The paper introduces Evidence‑Gated Regularization (EGR), a modality‑agnostic training objective that mitigates modality entanglement in Vision‑Language‑Action (VLA) policies. EGR uses per‑frame, per‑sensor task‑relevance signals to enforce invariance on low‑evidence sensors and single‑sensor sufficiency on high‑evidence ones, adding no inference‑time overhead. Evaluations on a BEHAVIOR‑1K benchmark and two real‑robot setups (bi‑manual Kinova arms with RGB cameras and a single‑arm MELFA ASSISTA with vision and GelSight tactile sensors) show significant improvements in success rates across various corruption and fallback scenarios.
arXiv:2608.25757v4 Announce Type: replace-cross Abstract: Large-scale vision--language--action (VLA) policies have advanced generalist robot control, yet most remain stimulus-to-action black boxes: a...
arXiv:2606. 26443v1 Announce Type: cross Abstract: A robot working alongside people must reason about what they have done, in what order, and with what intent.
LM‑X is a generalist vision‑language‑action policy that augments action prediction with three online, explicitly supervised signals: return‑to‑go (RTG) for task progress, event‑to‑go (ETG) for the next semantic transition, and heteroscedastic action flow for local reliability. By conditioning action generation on these signals, LM‑X embeds explainability directly into control rather than as a post‑hoc explanation. After a 20‑day pretraining run on 64 GPUs, LM‑X outperforms an action‑only backbone by 16.0 points and a single‑head variant by 10.8 points, and achieves 74.1 % success on 50 RoboTwin2.0 tasks and 68.6 % on seven real‑robot tasks, surpassing the GR00T N1.7 baseline.
INSPECT is a system that learns how a robot should choose its camera view during assembly inspection by observing a smart‑glasses assistant that answers part queries and guides the user. It uses techniques such as Presence‑Invariant TwinSwap for object evidence calibration, claim‑indexed supervision to separate evidence needs from camera changes, and object‑centered calibration to adapt view preferences to robot poses. In experiments on gearbox assemblies and angle‑grinder recordings, INSPECT outperforms other non‑oracle policies, improving view utility and decision accuracy.
arXiv:2511. 15669v3 Announce Type: replace-cross Abstract: Does Chain-of-Thought (CoT) reasoning genuinely improve Vision Language Action (VLA) models, or does it merely add overhead?
arXiv:2608. 11739v1 Announce Type: cross Abstract: The prevailing recipe for Vision-Language-Action (VLA) models couples a pretrained VLM with a separately trained flow-matching action expert.
Vision‑Language‑Action models (VLAs) often fail to follow language instructions because they rely on vision shortcuts learned from biased datasets. The authors introduce LIBERO‑CF, a counterfactual benchmark that tests language following by providing alternative task instructions for the same scene layouts, revealing widespread counterfactual failures. They propose Counterfactual Action Guidance (CAG), a dual‑branch inference scheme that amplifies language conditioning, which consistently improves language grounding and task success across diverse VLAs without requiring extra data or architectural changes.
The report evaluates the GPT-6-Astra multimodal large language model (MLLM) on the EBench platform, comparing two agent configurations: direct control, where the agent generates waypoint commands, and harness with policy, where it can invoke post‑trained vision‑language‑action (VLA) or world action model (WAM) policies. Direct control performs competitively overall, excelling on short‑horizon tasks and showing stable generalization, but falls behind on long‑horizon, tabletop, and high‑precision contact‑rich tasks. The harness with policy approach combines the strengths of both components, achieving success rates comparable to the best post‑trained policies while reducing token cost, yet still lags behind the standalone upper bounds due to suboptimal policy invocation and execution issues.
EmbodiedSkills is a skill-based agent framework that enhances frozen vision‑language‑action (VLA) policies with world action models for consequence‑aware action selection. It defines reusable skills for observation, context encoding, candidate generation, consequence prediction, outcome evaluation, candidate comparison, and action execution, organized into five stages that support both continuous action chunks and pick‑and‑place primitives. The framework improves success rates on MetaWorld MT50, CLIPort, and CALVIN benchmarks by incorporating outcome predictions and lightweight comparators while keeping the core policy and observation encoders frozen.
arXiv:2606. 02735v1 Announce Type: cross Abstract: Generalization remains a central bottleneck for vision-language-action (VLA) models: under distractors, appearance shifts, and semantically similar tasks, the policy must often infer local execution details from coarse instructions while also deciding which parts of the image matter for control.
arXiv:2609.38494v1 Announce Type: cross Abstract: Robotic manipulation integrates vision, touch, and language, whose importance shifts across stages: vision guides reaching, while touch, through its...
arXiv:2606. 14981v1 Announce Type: cross Abstract: Inference-time steering adapts pre-trained generative robot policies during deployment by verifying candidate actions before execution.