LG, cs.AI, cs.CV, eess.IV, eess.SP
arXiv: https://arxiv.org/abs/2610.03713v1

Continual learning treats degradation on previously seen data as evidence of failure, a convention inherited from settings with a stationary prediction target, where a correct label remains correct indefinitely. World models do not satisfy this condition. Their prediction target is the environment, which changes, so knowledge that was accurate when acquired may later become false, and discarding it is required behavior rather than a defect. Non-stationary ground truth is well studied in the concept drift literature and in the temporal factuality of language models, but has not been formulated for world models, which are distinctive in that they also encode knowledge that must never be revised. We argue that continual world models require retention stratified by invariance timescale, separating invariants such as physics and object permanence, which must never be revised, from instance-level facts that should be revised as soon as the environment changes. Standard forgetting metrics cannot distinguish a world model that has correctly revised outdated knowledge from one that has suffered catastrophic forgetting, and consequently rank a frozen model highest, while existing physical-reasoning benchmarks evaluate only frozen checkpoints. We propose differential retention, which reports invariant regression testing across the adaptation stream jointly with revision latency, without aggregation.

---

### RNADyn: A Benchmark for Generating and Understanding RNA Dynamics

Authors: Yiming Huang, Lennart Bastian, Hanqun Cao, Luis Vollmers, Tolga Birdal
Published: 2026-10-02
Categories: cs.LG, physics.bio-ph, q-bio.BM
arXiv: https://arxiv.org/abs/2610.03712v1

Ribonucleic acid (RNA) functions through conformational changes that are not fully captured by static structures. However, large-scale standardized RNA dynamics data remain limited, and existing approaches typically treat trajectory generation and dynamics understanding as separate objectives. Here, we introduce RNADynBench, a standardized RNA molecular dynamics (MD) benchmark with 2585 quality-controlled 100-ns all-atom trajectories and leakage-controlled splits. Building on RNADynBench, we develop RNADynNet, a unified model for RNA dynamics learning that uses a shared backbone for both trajectory generation and dynamics fingerprint extraction from a single conformer. It combines coordinate denoising, single-frame-to-trajectory alignment, and physical grounding to connect all-atom trajectory generation with dynamics representation learning. Physical grounding improves both generated dynamics and the physical information recoverable from these fingerprints. Across both test sets, including the high-flexibility challenge set, the generated trajectories achieve RMSF correlations of 0.875 and 0.766, while single-conformer predictions show comparable agreement with MD-derived dynamics. RNADynBench and RNADynNet together establish a benchmark and unified modeling framework for generating and understanding RNA dynamics.

---

### EyeRobot 2.0: Active Gaze for Precise Manipulation without Wrist Cameras

Authors: Kush Hari, Justin Kerr, Nidhya Shivakumar, Samarth Mahapatra, Carmelo Sferrazza, Jiahui Lei, Jitendra Malik, C. Karen Liu, Ken Goldberg, Angjoo Kanazawa
Published: 2026-10-02
Categories: cs.RO, cs.AI
arXiv: https://arxiv.org/abs/2610.03710v1

Inspired by human vision, we introduce a framework using active gaze to enable fine-grained bimanual manipulation with only a single stereo camera. EyeRobot 2.0 physically attends to a 3D fixation point in the scene by swiveling two eye viewpoints to center their gaze on it. The resulting images are processed foveally by allocating more visual tokens to the image centers, focusing computation on task-relevant features. Such Active Visual Fixation (AVF) requires carefully coordinated gaze during task execution, which we accomplish hierarchically by first training a low-level gaze servoing policy conditioned on a goal object, then training a target selector which emits fixation goals based on task progress. Both modules are trained with RL on real-world data: the first is trained with a dense geometric reward and the second co-trains with the BC gripper policy which allows it to discover fixation sequences that can resemble a human's fixation sequence while performing the task. EyeRobot 2.0 further takes advantage of fixation by canonicalizing gripper information into a fixation-relative SE(3) frame, which compacts the size of the action distribution to learn. We collect teleoperation data for 7 real-world and 6 simulated tasks, and conduct over 1000 physical and 1800 simulated robot trials comparing EyeRobot 2.0 against passive stereo and ego + wrist camera policies trained on the same data. Removing wrist cameras is costly for standard policies: with only passive stereo, real-world success drops from 52% to 27%. EyeRobot 2.0 closes this gap with only stereo, outperforming passive stereo by 40% in real and 20% in sim. It matches ego + wrist policies when their wrist views are clear (69% vs. 64%), and more than doubles their success when grasped objects occlude the wrist cameras (48% vs. 22%)

---

### From Mixing to Tearing: Graph Decomposition in Decentralized Optimization via Message Passing

Authors: Kuangyu Ding, Gesualdo Scutari
Published: 2026-10-02
Categories: math.OC, cs.LG
arXiv: https://arxiv.org/abs/2610.03709v1

We study the minimization of sums of smooth strongly convex functions over undirected graphs, with each function held by one agent and communication restricted to neighbors in the graph. Existing decentralized methods, whether based on gossip or on routing over spanning trees, typically use the network to mix or aggregate information to enable {\it prescribed} local optimization updates. What this communication-centered viewpoint lacks is a general framework that uses graph structure to {\it jointly} design the optimization subproblems and the cooperative computation and communication through which agents solve them cooperatively. We develop such a framework from first principles, jointly designing the linear representation of agreement constraints, the blocks of the resulting dual variables (jointly optimized), and connected cluster of agents that cooperatively solve each block subproblem over the assigned subgraph. GATE (Graph-Tearing message passing) is a first instance of this framework: one variable per edge and tree blocks. At each iteration, agents update their assigned edge variables by minimizing the sum of the two endpoint cost-to-go messages and relaxing the result. The messages are updated through local minimizations following the tree recursion. To reduce per-iteration computational and communication costs, we develop GATE-S, a surrogate variant using tractable local models and lightweight message parametrizations. We establish linear convergence with a rate explicit in the interplay among function regularity, network topology, and the chosen partition, revealing the effects of graph decomposition. Numerical experiments are conducted to validate the theoretical results and evaluate the efficiency of our algorithms.

---

### LESSER: Post-Training Data Selection with Output-Layer Gradients

Authors: Lyuxin David Zhang, Eric Wong, Surbhi Goel, Anton Xue
Published: 2026-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.03702v1

The choice of post-training data for large language models substantially affects downstream performance. Gradient-based data selection is a popular approach that ranks training data by how well their gradients align with those of a small validation set. However, ranking with full-parameter gradients requires an expensive backward pass on every sample, making computation intractable for large candidate pools. This raises a natural question: can we approximate full-gradient features at a fraction of the cost? Conveniently, we find that output-layer gradients suffice for effective data selection, yet require only the cheaper forward pass. We implement this as LESSER, a drop-in wrapper for selection methods that reduces the feature-extraction FLOP cost by $9.7\times$ for SFT and $3.0\times$ for RL benchmarks, while tracking full-gradient performance on downstream tasks. Empirically, we find that even when output-layer and full gradients rank individual samples differently, they select batches with aligned gradients.

---

### Decoding the Functional Roles of Register and High-Norm Patch Tokens in Vision Transformers

Authors: Neel Varma, Andrew Rufail, Dipika Khullar, Vasu Sharma
Published: 2026-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.03698v1

Self-supervised Vision Transformers (ViTs), such as DINOv2, learn rich visual representations, but the functions of their internal tokens remain poorly understood. Recent architectures introduce dedicated register tokens to reduce high-norm out- lier patch tokens that emerge in background re- gions, yet the semantic and functional roles of both token types have not been fully established. In this paper, we analyze these roles by training sparse autoencoders (SAEs) on register-token and outlier-token activations in DINOv2. Using an automated interpretability pipeline, UMAP clus- tering, and CLIP-space cross-checks, we find that register-token features are more strongly associ- ated with high-level semantic concepts. Outlier- token features, by contrast, are more often associ- ated with lower-level structural, background, and texture-dominant patterns. Causal ablations fur- ther reveal a substantial functional asymmetry: disrupting top-activating register-derived features produces a 48.17% drop in representation cosine similarity, whereas disrupting outlier-derived fea- tures produces only a 0.31% drop. Together, our results provide evidence for token specialization in self-supervised ViTs.

---

### Language Models that Play Chess and Explain Their Moves

Authors: Adithya Bhaskar, Jeffrey Cheng, Danqi Chen
Published: 2026-10-02
Categories: cs.CL
arXiv: https://arxiv.org/abs/2610.03695v1

Modern chess engines are silent experts: they play at a superhuman level, but do not offer explanations for their play. On the other hand, language models (LMs) can generate plausible-sounding explanations, but their weak playing strength limits the utility of their explanations. We introduce Queen, a 4B-parameter chess-language model that can explain its moves and plans while playing at the level of a typical Grandmaster. Our novel framework enables domain-specific reasoning through complementary components: an encoder-decoder architecture and an iterative distillation algorithm. This architecture integrates a silent expert chess encoder with an instruction-tuned LM through cross-attention, which we train via a question-answering curriculum to extract chess concepts from the encoder's representations. Building on this domain-adapted model, we iteratively improve its explanations with a natural-language analog of the Bellman update: the model analyzes the positions after its top candidate moves and consolidates them into an explanation of the current position, which is then distilled back into the model. Over seven iterations, our model gains over 900 Elo points (1782 to 2697), substantially surpassing all frontier models on both playing strength and puzzle accuracy, despite containing three orders of magnitude fewer parameters. Furthermore, LM-based evaluations show that our explanations are fluent and approach GPT-5.6-Sol (high) in coherence. The generality of our architecture and training procedure suggests a recipe for applying language models to domains where silent expert encoders are available, like games, robotics, and computer use.

---

### Transcriptome-informed multi-modal AI for predicting neoadjuvant therapy response from breast cancer biopsies

Authors: Jungkyu Park, Dhruva Biswas, Joseph Cappadona, Cerise Tang, Ken G. Zeng, Bartosz Machura, Chuwen Liu, Paolo Tarantino, Coral Omene, Francisco J. Esteva, Rohit Bhargava, Marcin Braun et al.
Published: 2026-10-02
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.03693v1

Scarcity of labeled data limits development of deep learning biomarkers in oncology. We develop a two-stage AI model predicting pathological complete response (pCR) to neoadjuvant therapy in breast cancer. The first stage learns the transcriptome from histopathology using 8,742 patients across 32 cancer types, corroborated by pathologist review and spatial agreement with measured expression. This simplifies the second stage to predicting pCR from inferred expression and clinical variables. Developed using 1,080 patients (five cohorts) and evaluated in 1,412 patients (nine cohorts), the model achieves a pooled AUROC of 0.79 (95% CI, 0.73-0.85), discriminating responders within molecular subtypes. It outperforms histopathological biomarkers, remaining stable across intratumoral sampling and with minimal biopsy tissue. Ablations show transcriptome-wide inference improves discrimination over clinical variables alone or one-stage pathology models, and robustness by avoiding genomic assays' gene selection constraints. These results indicate that biologically informed compression may generalize to data-sparse applications in precision oncology.

---

### FlowHMR: Physically Plausible Motion Capture from Video

Authors: Zhanke Wang, Chengfeng Zhao, Qing Shuai, Jingzhong Lin, Heng Li, Zeyu Ling, Yuxin Wen, Jing Li, Di Kang, Chunchao Guo, Linchao Bao
Published: 2026-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.03691v1

We present FlowHMR, a framework for recovering physically plausible global 3D human motion from monocular video. Previous learning-based methods typically regress human motion directly from video and train the network with geometric supervision. However, recovering human motion from monocular video is inherently ambiguous in depth, and direct regression tends to collapse toward an averaged solution. Moreover, the recovered motions are not guaranteed to be physically plausible, so physics-based tracking of them often fails. To address these challenges, we formulate video motion capture as a video-conditioned motion generation problem and first pretrain a flow matching model for this task. Given an input video, the pretrained model generates diverse motion candidates, but not all of them are faithful to the video or physically trackable. We therefore post-train the model using Group Relative Policy Optimization (GRPO) with two rewards. A fidelity reward encourages consistency with the input video. A tracking reward favors motions that a physics-based controller can track successfully. Together, these rewards shift the model's output preference, so the post-trained model stays faithful to the input video while producing more physically plausible motion. We further introduce Wild-4K, a large and diverse dataset of about 4K internet videos, for evaluating human motion recovery in the wild. Qualitative and quantitative experiments on Wild-4K show that our method outperforms state-of-the-art methods in overall motion fidelity and achieves a physical tracking success rate of 82.47%, compared with 62.82% for the strongest baseline, GVHMR.

---