Reinforcement_Learning
-
Geometry-Aware Reinforcement Learning for Close-Fitting Wearable Insertion
2026-10-23 04:17:15Takuma Tsukakoshi, Elle Miller, Tamon Miyake, Namiko Saito, Yushi Wang, Shigeki Sugano, Tetsuya Ogata, Sethu VijayakumarAbstract
Robot-assisted dressing is particularly challenging due to narrow openings and articulated body parts. Even small errors in the relative pose between the body part and garment opening can cause snagging, missed insertion, or incomplete dressing. To address these challenges, we formulate dressing as a geometry-aware insertion problem that represents task-relevant spatial relationships between the body part and object opening, rather than relying on the full deformable-object state. Based on this formulation, we propose a Reinforcement Learning (RL) framework combining a staged curriculum, which progressively guides the policy from coarse alignment to insertion and final placement, with a finger-wise insertion soft gate that provides continuous feedback on individual finger insertion progress and helps suppress premature insertion. We evaluate the framework on rigid and deformable bracelet dressing tasks through component ablations, systematic variations in the initial hand pose, and closed-loop execution of the learned policies on a real robotic system. On the post-training deformable evaluation, the full method reaches 84.7 +- 1.2% success. The insertion gate accounts for the largest gain, and adding the curriculum on top of that gate reaches the top score. On a mannequin hand the same full policy succeeds on 15/20 real-robot trials.
Abstract (translated)
URL
https://arxiv.org/abs/2610.32143
PDF
-
SyncPro: Synchronous Reasoning and Responding for Proactive Streaming Video Understanding
2026-10-23 03:58:17Xueyun Tian, Bingbing Xu, Jiaming Liu, Jinpeng Yu, Yang Shi, Yu Zhao, Ruihua Huang, Yuanzhuo Wang, Huawei ShenAbstract
Proactive streaming video understanding requires assistants to track unfolding events and respond when a user's requested condition is met. To make these decisions, assistants must accumulate relevant evidence and reassess response conditions as new observations arrive. Existing work offers complementary capabilities for this process: proactive video assistants learn when to respond, while streaming video reasoning methods generate scheduled intermediate text to organize observations and support subsequent answers. Together, these capabilities provide a foundation for proactive reasoning, but how to coordinate ongoing reasoning with timely responses remains underexplored. In a direct adaptation, fixed reasoning intervals can trigger updates when nothing relevant has changed or postpone useful updates until the next scheduled step. Sharing one generation path also couples two distinct requirements: preserving evidence for future decisions and producing only the response requested by the user. Moreover, responses generated after reasoning must wait for it to finish, even when sufficient evidence is already available. To address these challenges, we introduce SyncPro, a synchronous dual-stream framework with specialized generation paths for evidence tracking and user-facing responses. Its Inner and Response branches share causal history and activate independently, allowing relevant evidence changes to trigger internal updates and responses to proceed concurrently with ongoing reasoning. We construct 189K dual-stream examples and train SyncPro through supervised fine-tuning followed by trajectory-level reinforcement learning. Experiments across three proactive benchmarks show improved response timing and more accurate answers to user requests.
Abstract (translated)
URL
https://arxiv.org/abs/2610.32132
PDF
-
TravSUITE: Traversability via Self-Supervised, Uncertainty-Aware IRL and Terrain Estimation
2026-10-23 03:56:47Samuel Triest, Amirreza Shaban, David D. Fan, Wenshan Wang, Sebastian SchererAbstract
Traversability analysis in off-road settings remains a fundamental challenge for mobile robots. Key difficulties include constructing an accurate and expressive local map from multimodal sensor data, and using the map to design traversability rules that yield desirable navigation behavior. Importantly, this system must be resilient to the limited sensing regime brought about by complex environments and high speeds. In this paper, we present TravSUITE, a traversability system suitable for high-speed navigation in off-road environments. TravSUITE consists of two major components: 1) an efficient voxel mapper that leverages visual foundation models (VFMs) to build a rich geometric-semantic local map from streams of on-board sensor data, and 2) a unified neural network that jointly predicts traversability-relevant quantities in bird's eye view (BEV), including geometry, semantics, speed and cost. Our training strategy is entirely annotation-free and self-supervised, leveraging tasks such as map inpainting and inverse reinforcement learning (IRL) to learn both map representations and traversability. We perform a thorough ablation study and comparison to state-of-the-art approaches, and the results indicate that cost learning and auxiliary inpainting each contribute significantly to planning quality, and their combination is critical for achieving state-of-the-art performance in off-road path planning. We also design a simple risk adaptation mechanism to leverage our method's uncertainty estimates at deploy-time, and demonstrate that a combination of inpainting and risk estimation can result in 80% fewer navigation errors and 5% faster autonomous traversal speeds in real-world hardware experiments.
Abstract (translated)
URL
https://arxiv.org/abs/2610.32130
PDF
-
DiscoDemo: Discovering Efficient and Diverse Robot Demonstrations for Imitation Learning
2026-10-23 02:55:17Minho Park, Kinam Kim, Donghu Kim, Byungkun Lee, Dongyoon Hwang, Yongjae Shin, Junha Hyung, Hojoon Lee, Jaegul ChooAbstract
Behavior cloning for robotic manipulation is highly sensitive to the coverage of its training data, yet collecting diverse successful demonstrations remains costly. We study reinforcement learning (RL) not as a deployable controller, but as a data generator for downstream visuomotor imitation learning. Our key premise is that multiple successful solution modes cover a broader range of successful trajectories than repeatedly collecting a single narrow solution. We introduce DiscoDemo, which trains a privileged latent-conditioned RL generator with a parallel reverse curriculum and a success-gated diversity objective to discover diverse solutions while preserving task success. The trained generator produces multiple successful closed-loop behaviors, which are safety-filtered and exported as a static dataset for downstream imitation learning. Across four manipulation tasks, DiscoDemo produces broader successful-state coverage and improves downstream visuomotor policies in both simulation and the real world. Averaged across tasks, downstream success increases to 62% from 40% for the strongest baseline in simulation, and to 78% from 39% on the real robot, respectively. These results suggest that successful solution diversity improves imitation learning for a pretrained flow-matching VLA and remains complementary to corrective data collection. Project page: this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2610.32093
PDF
-
PPO vs. SAC: A Comparative Study for Cooperative Motion Control of Dual-Arm Space Robots
2026-10-23 01:22:07Nasim Fallahiarezoodar, Zheng H. ZhuAbstract
The growing population of orbital debris has intensified interest in active debris removal and on-orbit servicing, for which dual-arm space manipulators offer improved grasping stability and load distribution relative to single-arm systems. However, coordinating two manipulators under inter-arm collision avoidance, orientation alignment, and sensing uncertainty remains challenging. This paper develops a coordinated reinforcement learning framework for cooperative motion planning of a dual-arm space manipulator in a target-reaching task with inter-arm collision-avoidance constraints. A centralized state representation combines joint-space, end-effector, and target-relative information, while a multi-objective reward function promotes target-reaching accuracy, orientation alignment, motion smoothness, synchronized task completion, and collision avoidance. Within this framework, Proximal Policy Optimization (PPO) and Soft Actor-Critic (SAC) are compared as representative on-policy and off-policy algorithms, and the learned policies are further evaluated for generalization to previously unseen target configurations. Across five random seeds, SAC substantially outperforms PPO within the considered fixed-base simulation scenarios: it reaches an 80% success threshold in every seed (median 1.6M environment steps) whereas PPO does not reach it within a 4M-step budget, and it attains a higher mean success rate (78.6% versus 7.4%) with lower position and orientation errors and tighter bimanual synchronization. Both algorithms degrade on unseen boundary targets, but SAC retains most of its performance while PPO's collapses. The results also expose a failure mode of PPO: under the two-phase shaped reward, it tends to converge to a reward-loitering local optimum rather than committing to the precision task. This behavior highlights the sensitivity of on-policy learning to reward shaping.
Abstract (translated)
URL
https://arxiv.org/abs/2610.32034
PDF
-
Kolibri: A Sovereign European Model on the Pareto Frontier
2026-10-22 21:45:10Aleph Alpha, , Moritz Althaus, Artur Baranowski, Michael Barlow, Fabien C. Y. Benureau, Felix Berkenkamp, Amrita Bhattacherjee, Lukas Bl\"ubaum, Duncan Blythe, Augustijn de Boer, Bastian Boll, Thomas F. Burns, Bj\"orn Deiseroth, Niko D\"urr, B\'eni Bal\'azs Egressy, Niklas Finken, Niklas Freymuth, Fabian Gebhart, Max Hammer, Bastian Harren, Andreas Hartel, Johann Higl, Steffen Hirschmann, Pablo Iyu Guerrero, Yasser Jadidi, Andreas Jagusch, Mitja Kleider, Markus Klein, Jonas Knupp, Murali Kondragunta, Garwin Lechner, Dominik L\^e-Kellner, Jurek Leonhardt, Jiaze Li, Sohir Maskey, Omar Mehio, Pit Neitemeier, Letitia Parcalabescu, Alejandro Petry Pacheco, Pranav Ragupathy, Martin Reinhardt, Peter Rodenkirch Nfissi, Dylan Rodriquez, Leonard Salewski, Raziye Sari, Jordan Sassoon, Ivo Schaper, Ilhan Scheer, Frank Schneider, Philipp Scholl, Lukas Schott, Eric Schreiber, Pablo Schumacher, Alessio Serra, Philipp Siedler, Prabhu Teja Sivaprasad, Mohamed Soliman, Till Speicher, Volker Stampa, Simon Thel, Serge Thilges, Irina Vidal Migall\'on, Denis Wa{\ss}mann, Samuel Weinbach, Sascha Wirges, Michael W\"uthrich, Alexei ZacharovAbstract
This report introduces Kolibri, an English-German mixture-of-experts transformer with 78.1B total parameters and 3.46B active parameters per token, released as open weights under the Apache 2.0 licence. We built Kolibri for sovereign and specialised deployment, with a particular focus on German and on regulated domains such as public administration, industry, and aerospace. To support these settings, we control the full model-development process, including data, architecture, training infrastructure, post-training, and evaluation. We train Kolibri on 24T tokens across pre-training, mid-training, and long-context extension, with German accounting for more than 20% of the data mix by bytes. To reach this share, we complement German datasets with more than 2T German tokens that we curate from the web with our own pipeline and generate synthetically. Post-training combines supervised fine-tuning on curated German and English data with reinforcement learning on more than 1.2M internally curated tasks, which cover reasoning, tool use, instruction following, code, and retrieval. The model activates 4.4% of its parameters per token, and among the open models evaluated in this report, it lies on the Pareto frontier of quality and serving cost in English and German. Kolibri builds on Kolibri Origin, an earlier iteration that validated our full pipeline at scale. Both models come from our Model Factory, which connects data, experimentation, training, evaluation, and deployment through version-controlled workflows.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31945
PDF
-
Move, don't think: sample-efficient collision avoidance with lossy proximity sensors
2026-10-22 20:41:18Carson Kohlbrenner, Niraj Pudasaini, William Xie, Naren Sivagnanadasan, Nikolaus Correll, Alessandro RonconeAbstract
Data-driven robots must learn to avoid collisions with as few demonstrations as possible to avoid collecting large datasets of destructive collision examples. We explore whole-body proximity sensors as inductive biases for learning reflexive collision-dodging behavior and find that lossy sensors with low resolution were stronger biases than high-resolution coverage. We also found that, perhaps unintuitively, anticipating when and where contact happens on the body of a robot reduced its learning sample efficiency. We investigated eight candidate representations in a reinforcement learning environment, which included latent compression, future contact anticipation, lossy corruption, and privileged state estimation. The candidate representations were tested by training over 200 humanoid robots to play a simplified version of dodgeball end-to-end under an intentionally constrained sample and compute budget.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31891
PDF
-
Train4Merge: RL or SFT---Which Yields Better Teachers for OPD-Based Model Merging?
2026-10-22 20:37:53Jingyuan Huang, Zuming Huang, Yucheng Shi, Zhongzhi Li, Xiaoming Zhai, Wei Chu, Ninghao LiuAbstract
Domain experts trained from a shared checkpoint can transfer their specialized capabilities to a single student through on-policy distillation (OPD). Existing research primarily focuses on improving this merging process, while the algorithms used to train the experts have received limited systematic comparison. We investigate which training algorithm produces teachers better suited to OPD through controlled single-teacher comparisons of supervised fine-tuning (SFT) and reinforcement learning (RL) across Agentic, Reasoning, and Perception. We further examine whether this advantage extends to multi-teacher OPD (MOPD). Teachers and students share the same Qwen3.5-9B initialization, and the two teacher types are compared at similar task performance. Our single-teacher experiments show that RL teachers yield stronger students and higher recovery of teacher performance gains across all three domains. In MOPD, the RL teacher group also yields a student with a higher three-domain mean score and higher overall teacher gain recovery than the SFT teacher group. Our further analysis shows that RL teachers undergo smaller parameter displacements from the shared initialization than SFT teachers. Teacher--student Top-K overlap also increases faster early in OPD or remains consistently higher with RL teachers than with SFT teachers. These findings support the hypothesis that RL teachers' smaller departures from the student's starting point facilitate learning through OPD, resulting in stronger students.
Abstract (translated)
URL
https://arxiv.org/abs/2609.32303
PDF
-
Effectiveness of Curriculum Learning Depends on Reward Sparsity and Competing Optima
2026-10-22 20:25:52John J. Vastola, Ann Huang, Satpreet H. Singh, Samuel J. Gershman, Kanaka RajanAbstract
While animals and people tend to learn a task more quickly and reliably when they first train on simpler versions of it, curriculum learning's effectiveness in artificial settings varies, and appears substantially greater in reinforcement learning than in supervised learning. In this work, we construct and analyze a minimal model of policy learning to better understand why. We consider an open-loop, episodic target interception task whose parameters---especially how close the agent must be to the target to receive an informative reward signal---can be chosen so that the agent either receives informative rewards frequently ('dense' rewards) or rarely ('sparse' rewards). We mathematically and numerically analyze the dynamics of REINFORCE-based policy learning in both cases. In the dense reward case, curricula have only marginal benefits; in the sparse case, curricula can be required for learning to occur at all. We find that this is especially true when the agent has the conflicting goals of minimizing effort and accumulating high reward, since this can produce distinct reward landscape optima which compete with one another. Our characterization of useful curricula in this setting indicates that they address two issues: they (i) make rewards less sparse to speed up learning, and (ii) shape the reward landscape to steer agents away from bad local optima.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31884
PDF
-
Emergent Cooperative Thermoregulation in Lunar Swarm Robotics
2026-10-22 19:57:47Anais Jeger, Harry Holt, Lennart PuckAbstract
Swarms of small robots are an attractive way to survey lunar caves and lava tubes. Missions such as NASA's CADRE will fly cooperating rovers for distributed mapping, but only in daylight. The 354 h lunar night drops the surface towards 95 K, and the radioisotope heaters that let single rovers survive it cannot be supplied at swarm scale. In this study we investigate whether a swarm can instead use its own members as thermal shielding. We train one policy, shared by all twenty robots, by reinforcement learning with an evolution-strategies estimator. The policy receives a single reward, once, at the end of the night: the mean temperature of the five coldest robots. Nothing in that reward refers to position, spacing, exposure or motion. Two behaviors nonetheless emerge. The swarm aggregates, and robots circulate between the sheltered core and the exposed rim, the same pair of behaviors with which emperor penguins survive the Antarctic winter. We measure the survival threshold, the smallest stored energy per robot at which every robot is still alive at sunrise. Across ten runs it is 155.6 +/- 1.3 pcm-h, where one pcm-h is 7.2 kJ of stored heat. That is 37% below a motionless huddle and within 13% of a physics-based lower bound, and at a budget where the motionless huddle loses six robots, every robot survives. The behavior needs no shared spatial reference and no robot-robot contact, and transfers from twenty robots to forty without retraining.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31868
PDF
-
Aligning Large Language Models with Human Feedback: Mathematical Foundations and Algorithm Design
2026-10-22 19:50:27Siliang Zeng, Luca Viano, Chenliang Li, Jiaxiang Li, Volkan Cevher, Markus Wulfmeier, Stefano Ermon, Alfredo Garcia, Mingyi HongAbstract
This article provides an introduction to the mathematical foundations and algorithmic frameworks used to align Large Language Models (LLMs) with human intentions, preferences, and values. We discuss standard alignment techniques, such as fine-tuning (SFT), reinforcement learning with human feedback (RLHF), and direct preference optimization (DPO). We also explore the theoretical underpinnings of learning from human preferences, drawing connections to inverse reinforcement learning (IRL) and discrete choice models. We present state-of-the-art algorithms in a tutorial style, discuss their advantages and limitations, and offer insights into practical implementation. Our exposition is intended to serve as a comprehensive resource for researchers and practitioners, providing both a foundational understanding of alignment methodologies and a framework for developing more robust and scalable alignment techniques.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31862
PDF
-
SEVI: Safe Evidential Real-World Reinforcement Learning
2026-10-22 19:08:36Shaekh Mohammad Shithil, Lantao LiuAbstract
Real-world reinforcement learning has enabled contact-rich manipulation tasks to be trained directly on physical robots, but unsafe interactions during online learning remain a key barrier to practical deployment. We propose Safe Evidential Real-World Reinforcement Learning (SEVI), a framework for reliability-aware safety intervention during real-world policy learning. SEVI introduces an Evidential Safety Critic that estimates both future constraint-violation risk and epistemic uncertainty in a single forward pass. During online learning, SEVI evaluates each task-policy action before execution; if the action is predicted to be unsafe or the prediction is insufficiently supported by evidence, SEVI overrides it with an offline-trained recovery policy that guides the robot toward a lower-risk region. We evaluate SEVI on two simulated manipulation tasks and a real-world USB insertion task under in-distribution and distribution-shifted conditions. Across task-learning and safety-critic baselines, SEVI achieves the strongest trajectory-level safety prediction, the highest final task success, and the lowest unsafe-event counts. SEVI also reduces unsafe events by 13.4 to 57.1 percent relative to recovery-based safe RL while preserving low-latency execution-time filtering. The project page is available at: this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2610.31829
PDF
-
Free Lunch for LLMs Confidence: The RL Critic Already Knows
2026-10-22 18:57:52Shujin Wu, Heng JiAbstract
Reliable confidence estimation is essential for trustworthy deployment of large language models (LLMs), but dedicated confidence training stage requires additional data and compute. We ask a fundamental question: what prediction problem does confidence training actually solve? Inspired by prior work on external confidence estimation, we compare self- and external estimators of comparable capacity trained on the same (question, response, correctness) triples. We show that an external model trained solely on a policy's behavioral traces can estimate the policy's confidence as accurately as the policy itself, without accessing the policy's internal representations. This points to a deeper connection with value function learning in reinforcement learning (RL): under a correctness-based outcome, we prove that confidence estimation and critic training are equivalent, both predicting the likelihood that a policy's response will be correct from its trajectory. A simple consequence follows: for LLMs trained with value-based RL, the existing critic can be reused as an off-the-shelf confidence estimator with no additional training or data. The policy remains unchanged, so its task performance is preserved, and the critic provides confidence throughout generation, including intermediate reasoning steps. Across math, code, tool calling, and general reasoning spanning a dozen domains, the reused critic is the best confidence estimator on every in-distribution split and the best-calibrated on 14 of 16, without hurting the task performance.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31820
PDF
-
PrefO2O: Zero-Shot Preference-Based Reinforcement Learning via Offline-to-Online Latent Preference Transfer
2026-10-22 18:49:10Ruiqi Wang, Dezhong Zhao, Xuesu Xiao, Erdem Biyik, Byung-Cheol Min, Guohua ChenAbstract
Preference-based reinforcement learning (PbRL) has emerged as a promising paradigm for learning robot behaviors aligned with human intent. However, its scalability is fundamentally limited by the need to collect expensive human preference labels for every new task. In this work, we present PrefO2O, an offline-to-online latent preference transfer framework that enables zero-shot PbRL on unseen target tasks without requiring any human feedback. The core idea is to learn a task-invariant latent preference space from a small amount of offline preference data on source tasks, where transferable preference structures are disentangled from task-specific variations. This learned manifold allows PrefO2O to transfer preference structures and reward priors to new tasks under distribution shift, while confidence-weighted optimal transport alignment generates reliable pseudo-preference supervision for online reward adaptation. Extensive experiments on robotic manipulation and locomotion benchmarks demonstrate that PrefO2O substantially outperforms existing preference-transfer baselines and matches or exceeds PbRL methods trained with real human feedback, while generalizing across multi-to-one and one-to-one transfer settings as well as across different robot embodiments.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31812
PDF
-
DiffDreamer: Coupling World Models with Differentiable Simulation for Agile Quadrotor Control
2026-10-22 17:41:13Fanxing Li, Yuxiang Huang, Shengyang Wang, Fangyu Sun, Shuyu Wu, Tianbao Zhang, Dexin Zuo, Yan Yufei, Danping Zou, Wenxian YuAbstract
Tasks such as quadrotor dynamic-gate racing require agents to anticipate external motion while maintaining precise high-frequency control. World-model-based reinforcement learning provides a natural solution by learning latent dynamics for future imagination, yet it often entangles environmental evolution and the robot's own dynamics in a single learned rollout, where model bias can affect policy optimization efficiency. We propose DiffDreamer, a differentiable reinforcement learning framework that tightly couples a learned world model with differentiable simulation for actor optimization. In one optimization step, the learned world model propagates latent states that capture external dynamics and history-dependent features, while the differentiable simulator propagates the robot's explicit physical state. We further decompose the actor objective into a critic-driven pathway for non-differentiable task rewards and a simulator-driven pathway for differentiable motion rewards. Experiments on aggressive tracking and dynamic-gate racing show that DiffDreamer improves training convergence over the model-free, model-based, and differentiable-RL baselines, and final performance over these baselines and classical planning and control methods. Additional experiments assess reward decomposition and the odometry-conditioned relative-state representation through ablations. We also provide proof-of-concept real-world flight experiments for validation.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31617
PDF
-
Humanoid Mobility via Attention-based Contacts
2026-10-22 17:37:36Junzhe He, Kehan Wen, Chong Zhang, Clemens Schwarke, Zewei Zhang, Kaixian Qu, Siyuan Hu, Matthias M\"uller, Vladlen Koltun, Marco HutterAbstract
Humanoid mobility over complex terrain requires accurate foot placement and coordinated whole-body motion capable of robust adaptation. Reinforcement learning (RL) policies trained without motion references generalize well across terrains, yet eliciting effective whole-body coordination through reward design and exploration alone remains challenging. In contrast, reference-based methods yield natural motions but struggle to adapt beyond the predefined trajectories in their training data. We introduce a learning framework that bridges these two approaches by using motion data strictly as a training prior. An explicit, self-supervised, attention-based contact planner predicts future body-environment contacts, while a contact-conditioned tracker translates these plans into coordinated whole-body motion. Together, they establish contact as an interface between sensory observations and control actions. This abstraction allows the policy to internalize the contact-aware structure of locomotion rather than directly replicating kinematics. Consequently, a single policy can exploit data-driven motion priors for natural whole-body coordination while retaining the flexibility to discover environment-dependent behaviors through reinforcement learning. Building on the learned contact planner and tracker, a subsequent fine-tuning stage bootstraps their acquired capabilities to explore environment variations beyond those covered by the original motion data. Operating entirely from onboard observations without any reference inputs at deployment, this framework enables versatile humanoid mobility in challenging environments.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31608
PDF
-
Towards User Simulators That Stay On Task And Still Sound Natural
2026-10-22 16:26:48Mustafa Omer Gul, Philippe Laban, Tanya Goyal, Jennifer NevilleAbstract
Users increasingly rely on AI systems for extended, multi-turn tasks, yet evidence indicates that LLM-based system performance degrades over prolonged interaction. To evaluate and improve AI assistants in such settings, we need user simulators that faithfully follow complex user intents while exhibiting realistic conversational behavior. Existing approaches manifest a trade-off: prompted assistants can follow complex user intents but lack behavioral fidelity, while specialized user simulators better approximate conversational behavior but struggle to remain faithful to complex intents. In this paper, we propose a reinforcement learning (RL) recipe that jointly rewards task intent adherence and behavioral naturalness. A crucial contribution of the method is operationalizing naturalness with a lightweight discriminator trained in an alternating, adversarial loop with the user simulator. We release the resulting model, UserLM-2.0, which produces valid simulated conversations (i.e., meeting both task-faithfulness and naturalness criteria) 46.2% of the time, over three times the rate of the strongest prior user simulator (14.1%).
Abstract (translated)
URL
https://arxiv.org/abs/2610.31512
PDF
-
Learning to Be Monitored: Causal Bounds and Intervention-Aware Post-Training Recipes for Chain-of-Thought Monitorability
2026-10-22 16:00:17Qiang Fu, Xinyue Li, Andre Wibisono, Arman CohanAbstract
Chain-of-thought (CoT) monitoring aims to detect potentially harmful behaviors in large language models (LLMs) from their CoTs. Intervention-based evaluations measure this property by asking whether a monitor can detect the role of an intervention from the CoT when the targeted behavior occurs, typically using minimal-criterion sensitivity ($\mathsf{TPR_{mc}}$) proposed by~\citep{guan2025monitoring}. However, a target behavior observed under intervention is not necessarily caused by it, and we show that this metric can overestimate how often a monitor detects target behaviors actually caused by the intervention. Specifically, we establish $\mathsf{TPR_{mc}}$ as an upper bound on this causal detection rate and derive a complementary lower bound, which provides a tighter evaluation framework for CoT monitorability. With this framework, we study whether LLMs can be trained to make their CoTs easier to monitor. We first construct a training dataset of paired control--intervention tasks spanning multiple intervention types. Using the dataset, we design and apply three intervention-specific training methods based on off-policy self-distillation, reinforcement learning, and on-policy self-distillation to train separate variants of each base model. Across various model-monitor configurations, all three methods generally improve CoT monitorability over the corresponding base models under our extended evaluation framework. Our code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31733
PDF
-
Reinforcing intra-operative tumour localisation with weak supervision from pre-operative appearances
2026-10-22 13:33:10Sidhart Sami, Ali Sharjeel, Sri Atluri, Miao Zhao, Caroline Roney, Eirini Marouli, Shaheer U. SaeedAbstract
Limited tumour visibility on intra-operative images complicates surgical navigation, including liver hepatocellular carcinoma (HCC) targeting on computed tomography (CT) imaging. This makes labelled datasets difficult to curate for deep learning, and transferring pre-operative delineations through registration remains challenging due to intra-operative deformation and lack of paired pre- and intra-operative data. We propose to transfer tumour appearances learned from pre-operative magnetic resonance (MR) images to guide localisation on intra-operative CT through weak supervision. A classifier is trained to classify tumour presence on pre-operative images. This is applied to windows on intra-operative images, and despite the domain gap, it assigns higher tumour-presence probabilities on average to tumour-containing windows. These weak cross-domain signals guide a reinforcement learning agent to localise tumours on intra-operative images. We compare direct transfer, few-shot adaptation and synthetic transfer (trained using unpaired MR and CT), which translates CT into synthetic MR before localisation. On an external HCC cohort, synthetic transfer outperformed other tested methods and CT-only baselines. Our framework enables cross-modality tumour localisation without paired MR-CT training images.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31253
PDF
-
HiSTEVE 1.0: Harnessing Human-Interface VLA Agents for Open-World Games
2026-10-22 12:17:45Zecong Tang, Zian Guan, Yuncheng Guo, Zhanqiu Zhang, Yiwen Guo, Bo DaiAbstract
Building virtual embodied agents that operate in open-ended games like human players is a long-standing goal requiring both long-horizon autonomy and reliable native-interface control. As a procedurally generated open world, Minecraft provides a natural testbed for studying the integration of these capabilities. Existing agents typically provide either native visuomotor control without long-horizon autonomy or high-level planning executed through environment APIs or predefined skills. To bridge this gap, we propose HiSTEVE, a hierarchical VLA agent with a capability-aware agent harness. At the orchestration level, we equip the harness with structured memory for tracking progress and a VLM planner for selecting subtasks within the VLA's capabilities. To reliably determine subtask completion, we further introduce a policy-coupled task completion state detector that conditions on VLA policy context to signal candidate completion. At the control level, a single VLA executes all subtasks with vision-only perception and native keyboard-and-mouse actions. To mitigate interference between heterogeneous GUI and non-GUI interactions, we introduce regime-specific RMSNorm branches and separate mouse-movement token embeddings. We also train the VLA through world knowledge injection, imitation learning, and reinforcement learning with environment feedback to improve reliability and recovery. HiSTEVE achieves state-of-the-art atomic-task performance, scoring 85.0% on MineStudio Bench and a 37.4% average success rate on OpenHA, while also completing long-horizon tasks such as crafting an iron pickaxe from scratch like human players.
Abstract (translated)
URL
https://arxiv.org/abs/2610.31162
PDF
