# Research Papers — 2026-10 · bundle page 29 of 156

> Full-text markdown of papers #1401–1450 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/29
[← Previous bundle](https://aipod.dev/papers/archive/2026-10/page/28.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-10/page/30.md)

### Linear regression with probabilistic networks for network-linked data

Authors: Jingnan Zhang, Shiwen Ye, Danyang Huang
Published: 2026-10-13
Categories: stat.ME, stat.ML
arXiv: https://arxiv.org/abs/2610.18937v1

Network-linked data combine a network with covariates and responses observed for its nodes. Common regression approaches either model dependence among neighboring responses using the observed network or capture similarities associated with network structure without explicitly modeling response dependence. We propose a probabilistic network-linked data analysis framework that incorporates both features while accounting for randomness in network formation. The model uses normalized connection probabilities to weight peer responses and latent positions to capture network cohesion. A two-step likelihood procedure first estimates the network structure and then fits the response model. We establish theoretical guarantees for network recovery in sparse networks with heterogeneous node effects, and prove consistency and joint asymptotic normality of the regression estimators subject to suitable conditions. The analysis also clarifies when estimating the network preserves the classical convergence rate for spatial autoregressive models. Simulation studies demonstrate the estimation performance of the method and assess its inference accuracy. An analysis of diplomatic network data illustrates its practical application.

---

### Delay Complexity Dimension: Learning with Delayed and Censored Labels

Authors: Madhava Gaikwad
Published: 2026-10-13
Categories: stat.ML, cs.IT, cs.LG
arXiv: https://arxiv.org/abs/2610.18935v1

We study PAC learning of binary classifiers when labels arrive after random delays and are censored beyond an observation window $W$. How many samples does this observation model require? We propose a distribution-dependent answer, the Delay Complexity Dimension (DCD). The DCD combines the VC dimension with the exact label-information rate of the observation channel. Every learner needs $Ω(\mathrm{DCD}/\varepsilon^2)$ samples. An oracle pseudo-label ERM matches this rate on the instance family of the lower bound, up to constants and logarithmic factors. On that family the DCD is within a constant factor of the classical censored-ERM complexity, and whether the DCD rate is achievable where the two separate remains open. The theory also yields a quantization scheme under which delay information lowers the residual label variance, an expedition policy with a $(1-1/e)$ guarantee for its version-space objective, and a phenomenon with no analogue in the no-delay setting. Under a fixed collection horizon, a moderate window can strictly beat immediate labels. Experiments on synthetic data and the Criteo conversion logs support these predictions. A budget-matched test of a companion labeler-diversity prediction is reported as a negative result.

---

### An Initial Study of Physiological Markers for Deepfake Speech Detection

Authors: Vojtěch Staněk, Michal Frič, Kamil Malinka, Anton Firc
Published: 2026-10-13
Categories: cs.SD, cs.CR, cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2610.18934v1

Conventional deepfake speech detectors primarily target synthesis artifacts and often degrade under unseen attacks or recording conditions. Human speech, however, contains physiological markers associated with respiration, phonation, and temporal coordination arising from the speech-production system that synthesizers may not reproduce faithfully. This motivates the hypothesis that deepfake speech can be detected by identifying inconsistencies in these patterns, which we test using 36 handcrafted physiological markers organized into respiratory, vocal, and temporal categories. We evaluate the marker vectors with logistic regression and SVM classifiers on pooled English recordings. The best in-domain system achieves a 21.40% equal error rate (EER), while logistic regression achieves 32.77% on the out-of-domain dataset pair. Since the markers describe specific speech-production properties, the logistic-regression coefficients indicate whether higher marker values push the classifier toward bona fide or deepfake speech. These results provide initial evidence that physiological markers offer a useful, interpretable signal for deepfake speech detection.

---

### The Agentic Scientist: Orchestrating Coding Agents for Automated Scientific (Co-)Discovery

Authors: Robert Tjarko Lange, Luca Grillotti, Yutaro Yamada, Rujikorn Charakorn, David Ha
Published: 2026-10-13
Categories: cs.AI, cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2610.18933v1

Modern coding agents are fundamentally reshaping the role of human input for computational research: they write code, launch in silico experiments, analyze their results, and autonomously operate with ever-increasing time horizons. Despite these developments, fully autonomous research agents suffer from a limited ideation radius and myopic hill-climbing search dynamics. To bridge this gap, we introduce The Agentic Scientist, an end-to-end distributed pipeline for human researchers to conduct computational research alongside modern coding agents. Building on its predecessor, our system decomposes research into task drafting, idea generation, planning, sandboxed experimentation, result aggregation, baseline and ablation construction, post-hoc analysis, and report writing. Importantly, each step in the pipeline is powered by the next generation of coding agents and leverages agentic self-verification. The overall research process is driven by a meta-controller that manages parallel research agents and enforces rubric verification at each stage. This design enables the same orchestration layer to be reused across heterogeneous domains with different execution environments, agentic coding harnesses, and evaluation metrics. Our system can be used across a spectrum of autonomy: either as a co-scientist tool or to perform end-to-end scientific investigation. Within the co-scientist setting, we leveraged a development version of our system to submit a single research paper, which got accepted for publication at ALIFE 2026. We further apply The Agentic Scientist fully autonomously to open problem-solving tasks in machine learning, and demonstrate the system's capability of scientific discovery across different domains of computational research. Our ablations suggest that reliable scientific autonomy depends not only on base model capabilities, but also on the cross-agent orchestration layer.

---

### Why Static Skill Inconsistencies Do Not Equate to Runtime Risks: A Behavioral Analysis of LLM Agent Execution

Authors: Fengrui Fan, Huahao Wang, Jingyi Fang, Yutao Hu, Yueming Wu, Deqing Zou
Published: 2026-10-13
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.18927v1

LLM agent frameworks heavily rely on external skill assets to standardize tool usage, encode operational workflows, and supply executable runtime resources. Despite their widespread adoption, multi-component skill packages frequently suffer from cross-artifact inconsistencies, where documentation, scripts, and configuration files specify conflicting behavioral constraints. Although static auditing techniques can uncover such contractual discrepancies, static defect volume cannot reliably reflect real-world agent performance degradation. This gap leaves practitioners without rigorous evidence to assess the practical risk of flawed agent skills. This paper presents a systematic empirical study to demystify the runtime implications of skill inconsistencies. We conduct large-scale auditing over 575 community and benchmark skills, yielding 1,182 manually validated contract inconsistencies. We further build a controlled evaluation pipeline with 49 executable agent tasks and three representative LLMs across NoSkill, OriginalSkill, and FixedSkill experimental settings. Through fine-grained behavioral analysis of 441 paired execution trajectories, we quantify the actual performance impacts of resolved defects.

---

### Personalising Deformable Image Registration with Clinical Context and Deformation History for Adaptive Radiotherapy

Authors: Caiwen Jiang, Songlin Zhao, Yuzhen Ding, Jia Mi, Samir H. Patel, Alexandra Bennett, Keldon K. Lin, Terence T. Sio, Jonathan B. Ashman, Lisa A. McGee, Jean-Claude M. Rwigema, William G. Rule et al.
Published: 2026-10-13
Categories: eess.IV, cs.AI
arXiv: https://arxiv.org/abs/2610.18926v1

Adaptive radiotherapy must account for anatomical changes during treatment, but doing so depends on reliably mapping targets, organs at risk and dose across serial images. Existing registration methods treat each image pair independently, without using radiotherapy-specific context or the patient's preceding anatomical trajectory. We developed Clinical-context-Aware and Patient-adaptive Deformable Image Registration (CAP-DIR), integrating source contours, dose, clinical text and prior-registration features. We evaluated 16,361 CT examinations from 3,473 patients across four institutional cohorts and five public benchmarks. CAP-DIR increased internal-test organ-at-risk-region normalized cross-correlation from 84.76% to 96.92% relative to the strongest learning-based comparator, with low deformation folding, and remained superior across seven external cohorts. In a blinded review of 20 longitudinal cases, clinically acceptable contours increased from 45.0% to 82.5% for clinical target volumes and from 22.5% to 62.5% for organs at risk, while estimated editing time decreased. In a controlled benchmark, CAP-DIR reduced global dose-mapping error by 34.5% relative to the strongest comparator. These findings show that clinically informed, longitudinal registration can improve contour usability and dose mapping for adaptive radiotherapy.

---

### ChunkAct: Action Chunking for Agentic Robotics

Authors: Bingxin Xu, Wenrui Bao, Xinxin Liu, Jiawei Zhou, Yuzhang Shang, Emilio Ferrara
Published: 2026-10-13
Categories: cs.RO, cs.AI, cs.CL, cs.CV
arXiv: https://arxiv.org/abs/2610.18874v1

Agentic robotics places a large language model (LLM) agent in the robot control loop, where it composes perception, motion, and vision-language-action (VLA) tools under the guidance of a memory and thereby generalizes to new tasks without retraining. This generality comes at a steep price in efficiency. Every primitive requires a round of LLM reasoning over the full interaction history, so the cost of an episode grows with the number of primitives. We find this reasoning largely redundant: once a sub-goal is chosen, its primitives follow a routine already prescribed by memory. We present ChunkAct, which introduces action chunking to agentic robotics. Online, a single call launches a chunk of chained primitives, which executes autonomously and returns control only upon failure. Offline, ChunkAct compiles the memory into these chunks, so execution is memory-free at test time and the planner is reserved for decisions that require reasoning. LLM computation therefore scales with the number of decisions rather than the number of primitives. On LIBERO-Pro, ChunkAct improves task success by 5.0 points over the state of the art in agentic robotics while requiring 2.9x fewer LLM calls, 4.4x fewer input tokens, and 2.3x fewer output tokens.

---

### LoopWAM: Closed-Loop Generalist Policy Improvement from Deployment Rollouts

Authors: Wenxuan Guo, Yingping Liang, Zhaoguo Wang, Ziyuan Li, Angyuan Ma, Zhenyu Wu, Chengkun Wang, Bingyao Yu, Ze Chen, Borui Zhang, Xiuwei Xu, Meng Zhang et al.
Published: 2026-10-13
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2610.18873v2

Scaling expert demonstrations alone does not enable generalist robot policies to improve from deployment. Real-world rollouts contain successes, suboptimal behavior, and failures: imitating every recorded action can reinforce poor behavior, while filtering out failures discards evidence of physical interactions and their consequences. We introduce LoopWAM, a closed-loop world-action model that learns directly from this full quality spectrum, using rollouts from both itself and other deployed policies. Source-dependent noise sampling unifies behavior learning and action-conditioned world modeling, while quality steering and quality-weighted supervision retain full video supervision across qualities and regulate action imitation. Video and action tokens interact throughout parallel denoising, allowing learned consequences to inform action generation directly. On RoboChallenge Table30-V2, three deployment-and-training rounds raise success from 23.67% to 43.33%, a 19.7-percentage-point gain, with a final score of 56.80. LoopWAM also achieves 94.5% average success on RoboTwin 2.0 and 49.5% overall success on RoboCasa365 without additional embodied pretraining, obtaining the best overall results among the compared methods across all three benchmarks. On challenging shirt folding task, two rounds of deployment-driven improvement after single-task post-training raise success from 13.20% to 55.56%. These results demonstrate closed-loop improvement from mixed-quality experience under both multi-task learning and single-task post-training. Project page: https://loop-wam.github.io/

---

### LoopWAM: Closed-Loop Generalist Policy Improvement from Deployment Rollouts

Authors: Wenxuan Guo, Yingping Liang, Zhaoguo Wang, Ziyuan Li, Angyuan Ma, Zhenyu Wu, Chengkun Wang, Bingyao Yu, Ze Chen, Borui Zhang, Xiuwei Xu, Meng Zhang et al.
Published: 2026-10-13
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2610.18873v1

Scaling expert demonstrations alone does not enable generalist robot policies to improve from deployment. Real-world rollouts contain successes, suboptimal behavior, and failures: imitating every recorded action can reinforce poor behavior, while filtering out failures discards evidence of physical interactions and their consequences. We introduce \textbf{LoopWAM}, a closed-loop world-action model that learns directly from this full quality spectrum, using rollouts from both itself and other deployed policies. Source-dependent noise sampling unifies behavior learning and action-conditioned world modeling, while quality steering and quality-weighted supervision retain full video supervision across qualities and regulate action imitation. Video and action tokens interact throughout parallel denoising, allowing learned consequences to inform action generation directly. On RoboChallenge Table30-V2, three deployment-and-training rounds raise success from 23.67\% to 43.33\%, a 19.7-percentage-point gain, with a final score of 56.80. LoopWAM also achieves 94.5\% average success on RoboTwin~2.0 and 49.5\% overall success on RoboCasa365 without additional embodied pretraining, obtaining the best overall results among the compared methods across all three benchmarks. On challenging shirt folding task, two rounds of deployment-driven improvement after single-task post-training raise success from 13.20\% to 55.56\%. These results demonstrate closed-loop improvement from mixed-quality experience under both multi-task learning and single-task post-training. \href{https://loop-wam.github.io/}{\textcolor{orange}{Project Page}}

---

### EgoExplorer: A Vision-Motion-Touch Observatory for Human Loco-Manipulation In Situ

Authors: Yunsong Zhou, Xuekun Jiang, Haiqi Shen, Hangxu Liu, Chaojian Shi, Yujie Liu, Jingyuan Li, Haoran Yang, Jiafei Cao, Dongzhan Zhou, Dong Wang, Yao Mu et al.
Published: 2026-10-13
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2610.18872v1

Learning humanoid loco-manipulation from human demonstrations rests on observing how locomotion, posture, and object interaction evolve together as tasks unfold. Scaling collection across diverse environments, in turn, favors lightweight, portable capture systems. Yet, existing datasets provide only partial views of this coordination. Thus, we introduce EgoExplorer, a 100-hour vision-motion-touch observatory of human loco-manipulation built with a wearable system. With an emphasis on professional activities, its continuous recordings align egocentric video with reconstructed whole-body motion, fine-grained finger articulation, and tactile observations, accompanied by step-level descriptions. To achieve this coverage with sparse body sensing, we develop the Calibrate-Complete-Constrain pipeline. Concretely, reusable calibration fixes participant geometry and sensor attachments before motion inference to disentangle mounting offsets from articulation. A temporal prior then completes unobserved joint motion, while layered constraints promote measurement agreement, temporal continuity, and geometric consistency, properties that motion plausibility alone does not guarantee. Against tracker-fitted references, the method reduces body-point error by 71.9% relative to the strongest baseline; matched evaluations also show gains in hand stability and contact estimation. Preliminary studies explore multimodal generation and conditioning alongside humanoid retargeting with whole-body control in simulation.

---

### Cepheid: Decoupling Scale from Structure for Cross-Family Motion Completion

Authors: Chih-Hai Su, Tsung-Wei Ke, Yu-Lun Liu
Published: 2026-10-13
Categories: cs.CV, cs.GR, cs.LG
arXiv: https://arxiv.org/abs/2610.18871v1

Sparse-to-dense motion completion recovers dense mesh motion from sparse vertex displacements. Residual priors trained on one geometry family frequently fail to beat parameter-free harmonic extension off-family. Moreover, cross-family training meets a scaling barrier: residual amplitude varies by more than an order of magnitude across geometry families and collapses as observations become denser. Since this scale depends on unknown target motion, it cannot be measured directly at test time. We decouple residual structure from physical scale by learning a dimensionless prior, estimating amplitude from observables and a lightweight correction fitted only on training sources. Our unified model surpasses harmonic extension across all evaluated families, including unseen cloth and legged bodies, without skeletons, registration, or cross-mesh correspondence. The same prior applies without retraining to point-tracked animal video and a robot grasping a soft body. Project page: https://su-terry.github.io/cepheid-website/

---

### MiniDog: Dog-Breed Text-to-image Flow Matching from Pretraining to Fine-Tuning on 4 Consumer GPUs

Authors: Reyhaneh Esmailizadeh, Xingjian Leng, Zhanhao Liang, Liang Zheng
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18870v2

We generally perceive that training text-to-image diffusion models requires heavy compute: it is true if we aim at a comprehensive frontier model. This understanding prevents people from training their own models when getting started in the field. We develop MiniDog, a framework that enables pretraining diffusion transformers in 55 minutes and fine-tuning in another 15 minutes on four 3090 GPUs. For data, we use 26k dog images from ImageNet and caption them for pretraining and use 2k images generated by FLUX.1-schnell for fine-tuning. We precompute the image and text features and cache them for efficiency. For architecture, we use E2E-INVAE for its semantic latent space, a small transformer with 22M parameters, a small Qwen model for text encoding, and in-context conditioning. For training, we use iREPA and an EUPE perception encoder for acceleration. For evaluation, our repository supports FID, HPSv2 score, PickScore, and various VAE reconstruction metrics. The project is open-sourced at https://github.com/Reyhaneesmailizadeh/MiniDog. We do not claim technical novelty.

---

### Dog-Breed Text-to-image Flow Matching from Pretraining to Fine-Tuning on 4 Consumer GPUs

Authors: Reyhaneh Esmailizadeh, Xingjian Leng, Zhanhao Liang, Liang Zheng
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18870v1

We generally perceive that training text-to-image diffusion models requires heavy compute: it is true if we aim at a comprehensive frontier model. This understanding prevents people from training their own models when getting started in the field. We develop MiniDog, a framework that enables pretraining diffusion transformers in 55 minutes and fine-tuning in another 15 minutes on four 3090 GPUs. For data, we use 26k dog images from ImageNet and caption them for pretraining and use 2k images generated by FLUX.1-schnell for fine-tuning. We precompute the image and text features and cache them for efficiency. For architecture, we use E2E-INVAE for its semantic latent space, a small transformer with 22M parameters, a small Qwen model for text encoding, and in-context conditioning. For training, we use iREPA and an EUPE perception encoder for acceleration. For evaluation, our repository supports FID, HPSv2 score, PickScore, and various VAE reconstruction metrics.

---

### What Matters in Quality-Aware VLA Pretraining: A Benchmark and Empirical Study

Authors: Mingxuan Yan, Litian Gong, Dihong Huang, Zhixu Li, Tianyu Zhang, Wenqian Zhang, Hefeifei Jiang, Peihao Li, Jian Zhang, Mengfei Zhao, Yikai Tang, Hai Zhai et al.
Published: 2026-10-13
Categories: cs.RO, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2610.18869v1

How should demonstration quality be used during vision-language-action (VLA) pretraining to improve performance after target-task fine-tuning? We compare weighting, filtering, and conditioning across reward definitions and label granularities using 5,000 naturally suboptimal human demonstrations and 24 target tasks. Within each VLA, we match demonstrations and training budgets. Additional comparisons examine pretraining data size and the roles of quality tags during fine-tuning and inference. Across three VLAs, we find that approach-aware episode conditioning improves mean success over plain behavioral cloning; conditioning and filtering favor episode and chunk labels, respectively, on average, with model-dependent reversals; and constant-tag fine-tuning improves mean success after conditioned pretraining with either tagged or untagged inference. We introduce AXIS-Bench to support these controlled comparisons. Built on crowdsourced simulation demonstrations, it provides recorded object and contact states for recomputing quality labels on shared trajectories. Replayable trajectories support workspace-aligned rendering and user-defined real-world target tasks. In our real-world experiment, simulation pretraining raises success from 9.6\% to 31.7\%. Project page: https://axis-bench.github.io/.

---

### Tev: Scaling Long-Horizon Agentic Environments with Dense Progress Rewards

Authors: Wenbo Hu, Shraman Pramanick, Krishna Kumar Singh, Nanyun Peng, Kai-Wei Chang, Yong Jae Lee
Published: 2026-10-13
Categories: cs.CL, cs.AI, cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2610.18868v1

Long-horizon agents have drawn growing attention for their ability to work autonomously -- from hours to days -- on economically valuable tasks, a capability essential to automated discovery and recursive self-improvement (RSI). Yet open-source models still struggle on long-horizon tasks, largely for lack of both training data and a working recipe, and scaling both is hard: long, complex tasks are difficult to curate and verify, and their rewards are sparse. We address both. First, we introduce Tev-Gym, the first long-horizon agent training dataset with verifiable, dense, progressive rewards -- 1,170 verified task environments across 7 domains. It is built on a latent-world-first approach that first constructs a latent world: a complete and correct hidden specification of the task domain, workflow, and expected state after each step. Each task is then progressively rendered from this specification into an agent-facing environment, enabling controllable complexity and a natural progress metric. Second, we turn this metric into a training recipe: via progress-potential reward shaping (PPRS), we convert progress into a dense per-step reward that improves training efficiency and makes otherwise-intractable tasks trainable. Training with this recipe yields Tev-27B, which achieves substantial gains on the Agents' Last Exam computing and math sciences domains (ALE-CM), matching GPT-6 Astra and Claude Opus 5 at Pass@5. Despite training only on ALE-related data, it generalizes to the Terminal-Bench (TB) 3.0 and 4.0 subsets, surpassing GLM 5.2 and Gemini 3.7 Flash, and reaches Claude Opus 4.8-level performance on middle-horizon tasks such as TB 2.1.

---

### PLAYGROUND: Progressive Layout Generation with Render-Grounded Planning

Authors: Jaeho Lee, Junhwan Heo, Junghyun Park, WonJun Moon, Eunju Yang, Seungho Jang, Seongchan Kim, Seungryong Kim
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18865v1

Generating editable graphic designs requires arranging visual and textual elements into a coherent composition while preserving their content and editability. Although recent MLLM-based approaches employ iterative generation and visual refinement to reflect the real-world design workflows, their high-level plans are predetermined prior to execution, limiting their adaptability as the design evolves. To address this limitation, we propose PlayGround, a progressive layout generation framework that dynamically adapts to the current rendered canvas. PlayGround establishes a progressive design loop comprising a Progressive Planner, a Layout Generator, and a Visual Verifier, guided by a Structure-Aware Retriever. Given an initial set of unordered elements, the Structure-Aware Retriever initially infers a coarse layout structure to retrieve structurally similar reference designs. Subsequently, at each turn, the Progressive Planner examines the current rendered canvas alongside the remaining elements, forms the next functional group, and plans its composition. The Layout Generator then predicts editable layout attributes, while the Visual Verifier evaluates rendered candidates and provides feedback when refinement is necessary. To train the Progressive Planner with step-wise supervision, we construct a two-stage data generation pipeline that extracts functional groups and turn-level planning trajectories from completed designs. Experiments on Crello and LICA demonstrate the effectiveness of PlayGround over existing methods. Our code and weights will be publicly released.

---

### Data Free Training of Diffusion Models for Lattice Field Theories

Authors: Dennis Grinwald, Ankur Singha, Khaled Kahouli, Karl Jansen, Shinichi Nakajima
Published: 2026-10-13
Categories: hep-lat, cs.LG
arXiv: https://arxiv.org/abs/2610.18863v1

Sampling equilibrium field configurations remains a major computational bottleneck in lattice field theory, motivating the development of generative samplers that can produce independent configurations. Diffusion models are promising for this purpose, but their standard denoising score matching objective requires samples from the target distribution, seemingly reintroducing the original sampling problem. Here, we show that diffusion models can instead be trained from their own generated samples using the target score identity, provided access to the unnormalized probability density. Importance-weighted formulations of this identity yield a consistent target throughout training, while unweighted variants define a self-consistent fixed-point training scheme. We benchmark the different approaches in two-dimensional \(φ^4\) theory and identify effective choices that enable efficient training and sampling across the symmetric, critical, and broken phases. Ultimately, our data-free training approach achieves comparable performance to training on HMC reference data.

---

### DoubtLess: Training-Free Contrastive Decoding for Efficient Reasoning

Authors: Minghao Luo, Liang Chen
Published: 2026-10-13
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2610.18862v1

Reasoning models achieve strong performance through long chains of thought (CoT), but often spend more tokens than needed to solve a problem. Existing ways to shorten the CoT require additional training for each model, add a fixed steering vector to the model's hidden states, or use decoding heuristics such as stopping early or banning words like "wait", which can reduce accuracy. We propose DoubtLess, a training-free decoding method in which a much smaller reasoning model steers the large one towards brevity. The small model reads the large model's CoT so far twice, once prompted to be brief (the brief thinker) and once prompted to be verbose (the verbose thinker), and the difference between its two logit vectors, scaled by a strength, is added to the large model's logits at every step of the CoT. On seven benchmarks of competition mathematics, graduate-level science and competitive programming, DoubtLess shortens the CoT of Qwen3-32B by 31% with unchanged accuracy, outperforming five inference-time baselines. We also show that DoubtLess reduces thinking tokens per turn by 20% in a self-debugging coding agent with unchanged task performance. It also generalizes to two other model families with unchanged accuracy. When combined with training-based compression, it further shortens the CoT of a model already trained for brevity. Ablations and behavioral analysis show that DoubtLess reduces repeated checking, with most token savings coming after the correct answer first appears in the CoT. Our code is available at https://github.com/leoluolol/DoubtLess.

---

### Hybrid Reasoning Agents

Authors: Jiaru Zou, Yunzhe Qi, Sirui Chen, Federico Bianchi, Shang Zhu, Pan Lu, Jingrui He, James Zou
Published: 2026-10-13
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2610.18861v1

How can we build an agent that thinks beyond human language, yet still acts through it in the external environment? Modern LLM agents still reason and interact through text, while latent reasoning offers an efficient internal medium but lacks native tool access and reasoning transparency. We bridge these two directions with Hybrid Reasoning Agents (HybridAgents), which dynamically navigate between latent reasoning and text-based interaction throughout a unified generation process. Architecturally, a lightweight HybridLink enables seamless transitions between text and latent modes, while the agent's continuous working memory preserves lossless hybrid context across multiple turns. These designs preserve tool-integrated agentic capabilities while making latent thoughts directly inspectable. Building on this architecture, we further develop dedicated post-training algorithms for hybrid reasoning: HybridSFT stabilizes tool-integrated hybrid generation by teaching the backbone to coordinate with HybridLink, then HybridRL scales reasoning by learning when to think in latent space and when to act through text. We evaluate HybridAgent at 3 model scales across 11 benchmarks covering mathematics, science, medicine, search, and agentic reasoning. Compared with frontier LLM Agents and agentic frameworks spanning both text-based and latent methods, HybridAgent consistently improves performance by 10.4%, while delivering 35.5%-54.7% token reduction, 1.8x-2.6x faster inference, and 2.7x-3.7x cost savings. Code and data are available at https://github.com/togethercomputer/HybridAgent.

---

### AesCode: Aesthetic Code Generation with Decoupled Cross-Modal Rewards

Authors: Xin Gao, Yifan Yang, Xin Li, Pan Wang, Yan Li, Muzhao Tian, Mingxi Cheng, Bei Liu, Kai Qiu, Qi Dai, Dong Chen, Yue Dong et al.
Published: 2026-10-13
Categories: cs.AI
arXiv: https://arxiv.org/abs/2610.18860v1

Generating information-rich visual artifacts such as slides, posters, and dashboards as code is appealing because the outputs are structured, editable, and verifiable. Yet code models cannot see how their choices of layout and color come together on the canvas, often limiting spatial planning and aesthetic quality. Image generators excel at visual design but can misrender requested content, making their outputs unreliable targets for direct reproduction. We therefore introduce AesCode, which generates code grounded in the content prompt while using an image generated from the same prompt as an aesthetic guidance. To do so, AesCode represents the target as a design graph that describes the whole canvas and makes each property attributable. From this graph, we derive complementary cross-modal rewards: deterministic verifiers score properties parsable from code and rendering, while a VLM judge assesses the non-parsable ones with a sample-specific Visual Graph Rubric (VGR) tied to graph elements and relations. We optimize these decoupled rewards through reinforcement learning with per-channel normalization to jointly improve spatial planning, aesthetic quality, and verifiable correctness. To evaluate this setting, we build a benchmark that scores verifiable correctness and visual quality separately, with visual scores highly consistent with human judgments. We find that reference images substantially improve the visual quality of code generated by existing models, and ablations show that AesCode benefits from references both as model input and as a source of visual rewards. Relative to same-scale Qwen3-VL baselines given the same references, AesCode-8B and AesCode-32B improve visual quality by 31.1 and 22.4 percentage points, reaching scores comparable to those of closed-source frontier models such as GPT-5.5 and Claude Opus 4.8.

---

### Visual Tokens on Demand: Visual Budgeting for Scalable and Efficient Multi-Image Reasoning in Multimodal LLMs

Authors: Aimon Rahman, Joymallya Chakraborty, Vidit Bansal, Abhishek Dan
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18858v1

Multimodal large language models have advanced rapidly, but most remain optimized for single-image inputs. Multi-image reasoning is costly because each image adds many visual tokens, increasing encoding and attention demands. Existing techniques that address the above issue, including visual token pruning, token merging, visual token compression, learned resampling, hierarchical compression, retrieval-augmented visual reasoning, and memory-based architectures, have largely been designed for single-image scenarios, where redundancy is primarily spatial and local. Multi-image reasoning introduces additional challenges: redundancy may occur across images, task-relevant evidence may be sparse or distributed, and aggressive per-image compression can remove details needed for comparison, counting, temporal reasoning, retrieval, or evidence grounding. Moreover, existing token-reduction methods are often either (i) training-free but inference-inefficient, typically relying on accessing attention matrices to get token-importance that may be impractical to access with memory-efficient attention implementations such as FlashAttention, or (ii) training-dependent, requiring the base MLLM to be retrained or adapted to a new compressed-token paradigm. In this work, we study scalable multi-image reasoning and understanding without retraining the base model and without accessing attention matrices at inference time. To this end, we propose mmVB, a lightweight plug-and-play module that assigns query-conditioned per-image token budgets via cross-image context, adding only 1.6M trainable parameters while reducing visual input tokens by 9x. This reduction yields 5x faster time-to-first-token and 7x fewer prefill FLOPs, while maintaining 94.4% of full-resolution accuracy across eight multi-image reasoning benchmarks.

---

### PixelMoE: Scaling Pixel-Space Diffusion with LLM-Style Mixture-of-Experts

Authors: Ruixiao Dong, Jianmin Bao, Bozhi Luan, Zijie Meng, Xin Tao, Pengfei Wan, Kun Gai, Li Li, Houqiang Li
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18857v1

Pixel-space diffusion models provide an appealing alternative to latent-space diffusion models by directly modeling raw images without relying on a learned tokenizer. However, scaling pixel-space models has predominantly followed dense strategies, tightly coupling model capacity with active computation. In this work, we show that standard LLM-style Mixture-of-Experts (MoE) architectures can be effectively transferred to pixel-space diffusion. Building on this finding, we introduce PixelMoE, a pixel-space generative framework that sparsely scales model capacity with standard LLM-style MoE, without diffusion-specific routing mechanisms or auxiliary routing objectives. Specifically, PixelMoE combines standard Token-Choice (TC) routing with fine-grained routed experts for input-adaptive sparse computation, a shared expert for capturing common representations, and a bias-based load-balancing strategy. To complement sparse capacity scaling, PixelMoE further introduces Pixel-Space Hierarchical Flow Matching, which varies spatial resolution along the denoising trajectory from global structure formation to fine-grained detail refinement. At comparable active parameter counts, PixelMoE consistently outperforms its dense counterparts across model scales, demonstrating effective sparse capacity scaling in pixel space. On class-conditional ImageNet-1K, PixelMoE achieves FID scores of 1.47 at 256x256 and 1.53 at 512x512 without representation alignment or perceptual losses, comparing favorably with leading pixel-space and latent-space generative models. Code and models will be released soon.

---

### Exact Topological Compliance: Generating Persistence-Equivalent Graphs

Authors: Mattie Ji, Indradyumna Roy, Vikas Garg
Published: 2026-10-13
Categories: cs.LG, math.AT
arXiv: https://arxiv.org/abs/2610.18856v1

Persistent homology summarizes the topological evolution of a graph as persistence diagrams (PDs). Generating graphs whose topology matches a target descriptor is increasingly important, yet existing topology-aware generative models rely on soft regularization, producing graphs whose PDs may only be approximately similar. We initiate a principled investigation of the generation problem: produce graphs that realize a target PD exactly. We propose two approaches guided by the implicit local and global constraints a target PD imposes. The local approach builds degree-based PD-equivalent graphs iteratively, and we prove that every output realizes the target PD and that every degree-based PD-equivalent graph is reachable. However, as shown thereafter, local generation necessarily trades off between sample diversity and reaching infeasible states. We therefore recast generation as a global constraint satisfaction problem with a complete encoding solvable by modern constraint programming solvers. This is applicable to any vertex-based permutation equivariant filtration scheme and is minimal under the degree filtration. Experiments on four standard graph generation benchmarks confirm that both methods realize the target PD exactly. Together, our methods provide a principled treatment of strict topology-compliant graph generation. Code is available at https://github.com/Aalto-QuML/ExactCompliance.

---

### Two Clocks: Why Larger Rollout Batches Learn Slower but Fail on Schedule in Reinforcement Learning with Verifiable Rewards

Authors: Yuante Li, Junwei Quan, Kai Qiu, Kate Zhang, Yicheng Tao
Published: 2026-10-13
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2610.18853v1

Reinforcement learning with verifiable rewards (RLVR) relies on costly model-generated responses, motivating larger rollout batches and square-root learning-rate scaling. Reported to approximately preserve progress per sample with one optimizer step per rollout, this rule is less understood in practical multi-step training. We develop a two-clock framework that separates the samples needed to make progress from those consumed before collapse. Under a local quadratic model, the square-root rule preserves first-order drift per sample but not progress: larger batches take fewer, longer updates and lose more to curvature, yielding a learning-rate-dependent critical batch beyond which sample efficiency degrades. A drift-threshold hypothesis predicts a complementary failure clock whose sample count remains approximately invariant along the square-root rule. Controlled GRPO experiments on Qwen2.5-Math-1.5B across a $16\times$ batch range support both predictions. Ratio clipping delays failure, especially for larger batches, but the progress gap persists without it. Widening the clip trades stability for a transient speed gain. The progress gap reappears on another dataset, a general-purpose checkpoint, and a 7B model. Finally, we use the two-clock framework to derive a simple multi-step RLVR schedule: start below the critical batch and switch to a larger batch before failure. Our code is available at: https://github.com/WinstonLiyt/two-clocks-rlvr.

---

### MFG: Multi-turn Feedback Guided Molecular Optimization

Authors: Trieu Nguyen, Frazier N. Baker, Ruoxi Gao, Daniel Adu-Ampratwum, Xia Ning
Published: 2026-10-13
Categories: cs.AI, cs.LG, q-bio.QM
arXiv: https://arxiv.org/abs/2610.18851v1

Multi-objective molecular optimization is a critical process in drug discovery that aims to modify a given molecule to simultaneously improve multiple properties while maintaining a similar overall structure. This process often requires iterative modifications guided by computational or experimental feedback. Recent large language model (LLM)-based methods have begun to explore this iterative approach; however, these methods are not explicitly trained to leverage such feedback, while existing approaches primarily focus on single-turn molecular optimization. To address this gap, we introduce Multi-turn Feedback Guided (MFG) molecular optimization, a reinforcement learning framework that trains an LLM to progressively optimize molecules using feedback from previous turns. MFG introduces a feedback vector capturing both property improvements and source molecule similarity, alongside a feedback-aligned reward function that promotes progressive improvement across turns. We apply MFG over $\text{GeLLM}^{3}\text{O}$, a state-of-the-art LLM model for multi-objective molecular optimization, evaluate $\text{GeLLM}^{3}\text{O-MFG}$ on ten multi-objective molecular optimization tasks. The results demonstrate that $\text{GeLLM}^{3}\text{O-MFG}$ achieves the highest average similarity-constrained success rate ($81.50%$), outperforming the base model $\text{GeLLM}^{3}\text{O}$ ($70.81%$).

---

### WorldRecon: Generative Multi-View Mesh Reconstruction for Large-Scale 3D World

Authors: Mengchen Zhang, Yang Li, Chunshi Wang, Tong Wu, Chunchao Guo
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18846v1

Reconstructing complete, high-quality meshes of large-scale scenes from unposed multi-view images requires faithful geometric grounding while plausibly completing surfaces that are unobserved or only sparsely captured. We present WorldRecon, a generative multi-view reconstruction framework for recovering complete, large-scale 3D scene meshes from unposed RGB images. WorldRecon uses a coarse, noisy global point cloud and camera poses estimated from the input images by a reconstruction model as a global scaffold for whole-scene generation and partitions it into overlapping spatial blocks. To robustly leverage these imperfect geometric priors, we equip the native 3D generative model TRELLIS.2 with a dedicated reconstruction-based conditioning architecture that generates each block from its local point-cloud crop and block-relevant image observations, producing sparse structure, detailed geometry, and PBR materials. To maintain scene-level coherence, we introduce neighbour-aware latent inpainting and a causal world-generation procedure that propagates structural, geometric, and material context across overlapping blocks. Together, these designs scale multi-view generative reconstruction to large-scale scenes, yielding coherent, unified meshes. Extensive experiments demonstrate that WorldRecon achieves state-of-the-art reconstruction performance, reducing Chamfer distance by 14.9% and 43.1% while improving F-score by 17.3% and 12.0% over the strongest baselines on synthetic and real-world indoor scenes, respectively, thereby demonstrating faithful and complete generative reconstruction as well as strong sim-to-real generalization. Project page: https://kszpxxzmc.github.io/WorldRecon-website/.

---

### The Last Remote Sensing Model Built by Humans? Casting Any Earth Observation Task as Recursive Self-Improvement

Authors: Sijie Zhao, Zeqi Zhou, Heli Qi, Xuanchao Fu, Zhuo Zheng, Junshi Xia, Junjue Wang, Weihao Xuan, Naoto Yokoya
Published: 2026-10-13
Categories: cs.CV, cs.CL
arXiv: https://arxiv.org/abs/2610.18845v2

Earth observation (EO) models are still built by hand. For each new task, experts design a training recipe and tune it against a benchmark. We show that this loop can be handed to an agent. Any verifiable task can be cast as a harness and a verifier, where the harness defines what the agent may use and the verifier defines what counts as success. Every EO task can be made verifiable, through labeled benchmarks, independent sensors, physical models or the repeated observation of the same place. Given only these two components, a general-purpose coding agent builds the EO model and improves it recursively on its own outputs. Settings long treated as separate research problems, from learning with a handful of labels to adapting across domains, become choices of harness and verifier. On six EO benchmarks spanning optical, SAR and multispectral time-series imagery, with outputs that range from land-cover maps and ship instances to canopy heights and answers about large images, agent-built models surpass the state of the art on every one, and on land-cover mapping, they do so from 11% of the training labels. This establishes a new paradigm for EO, in which people design harnesses and verifiers and agents build the models, so EO models improve as agents do.

---

### The Last Remote Sensing Model Built by Humans? Casting Any Earth Observation Task as Recursive Self-Improvement

Authors: Sijie Zhao, Zeqi Zhou, Heli Qi, Xuanchao Fu, Zhuo Zheng, Junshi Xia, Junjue Wang, Weihao Xuan, Naoto Yokoya
Published: 2026-10-13
Categories: cs.CV, cs.CL
arXiv: https://arxiv.org/abs/2610.18845v1

Earth observation (EO) models are still built by hand. For each new task, experts design a training recipe and tune it against a benchmark. We show that this loop can be handed to an agent. Any verifiable task can be cast as a harness and a verifier, where the harness defines what the agent may use and the verifier defines what counts as success. Every EO task can be made verifiable, through labeled benchmarks, independent sensors, physical models or the repeated observation of the same place. Given only these two components, a general-purpose coding agent builds the EO model and improves it recursively on its own outputs. Settings long treated as separate research problems, from learning with a handful of labels to adapting across domains, become choices of harness and verifier. On four EO benchmarks spanning optical, SAR and multispectral time-series imagery, agent-built models surpass the state of the art and keep most of that accuracy from about 1% of the training labels. This establishes a new paradigm for EO, in which people design harnesses and verifiers and agents build the models, so EO models improve as agents do.

---

### RepProbe: Can Video Language Models Count Repeated Actions? A Diagnostic Benchmark

Authors: Mahsa Khoshnoodi, Sarah Adel Bargal
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18844v1

Repetition counting requires identifying and enumerating action cycles over time. While state-of-the-art video language models describe actions fluently, they fail to reliably count how many times an action repeats. To investigate this limitation, we introduce RepProbe, a benchmark of 42K real and synthetic video clips containing exactly k in (1, 20) complete repetitions, with controls for clip position, temporal coverage, and duration. These controls allow us to study the counting failure directly and locate where in the model it arises. Counting accuracy collapses at high counts across the video language models we evaluate, and controlled experiments show this collapse is not explained by frame sampling alone. We compare model outputs with linear probes trained on the final-layer representations used to generate those outputs. We find a consistent representation-to-output dissociation: substantial count information remains linearly decodable even when models fail to report the correct count. Further analysis shows this failure arises primarily at the output stage, where answers concentrate on a small set of preferred numbers rather than the available count information. The dissociation holds on both real and synthetic videos and is not explained by clip duration, position, or scene content. These results indicate that models encode repetition structure but do not reliably report it in their outputs. Project page: https://arenaa.github.io/RepProbe-webpage/

---

### LIBERO-RIG: A Hardware-Setup Robustness Benchmark for Vision-Language-Action Models

Authors: Andrea Yaoyun Cui, Hanyu Wang, Xinyi Wang, Bangzheng Li, Lin Zhang, Zhenggang Tang, Shyamal Buch, Dejia Xu, Matthew Tancik
Published: 2026-10-13
Categories: cs.RO, cs.AI
arXiv: https://arxiv.org/abs/2610.18843v1

Robot policies are expected to generalize across deployment environments, including variations in the robot itself: cameras are bumped and remounted, calibrations drift, controllers run at different rates, and end-effector pads wear down. We present LIBERO-RIG, a benchmark that perturbs only the robot's hardware and setup while holding every LIBERO task fixed. The benchmark spans eleven perturbation axes, including camera mounting and optics, camera-to-arm calibration, actuation gain, control rate, proprioceptive noise, base pose, table height, and gripper pads. Each axis has five severity levels, yielding 210 conditions across the four LIBERO suites. We evaluate nine recent vision-language-action models using 500 episodes per condition and substantiate each finding with rigorous statistical tests. Pooled success falls from 0.932 to 0.767 at the middle severity and 0.509 at the highest severity. Performance degradation varies substantially across perturbation axes and action interfaces, and training on visual perturbations does not mitigate faults that leave camera images unchanged. We release the benchmark together with manifests specifying the fixed perturbation parameters for each episode.

---

### Who Improves the Robot? Toward Physical Recursive Self-Improvement

Authors: Hai Zhou, Fan Yang, Zeyi Chen, Lejun Zhang, Chenyang Fu, Jianping Li, Jiayi Zhu, Xianyang Liu, Yingxuan Yang, Lili Meng, Lu Wang, Xinyang Han et al.
Published: 2026-10-13
Categories: cs.AI, cs.LG, cs.RO
arXiv: https://arxiv.org/abs/2610.18842v1

Robots are increasingly autonomous in task execution, yet their improvement remains largely human-driven. This study investigates Physical Recursive Self-Improvement (Physical RSI), in which physical experience guides lasting updates to both a robot's capabilities and its improvement procedure, making subsequent improvement more effective across tasks. Physical RSI faces challenges from noisy observations, continuous dynamics, costly trials, and irreversible consequences. We present an L0-L5 framework that differentiates capability-level from meta-level updates and within-task from across-task gains, reserving L5 for independently verified recursive gain. We study existing methods around harness adaptation and model training, alongside safety and control mechanisms that govern execution and successive updates. Our analysis reveals that, despite substantial progress in self-improvement, the surveyed literature does not yet demonstrate Physical RSI as defined by our framework. We identify research priorities in reliable feedback, causal attribution, safety preservation, and benchmarking, outlining a path toward robots that autonomously improve how they learn while ensuring safety. The GitHub repository for Physical RSI is available at https://github.com/SAIL-Research-Lab/Physical-RSI.

---

### Fast Rates for Offline Q-Learning with KL Regularization

Authors: Yize Hao, Guangxin Jiang, Shuang Qiu
Published: 2026-10-13
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2610.18839v1

This paper investigates offline Q-learning for finite-horizon Markov decision processes with Kullback-Leibler (KL) regularization. Recent studies have shown that KL regularization can yield fast rates in offline decision-making problems. However, it remains unclear whether Q-learning, which adopts incremental stochastic approximation updates, can achieve such fast rates in the offline setting. We first propose KL-regularized Q-Learning (KLQ), a simple variant of classical Q-learning. By exploiting KL regularization, we prove that KLQ achieves a sample complexity of $\widetilde{O}(1/ε)$ without explicit pessimism. This guarantee, however, has an exponential dependence on the KL regularization parameter and the horizon. To address this limitation, we further propose KL-regularized Pessimistic Q-Learning (KLPQ), which incorporates pessimism into KLQ. We prove that KLPQ retains the same \(\widetilde O(1/ε)\) rate while removing the explicit exponential dependence, thereby further improving the sample complexity. To the best of our knowledge, these results establish the first fast-rate guarantees for offline Q-learning in tabular finite-horizon KL-regularized MDPs.

---

### Scaling Online Optimal Transport Minibatches for Flow Matching

Authors: Peter Chen, Sergey Levine
Published: 2026-10-13
Categories: cs.LG, cs.CV
arXiv: https://arxiv.org/abs/2610.18838v1

A central challenge in scaling flow models is maintaining stability in training and integration. In flow matching, this stability depends critically on how source samples are paired with data. Independent random pairing introduces substantial stochasticity, motivating minibatch optimal transport (OT) to construct online source-target couplings that reduce this randomness (Tong et al., 2024). However, as applications extend to larger and higher-dimensional datasets, small minibatches can limit these benefits, while scaling the coupling to substantially larger minibatches at every iteration is computationally prohibitive. Moreover, how OT coupling affects the full stochastic gradient and integration remains theoretically unresolved. In this work, we develop a scalable online entropic OT approximation that enables substantially larger source-target couplings while preserving the benefits of OT. We close the theoretical gap for both unregularized and entropic OT by proving strict reductions in full stochastic-gradient variance and integration error relative to independent pairing, and further characterize how these improvements strengthen as the online coupling minibatch grows. Experiments on single-cell gene interpolation-flow and image-flow training validate these predictions, showing that larger online couplings consistently improve training stability and final generation performance.

---

### Plan-Bench: Can Existing VLMs Go Well Beyond Perception to Reason and Plan in Embodied Worlds?

Authors: Juntong Feng, Qingan Liu, Jingru Dong, Yue Zhu, Weiming Zheng, Yujie Cheng, Xinzhe Han, Shuai Hao, Yunzhi Zhuge, Haiwen Diao, Huchuan Lu
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18836v2

Vision-Language Models (VLMs) have achieved strong performance in visual perception and spatial reasoning, leading to growing interest in using them as high-level planners for embodied AI. However, physical tasks require models to integrate spatial information, reason over temporal states, and generate multi-step plans under various constraints, raising the question of whether current VLMs can reliably support such reasoning and planning independently of execution. We introduce Plan-Bench, a benchmark for evaluating the high-level reasoning and multi-step planning capabilities of VLMs relevant to embodied worlds, while isolating the influence of action execution. It covers three complementary dimensions, spanning 16 subclasses and 699 curated samples. Extensive evaluation of 28 VLMs reveals room for improvement, with even the strongest model achieving only 75.0% accuracy. We further find that model scaling generally improves these capabilities. Ablation studies also indicate larger models show increasing potential as high-level planners. Overall, Plan-Bench provides a lightweight probe for evaluating VLMs' reasoning and planning capabilities relevant to embodied environments and assessing their potential as decision makers for physical tasks. Our benchmark and code are available at https://github.com/MinneFeng/Plan-Bench.

---

### Plan-Bench: Can Existing VLMs Go Well Beyond Perception to Reason and Plan in Embodied Worlds?

Authors: Juntong Feng, Qingan Liu, Jingru Dong, Yue Zhu, Weiming Zheng, Yujie Cheng, Xinzhe Han, Shuai Hao, Yunzhi Zhuge, Haiwen Diao, Huchuan Lu
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18836v1

Vision-Language Models (VLMs) have achieved strong performance in visual perception and spatial reasoning, leading to growing interest in using them as high-level planners for embodied AI. However, physical tasks require models to integrate spatial information, reason over temporal states, and generate multi-step plans under various constraints, raising the question of whether current VLMs can reliably support such reasoning and planning independently of execution. We introduce Plan-Bench, a benchmark for evaluating the high-level reasoning and multi-step planning capabilities of VLMs relevant to embodied worlds, while isolating the influence of action execution. It covers three complementary dimensions, spanning 16 subclasses and 699 curated samples. Extensive evaluation of 28 VLMs reveals room for improvement, with even the strongest model achieving only 75.0% accuracy. We further find that model scaling generally improves these capabilities. Ablation studies also indicate larger models show increasing potential as high-level planners. Overall, Plan-Bench provides a lightweight probe for evaluating VLMs' reasoning and planning capabilities relevant to embodied environments and assessing their potential as decision makers for physical tasks. Our benchmark and code are available at https://github.com/MinneFeng/Plan-Bench.

---

### RoboSimStudio: An Interactive Toolkit for Human and Agent Rapid Real2Sim of Robot Manipulation Tasks

Authors: Kaiyuan Chen, Shuangyu Xie, Uksang Yoo, Xinyue Zhu, Cael Magner, Alvin Zhu, Daehwa Kim, Eric Heiden, Erwin Coumans, Miles Macklin, Spencer Huang, Ken Goldberg
Published: 2026-10-13
Categories: cs.RO, cs.AI, cs.GR, cs.LG
arXiv: https://arxiv.org/abs/2610.18835v2

It has been difficult to accurately model the physics of the interaction between a robot and the objects it manipulates, including contact mechanics, coupled contact dynamics, deformations, stiffness and friction. We present RoboSimStudio (RSS), an interactive toolkit that makes choices of solver, contact model, and physical parameters explicit and editable, allowing both humans and coding agents to quickly create and iteratively refine simulations based on observed physical behavior. We implement RSS with the open Newton physics simulation system and describe experiments with human-guided real2sim using (1) visual evaluation, (2) force-feedback and (3) agentic Real2Sim2Real. Code, video and additional information can be found on the project website: https://robosimstudio.github.io

---

### RoboSimStudio: An Interactive Toolkit for Rapid Real2Sim of Robot Manipulation Tasks

Authors: Kaiyuan Chen, Shuangyu Xie, Uksang Yoo, Xinyue Zhu, Cael Magner, Alvin Zhu, Daehwa Kim, Eric Heiden, Erwin Coumans, Miles Macklin, Spencer Huang, Ken Goldberg
Published: 2026-10-13
Categories: cs.RO, cs.AI, cs.GR, cs.LG
arXiv: https://arxiv.org/abs/2610.18835v1

It is difficult to accurately model the physics of the interaction between a robot and the objects it manipulates, including contact mechanics, coupled contact dynamics, deformations, stiffness and friction. We present RoboSimStudio (RSS), an interactive toolkit that makes choices of solver, contact model, and physical parameters explicit and editable, allowing both humans and coding agents to iteratively refine simulations based on observed physical behavior. We implement RSS with the open Newton physics simulation system and describe experiments with human-guided real2sim using (1) visual evaluation, (2) force-feedback and (3) agentic Real2Sim2Real. Code, video and additional information can be found on the project website.

---

### Game2Real-Bench: Do Toy Games Reveal and Transfer to Real-World Visual Reasoning?

Authors: Qingan Liu, Juntong Feng, Yue Zhu, Yujie Cheng, Xinzhe Han, Shuai Hao, Yunzhi Zhuge, Haiwen Diao, Huchuan Lu
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18834v1

Visual reasoning is an important capability for vision-language models (VLMs) to understand visual evidence, spatial structure, constraints, and state changes. Existing evaluations mainly rely on real-world data or procedurally generated tasks. Real-world data preserve natural complexity, while procedural games provide controllable generation and verifiable solutions. However, it remains unclear whether abilities measured in toy games correspond to real-world visual reasoning and whether game examples can support transfer. We introduce Game2Real-Bench, a benchmark that studies these connections across four capability groups, including fine-grained perception, spatial reasoning, constraint reasoning, and sequential planning. We construct 1,000 procedurally generated game samples with executable verification and curate 400 real-world samples aligned by shared reasoning requirements rather than visual similarity. We evaluate 25 representative VLMs and observe strong game-to-real consistency, with Pearson and Spearman correlations of 0.935. We further study one-shot game-to-real transfer on four models. Constraint reasoning improves for several models, while fine-grained perception declines with game demonstrations. We also construct Game2Real-Dataset with nearly 24K verified game samples across 18 tasks and 65 QA families. Our benchmark, dataset, and code are available at https://github.com/QingAnLiu/Game2Real-Bench.

---

### OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understanding

Authors: Xuyang Liu, Shihao Wang, Xinyu Wei, Minghan Li, Zeqing Wang, Zuhao Yang, Zichen Wen, Lei Zhang
Published: 2026-10-13
Categories: cs.CV, cs.CL, cs.MM
arXiv: https://arxiv.org/abs/2610.18829v1

With the rapid development of multimodal large language models (MLLMs), evaluating their video understanding capabilities has become increasingly important. Existing video understanding benchmarks cover only part of the evaluation dimensions, limiting systematic diagnosis of omni-modal video understanding. We introduce OmniCompass, a holistic benchmark for omni-modal video grounding and understanding. OmniCompass defines two evaluation tracks: OmniCompass-VL assesses vision-language models on tasks requiring visual evidence, while OmniCompass-AV assesses omni-modal models on tasks requiring joint audio-visual evidence. Each track includes three complementary tasks: temporal grounding (TG), spatial grounding (SG), and question answering (QA), assessing when a referenced target appears, where it is, and what happens. QA covers five perception and three reasoning categories in OmniCompass-VL, and three perception and two reasoning categories in OmniCompass-AV. OmniCompass-VL contains 820 videos and 2,460 instances across domains and three duration types: short, medium, and long. A narrative- and audio-rich subset forms OmniCompass-AV with 240 videos and 720 instances. Each instance is evaluated under two paired settings: text-only and text+image, specifying the same target through text and visual-language guidance. All instances are manually annotated with multi-round quality verification. Evaluations of 50 MLLMs show that open-source models still lag behind proprietary counterparts, particularly in audio-visual understanding. Precise temporal and spatial grounding remains challenging even for leading models, and their predictions still exhibit inconsistencies across the paired target specifications.

---

### WorldOracle: Benchmarking Action-Conditioned World Models against Real-World Executions

Authors: Haoyi Duan, Liliang Chen, Nan Wang, Yang Zhang, Yuhang Lu, Xiaotong Chen, Wei-Cheng Tseng, Shenyuan Gao, Zhaoshuo Li, Jingyi Jin, Pooya Jannaty, Dhruv Shah et al.
Published: 2026-10-13
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2610.18826v1

We introduce WorldOracle, a benchmark specifically designed to evaluate action-conditioned world models (ACWMs) against paired real-world executions. Given the same initial observation and low-level action sequence, WorldOracle asks whether a predicted rollout reproduces what happens in the real world, including physical consequences and task-state transitions. We evaluate models along four complementary dimensions: physical state correctness, interaction spatial alignment, key interaction area fidelity, and multi-view action coherence. To support this evaluation, we curate real-world training and evaluation suites that cover different interaction dynamics and embodiments. The evaluation suite contains long, multi-segment episodes with both successful and failed executions, enabling evaluation across varied action outcomes and behavior distributions. Experiments on seven representative ACWMs reveal distinct model strengths and failure modes. WorldOracle provides a systematic foundation for measuring and advancing action-conditioned world modeling for manipulation. We will release the datasets, evaluation code, and leaderboard.

---

### Joint Neural Posterior Estimation and Model Selection for Chemical Reaction Networks

Authors: J. Arturo Esquivel, Jorge A. Campos-Gonzalez-Angulo, Sergio Pablo-García, Alán Aspuru-Guzik, Vianey Leos-Barajas
Published: 2026-10-13
Categories: stat.ML, cs.LG, physics.chem-ph
arXiv: https://arxiv.org/abs/2610.18825v1

Simulation-based Bayesian inference for Chemical Reaction Networks (CRNs) must simultaneously solve two computationally demanding problems: discriminating between overlapping reaction pathways and estimating reaction rate constants. Classical approaches, such as Sequential Monte Carlo ABC, rely on handcrafted summary statistics and scale poorly across large parameter spaces. To overcome this, we introduce a neural inference architecture. Our pipeline utilizes shared 1D-CNNs and a permutation-invariant Set Attention mechanism to learn low-dimensional summary statistics from variable-length, multi-experiment time-series data. These statistics simultaneously condition a model selection network to compute posterior model probabilities and model-specific conditional Invertible Neural Networks (cINNs) that estimate parameter posterior distributions. We demonstrate the practical utility of our approach on a real protodeboronation example, illustrating how it handles structural ambiguity and extracts kinetic parameters from competing catalyst deactivation pathways.

---

### Read the Room Before Your Model Generates

Authors: Muyao Kong, Zhihao Lin
Published: 2026-10-13
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2610.18820v2

Large language models (LLMs) generally rely on sampling temperature to regulate the amount of randomness in their outputs. Mainstream methods for selecting temperature typically involve fixed, heuristic global settings and context or token-level calibration, rather than a geometrically interpretable representation of the communicative scenario. In this work, we introduce a simple and interpretable temperature selection framework, the Scenario-Actor, which adapts decoding to the inferred situational characteristics. Scenario-Actor extracts scenario-relevant structures from Transformer hidden states and determines the temperature from scenario embedding, a structured geometric representation of the communicative setting. Empirical results show that on Llama-3.1-8B-Instruct, Scenario-Actor achieves strong goal-completion performance in interactive social scenarios, significant gains on open-ended instruction following, and competitive performance on multi-turn conversation compared to multiple baselines.

---

### Read the Room Before Your Model Generates

Authors: Muyao Kong, Zhihao Lin
Published: 2026-10-13
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2610.18820v1

Large language models (LLMs) generally rely on sampling temperature to regulate the amount of randomness in their outputs. Mainstream methods for selecting temperature typically involve fixed, heuristic global settings and context or token-level calibration, rather than a geometrically interpretable representation of the communicative scenario. In this work, we introduce a simple and interpretable temperature selection framework, the \textbf{Scenario-Actor}, which adapts decoding to the inferred situational characteristics. Scenario-Actor extracts scenario-relevant structures from Transformer hidden states and determines the temperature from scenario embedding, a structured geometric representation of the communicative setting. Empirical results show that on Llama-3.1-8B-Instruct, Scenario-Actor achieves strong goal-completion performance in interactive social scenarios, significant gains on open-ended instruction following, and competitive performance on multi-turn conversation compared to multiple baselines.

---

### DuplexBridge: Coordinating Understanding and Generation for Unified Image Editing

Authors: Yidi Liu, Jie Xiao, Chaojie Mao, Jingfeng Zhang, Xueyang Fu
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18819v2

Complex image editing requires faithful task interpretation and accurate execution, but a failed edit does not reveal which responsibility failed. We present DuplexBridge, which uses a unified model's own understanding capability to enhance prompts for its own generation capability, without an external understanding module. A source-grounded bridge specifies targets, final-state changes, and preservation constraints. Understanding is supervised on fidelity to the original request without observing the generated image; generation is supervised on execution of the condition it actually receives. Reciprocal trust checks selectively admit global task feedback to each role while retaining independently valid local supervision. We introduce CE-Bench, a 1,200-task benchmark measuring joint satisfaction of task-derived change and preservation requirements. Across three model configurations, the full system improves complete compliance over direct editors by 21.75-25.75 percentage points, with gains also observed on KRIS-Bench, RISEBench, ImgEdit, and GEdit-Bench. From a common supervised Wan-Image initialization, gated joint training reaches 77.08% complete compliance, compared with 70.92% without gates and 75.00% when generation-local supervision instead checks the original task with both gates retained. These results support distinct local responsibilities and selective shared feedback for joint understanding-generation adaptation.

---

### DuplexBridge: Coordinating Understanding and Generation for Unified Image Editing

Authors: Yidi Liu, Jie Xiao, Chaojie Mao, Jingfeng Zhang, Xueyang Fu
Published: 2026-10-13
Categories: cs.CV
arXiv: https://arxiv.org/abs/2610.18819v1

Complex image editing requires faithful task interpretation and accurate execution, but a failed edit does not reveal which responsibility failed. We present DuplexBridge, which uses a unified model's own understanding capability to enhance prompts for its own generation capability, without an external understanding module. A source-grounded bridge specifies targets, final-state changes, and preservation constraints. Understanding is supervised on fidelity to the original request without observing the generated image; generation is supervised on execution of the condition it actually receives. Reciprocal trust checks selectively admit global task feedback to each role while retaining independently valid local supervision. We introduce CE-Bench, a 1,200-task benchmark measuring joint satisfaction of task-derived change and preservation requirements. Across three model configurations, the full system improves complete compliance over direct editors by 21.75-25.75 percentage points, with gains also observed on KRIS-Bench, RISEBench, ImgEdit, and GEdit-Bench. From a common supervised Wan-Image initialization, gated joint training reaches 77.08% complete compliance, compared with 70.92% without gates and 75.00% when generation-local supervision instead checks the original task with both gates retained. These results support distinct local responsibilities and selective shared feedback for joint understanding-generation adaptation.

---

### Back-Translation Still Matters: Leveraging Monolingual Data for Low-Resource Machine Translation with LLMs

Authors: Divyanshu Bathla, Parag Singla, Jitin Singla
Published: 2026-10-13
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2610.18818v2

Pre-trained Large Language Model (LLMs) are effective on many tasks in zero-shot and few-shot settings. For machine translation of low-resource languages, the obvious options are few-shot prompting and fine-tuning on a small parallel set. We ask: what if moderate amounts of low-resource monolingual data are also available? We argue that, rather than continued pretraining, a more effective use of this data is the pre-LLM era technique of back-translation (BT). The current fine-tuned LLM translates monolingual sentences in each language, and the resulting pairs train it in the reverse direction. Updates on the parallel set are interleaved for stable learning. With 1K parallel pairs on Odia and Swahili, BT beats continued pretraining into the low-resource language by 7.7 to 16.4 XCOMET-XXL points in domain. With beam-decoded intermediate translations, it also beats a model with 10 time more parallel data. In-context examples retrieved from the parallel data further improve BT. We also analyze the source, quality and amount of monolingual text. Across five low-resource languages, with Hindi as a high-resource control, and two 3B backbones, BT improves translation into the low-resource language by 16 to 173% XCOMET-XXL over parallel-only fine-tuning. Blind human evaluation on Odia and Sanskrit agrees with the automatic metrics.

---

### Back-Translation Still Matters: Leveraging Monolingual Data for Low-Resource Machine Translation with LLMs

Authors: Divyanshu Bathla, Parag Singla, Jitin Singla
Published: 2026-10-13
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2610.18818v1

Pre-trained Large Language Model (LLMs) are effective on many tasks in zero-shot and few-shot settings. For machine translation of low-resource languages, the obvious options are few-shot prompting and fine-tuning on a small parallel set. We ask: what if moderate amounts of low-resource monolingual data are also available? We argue that, rather than continued pretraining, a more effective use of this data is the pre-LLM era technique of back-translation (BT). The current fine-tuned LLM translates monolingual sentences in each language, and the resulting pairs train it in the reverse direction. Updates on the parallel set are interleaved for stable learning. With 1K parallel pairs on Odia and Swahili, BT beats continued pretraining into the low-resource language by 7.7 to 16.4 XCOMET-XXL points in domain. With beam-decoded intermediate translations, it also beats a model with 10 time more parallel data. In-context examples retrieved from the parallel data further improve BT. We also analyze the source, quality and amount of monolingual text. Across five low-resource languages, with Hindi as a high-resource control, and two 3B backbones, BT improves translation into the low-resource language by 16 to 173% XCOMET-XXL over parallel-only fine-tuning. Blind human evaluation on Odia and Sanskrit agrees with the automatic metrics.

---

### Self-Evolving Agent Harness

Authors: Chuyang Wei, Haoxiang Guan, Minghang Zhu, Kexin Zhou, Jiyan He, Xiren Zhou, Huanhuan Chen, Yifei Shen
Published: 2026-10-13
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2610.18813v1

Harness self-evolution revises an agent's software scaffolding, including prompts, memory, workflows, and executable components, while keeping model weights fixed. We formalize a feedback-driven process in which accepted revisions persist across runs. The framework separates update targets from proposal and admission mechanisms. It distinguishes changes to interpreted state from revisions to tools and code, repository-level control logic, and plugin composition. We characterize search through evolutionary optimization and examine its reliability through continual PAC-learning. Under explicit assumptions, the analysis yields an expected generalization bound, a retention bound conditional on historical regression budgets, and a high-probability admission rule using fresh, independent validation data. We examine how persistent harness revisions affect knowledge sharing and organizational workflows, with implications for scientific reliability, physical safety, and institutional accountability. Cases from prior work illustrate model--harness co-design, distributed execution, and plugin reuse. Reliable deployment requires continual validation as task distributions and evaluators change, with independent safeguards for learning, runtime privileges, and consequential actions. We provide a comprehensive 66-page discussion of related work, theory, algorithms, evaluation, applications, and the broader impact of Harness RSI, together with continuously updated code and supplementary materials, at https://github.com/zgcagi/Harness-RSI.

---

### Statistical Vignettes on the Grokking Phenomenon

Authors: Baye Malick Gning, Elvis Dohmatob
Published: 2026-10-13
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.18811v1

In this work, we study grokking in a simple, analytically tractable setting whose rich phenomenology allows us to disentangle the effects of initialization, signal strength, weight decay, sample size, and input dimension. These factors can produce superficially similar learning curves through distinct mechanisms, complicating their interpretation. Our analysis distinguishes early fitting through nuisance features, core feature learning, and delayed generalization. We identify sharp conditions under which grokking provably occurs and derive analytic asymptotic expressions for the lengths of the generalization plateaus. Our results also characterize regimes of early generalization and failure to reach a prescribed test accuracy target. We show that the interpolation edge constitutes a distinct dynamical regime, where vanishing spectral gaps separate core learning from generalization, helping clarify the role of proximity to interpolation in existing accounts of grokking. The resulting picture is analogous to fast and slow phases in chemical kinetics and suggests that the lazy-to-rich transition discussed in the literature may reflect a similar separation of dynamical time scales. Our theoretical results are faithfully supported by extensive experiments.

---

### Depth Memory

Authors: David González-Martínez, Philipp Nazari, T. Konstantin Rusch, Shiwei Liu
Published: 2026-10-13
Categories: cs.LG
arXiv: https://arxiv.org/abs/2610.18809v1

Residual connections propagate information through a network by additively accumulating layer outputs. Recently proposed alternatives, such as Manifold-Constrained Hyper-Connections (mHC) and Attention Residuals (AttnRes), instead use more expressive depth aggregation mechanisms that improve language modeling performance at scale. In this work, we consider the mixing of representations over depth as a problem of storing and retrieving information across layers. Formalizing this perspective, we introduce Depth Memory (DM), a unified framework that describes the standard pre-layer-norm (PreLN) transformer, mHC, and AttnRes in terms of streams that carry representations through the network. One such stream, for example, could be an explicit memory stream that is queried to enrich the input to every layer, and whose output is subsequently written back into the memory for use by later layers. For AttnRes, this memory stream is a growing cache of explicit layer representations. For mHC, it is an expanded residual stream. Building on this view, we propose Linear Depth Memory (LDM), which formulates information retrieval across depth as a regression problem solved using recursive least squares. The resulting architecture propagates three streams across depth: the standard residual stream, a linear associative memory stream, and an inverse covariance stream. For high-throughput training and inference, we propose a hardware-aware implementation of LDM that reduces peak activation memory and memory traffic. LDM consistently achieves the lowest perplexities among PreLN, mHC and AttnRes baselines at the 400M and 1.5B parameter scales, alongside competitive zero-shot accuracy.

---