Embodied AI Evidence Hub具身智能证据知识库

数据工程与质量 · Data Engineering and Quality

具身数据感知误差与认知误差区别

Perception Errors vs. Cognitive Errors

更新于

30 秒结论

先读这个判断

版本说明：本轮以 15 篇可获取完整正文的论文为论证主干，逐篇核对问题、方法、结果与限制；未能取得可读全文的论文不再承担正文结论。

精读论文
15
正式证据事件
15
知识单元
EA-DATA, EA-SENSOR, EA-EVAL, EA-MODEL, ERR-EMBODIED
审计状态
当前结算 run 已通过仓库校验

v0 · 完整正文

研究综述

研究边界

版本说明：本轮以 15 篇可获取完整正文的论文为论证主干，逐篇核对问题、方法、结果与限制；未能取得可读全文的论文不再承担正文结论。

  • 时间范围：2026 年 1 月 14 日至 7 月 14 日。
  • 覆盖规模：936 条去重候选、130 篇可用全文、32 篇直接相关论文；检索同时覆盖感知可观测性、推理—动作转译、失败恢复、世界模型评估和跨本体接口。
  • 停止依据：候选量、全文量、关键维度与连续低新增轮次均已达标。文章只保留能够区分“看错”和“想错”的代表性证据。
  • 重要限制：公开论文以预印本为主，两类误差虽可实验区分，却会沿闭环相互传播。

中心判断

近半年论文对"感知误差 vs 认知误差"给出的最有解释力的切分,不是按模型模块(视觉编码器 vs 语言头),而是按第一处可证伪偏离点:

一句话:感知误差问"机器人看到/记录到的世界对不对";认知误差问"基于这个世界表征,它理解和打算做的事对不对"。

这个切分之所以成立，是因为近半年出现了能把两层实验性分开的证据——最典型的探针实验显示，扰动场景下 VLA 的视觉骨干仍保持准确的空间表征，失败瓶颈只在动作头塌缩回记忆轨迹。感知对了、行为错了，两层解耦从推测变成了可测量的事实。这是本文综合 ProbeAct: Probe-Guided Training-Free Failure Recovery in Vision-Language-Action Models 与 Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation 得出的判断。

四个派生张力

1. 感知增强 vs 动作生成:更好的"看"不自动带来更好的"做"

系统研究的第一句结论就是:enhanced perception and broader semantic coverage do not inherently guarantee better action generation。高层语义推理必须转译成 end-effector motion、image-space trajectory 等动作相关表示才有用;把显式 CoT 当动作前缀会引入 compounding errors。

同一方向上, 证明认知处理可以压缩为 潜在状态 推理而保持长程规划与失败恢复——说明"认知层"是一个可独立优化、可压缩的处理阶段,而不是感知的附属品（Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation, Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning）。

结论：感知与认知之间存在一个转译界面（语义→动作表示），大量“认知误差”实际发生在这个界面上，而不是理解本身。Fast-ThinkAct 与 Embodied CoT 研究从推理压缩和动作耦合两侧支持这一判断（Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning, Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation）。

2. 感知对了 vs 执行错了:解耦的直接证据

的 probing 证据是本次证据集里对两类误差最锋利的区分:视觉骨干维持准确空间表征、动作头塌缩(memory trap)。它的工程含义同样清晰——既然内部感知没坏,恢复就不需要外部 3D 传感器,从 hidden states 提取几何参照 + 最小动作修正即可。

反过来, 的附录提醒:dense grounding 字段(boxes、gripper 位置)本身受 detector error、calibration bias、occlusion 污染——认知层的输入质量仍受感知层制约,两层可区分但不独立（ProbeAct: Probe-Guided Training-Free Failure Recovery in Vision-Language-Action Models, Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation）。

结论：诊断顺序应当是“先用 probing 或遥测确认感知层是否保真，再归因认知层”，而不是从表象猜；ProbeAct 的干预实验直接支持这一步骤（ProbeAct: Probe-Guided Training-Free Failure Recovery in Vision-Language-Action Models）。

3. 失败恢复:两类误差最好的实验场

恢复任务天然要求区分"哪里错了"。ReCoVLA: VLM-Guided Reward Compilation for Failure Recovery in Vision-Language-Action Policies 把认知层(VLM 只推断 failure mode、recovery stage、reward 选择)与控制层(residual 纠正)显式分开,其 Limitations 把 VLM failure-classification mistakes 与 perception errors 并列为独立失败来源——认知误差第一次在系统设计里有了自己的"账户"。

ActionReasoning: Robot Action Reasoning in 3D Space with LLM for Robotic Brick Stacking 从反方向补充:假设感知已准确、让 LLM 专注 3D 动作推理确实可行且省数据,但认知层误差会跨阶段传播,需要 inter-stage verification 拦截。感知侧证据同样支持这个实验场定位:TacForeSight: Force-Guided Tactile World Model for Contact-Rich Manipulation 表明学恢复必须显式采集 recovery interaction data——数据里没有恢复,认知层就无从学习恢复。

结论：恢复失败的归因链是“感知层（状态可见吗）→数据层（恢复数据存在吗）→认知层（失败阶段判对了吗）→控制层（纠正可执行吗）”，四层各有独立证据字段。ReCoVLA、ActionReasoning 与 TacForeSight 分别提供阶段判断、动作推理和恢复数据证据（ReCoVLA: VLM-Guided Reward Compilation for Failure Recovery in Vision-Language-Action Policies, ActionReasoning: Robot Action Reasoning in 3D Space with LLM for Robotic Brick Stacking, TacForeSight: Force-Guided Tactile World Model for Contact-Rich Manipulation）。

4. 世界模型横跨两层:预测误差是感知型,评估误差是认知型

把 reactive VLA 的长程推理与误差复合问题归因于缺少显式预测结构,并把世界模型的 evaluator 角色(候选动作 推演 排序/拒绝/安全过滤)单列一节——这属于认知层验证。但同一综述也指出 pixel-based 推演 的长程误差积累是世界模型自身的感知型缺陷。 给出评估侧的判据:evaluator 可靠性取决于长程动作忠实的 推演 一致性,不是视觉真实感。

展示了纯认知型规划误差的干净样本:重建地图没错,但不对"洪水后地形"做 what-if 推理,路线在执行时不可达——感知无过错,规划有责任（World Model for Robot Learning: A Comprehensive Survey, GigaWorld-1: A Roadmap to Build World Models for Robot Policy Evaluation, Path Planning in Physically Viable World Models）。

结论：评估世界模型时要分别记账——预测保真度（感知型）与决策有效性（认知型）各自有指标，混在一个分数里会掩盖问题来源（World Model for Robot Learning: A Comprehensive Survey, GigaWorld-1: A Roadmap to Build World Models for Robot Policy Evaluation, Path Planning in Physically Viable World Models）。

误差账本:两类误差的可操作区分

判据感知误差认知误差
第一处偏离原始世界 → 状态表征状态表征 → 意图/计划/动作选择
典型证据字段关键点可见率、接触事件、时间同步误差、action-state 一致性、episode 遥测probing 保真度、failure-mode 分类正确率、子目标/阶段判断、plan feasibility、推演 一致性
快速检验换传感器/补模态后失败是否消失感知输入不变、换推理/规划头后失败是否消失
治理手段多模态冗余、同步标定审计、episode 质量反馈(Closing the Loop in Teleoperation: Episode-Level Data Quality Assessment and Feedback for High-Quality Demonstration Collection)、监督可靠性分级(τ0-WM: A Unified Video-Action World Model for Robotic Manipulation)推理-动作转译监督(Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation)、阶段验证(ActionReasoning: Robot Action Reasoning in 3D Space with LLM for Robotic Brick Stacking)、失败条件化恢复(ReCoVLA: VLM-Guided Reward Compilation for Failure Recovery in Vision-Language-Action Policies)、what-if 评估(Path Planning in Physically Viable World Models)
跨本体附注传感器配置差异动作语义差异(SPACE: Enabling Learning from Cross-Robot Data Toward Generalist Policies):recorded action 不是通用监督信号,错配会伪装成任一类误差

条件、限制与未解决问题

对后续研究的启发

感知误差与认知误差的区别，不在“哪个模块出错”，而在第一处可证伪偏离发生在世界→表征，还是表征→行动。近期文献的实质进展是让这条边界可测量：probing 能证明感知保真时行为仍错，恢复系统能把失败判断与纠正执行分账，世界模型评估能把预测保真与决策有效分账。对数据工作的含义是，误差账本要同时记录两层证据字段，并保留“感知输入不变、只换认知头”这类对照实验的可能性（ProbeAct: Probe-Guided Training-Free Failure Recovery in Vision-Language-Action Models, ReCoVLA: VLM-Guided Reward Compilation for Failure Recovery in Vision-Language-Action Policies, GigaWorld-1: A Roadmap to Build World Models for Robot Policy Evaluation）。

诊断顺序

区分两类误差时，最有效的方法是逐层提供真值。先用真实物体状态替换感知输出，若动作恢复，问题主要在感知；若仍失败，再提供人工计划，观察动作头与控制器。这样的干预能把“看错”“想错”和“做错”从同一次失败中拆开。

账本还应记录错误传播。一个轻微位姿偏差可能诱发错误子任务，错误子任务又把机器人带到更严重遮挡中。第一偏离点决定最早修复位置，传播链决定需要哪些防线。只修最后动作，可能在下一个场景重新出现同类失败。

References

Accepted evidence

去重论文引用

  1. $τ_0$-WM: A Unified Video-Action World Model for Robotic ManipulationPengfei Zhou, Shengcong Chen, Di Chen, Jiaxu Wang, Rongjun Jin, Bingwen Zhu, Yike Pan, Songen Gu et al. · 2026
  2. ActionReasoning: Robot Action Reasoning in 3D Space with LLM for Robotic Brick StackingGuangming Wang, Qizhen Ying, Yixiong Jing, Olaf Wysocki, Brian Sheil · 2026
  3. Closing the Loop in Teleoperation: Episode-Level Data Quality Assessment and Feedback for High-Quality Demonstration CollectionGokul Narayanan, Yash Shahapurkar, Melih Erdogan, Brian Zhu, Eugen Solowjow · 2026
  4. Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent PlanningChi-Pin Huang, Yunze Man, Zhiding Yu, Min-Hung Chen, Jan Kautz, Yu-Chiang Frank Wang, Fu-En Yang · 2026
  5. GigaWorld-1: A Roadmap to Build World Models for Robot Policy EvaluationGigaWorld Team, Angyuan Ma, Boyuan Wang, Bohan Li, Chaojun Ni, Guo Li, Guan Huang, Guosheng Zhao et al. · 2026
  6. HapTile: A Haptic-Informed Vision-Tactile-Language-Action Dataset for Contact-Rich Imitation LearningAmirhosein Alian, Yongqiang Zhao, Shiyi Gu, Xuyang Zhang, Zhuo Chen, Christopher E. Mower, Haitham Bou-Ammar, Shan Luo · 2026
  7. Lift3D-VLA: Lifting VLA Models to 3D Geometry and Dynamics-Aware ManipulationJiaming Liu, Qingpo Wuwu, Nuowei Han, Hao Chen, Zhuoyang Liu, Fan Fei, Yueru Jia, Chenyang Gu et al. · 2026
  8. Path Planning in Physically Viable World ModelsSu Ann Low, Cheng-Hsi Hsiao, Xingjian Li, Adam J. Thorpe, Ufuk Topcu, Krishna Kumar · 2026
  9. ProbeAct: Probe-Guided Training-Free Failure Recovery in Vision-Language-Action ModelsFan Zhang, Seongbin Park, Baharan Mirzasoleiman, Shariar Talebi, Nader Sehatbakhsh · 2026
  10. ReCoVLA: VLM-Guided Reward Compilation for Failure Recovery in Vision-Language-Action PoliciesHaodi Hu, Chung-Ta Huang, Jing Liu, Ye Wang, Kei Suzuki, Matthew Brand, Toshiaki Koike-Akino · 2026
  11. Revisiting Embodied Chain-of-Thought for Generalizable Robot ManipulationNan Sun, Yuan Zhang, Yongkun Yang, Wentao Zhao, Peiyan Li, Jun Guo, Wenxuan Song, Pengxiang Ding et al. · 2026
  12. SPACE: Enabling Learning from Cross-Robot Data Toward Generalist PoliciesHaeone Lee · 2026
  13. TacForeSight: Force-Guided Tactile World Model for Contact-Rich ManipulationYujie Zang, Yuhang Zheng, Xian Nie, Yupeng Zheng, Shuai Tian, Songen Gu, Chen Gao, Zining Wang et al. · 2026
  14. TACO: TActile World Model as a Self-COrrector forScalable VLA Post-TrainingShengbang Liu, Yueru Jia, Yuyang Yan, Jiaming Liu, Xinran Zhang, Qiuxuan Feng, Yandong Guo, Shiji Zhou et al. · 2026
  15. World Model for Robot Learning: A Comprehensive SurveyBohan Hou, Gen Li, Jindou Jia, Tuo An, Xinying Guo, Sicong Leng, Haoran Geng, Yanjie Ze et al. · 2026
完整证据附录（15 条正式证据事件）
  • Time range: 2026-01-14..2026-07-14
  • Events: 15
  • 每个事件一节,标题即锚点;trace-map 中的 event 链接跳转到这里。

ERR-PVC-READ-0009

  • Claim: 长程规划、失败后自我纠正、少样本适应属于认知层能力;显式 CoT 能提升它们但推理延迟高,认知处理可以压缩为 latent 推理而保持规划与恢复能力。
  • Stance: support | Confidence: direct
  • Paper: 2601.09708 Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning
  • Locator: 5 Conclusion
  • Evidence: 论文指出 VLA 靠动作监督擅长基本技能,但在长程规划、失败自我纠正、新场景适应上泛化差;Fast-ThinkAct 用 preference-guided 蒸馏把冗长文本推理压缩为紧凑 latent CoT,在保持 long-horizon planning、few-shot adaptation 和 failure recovery 的同时推理延迟最多降 89.3%。
  • Quote: “By distilling lengthy textual reasoning into compact latent representations via preference-guided distillation and visual trajectory alignment, our approach bridges high-level embodied reasoning with low-level action execution through reasoning-enhanced policy learning. Extensive experiments across diverse robotic manipulation and embodied reasoning benchmarks demonstrate that Fast-ThinkAct achieves strong performance with significantly reduced inference latency while enabling effective long-hor”
  • Authors: chi-pin-huang; yunze-man; zhiding-yu; et al.

ERR-PVC-READ-0010

  • Claim: 纯反应式 VLA 在复杂物理环境中仍受长时程推理、时序归因和误差累积限制，这构成引入显式预测结构的主要动机。
  • Stance: support | Confidence: direct
  • Paper: 2605.00080 World Model for Robot Learning: A Comprehensive Survey
  • Locator: 1 Introduction
  • Evidence: 引言直接将纯反应 VLA 的三类困难列为长时程推理、temporal credit assignment 与 compounding errors。
  • Quote: “Yet despite strong scaling trends ( Xiao et al. , 2025 ; Li et al. , 2025b ; Zhu et al. , 2026 ) , purely reactive VLA policies remain limited in complex physical environments, where they often struggle with long-horizon reasoning, temporal credit assignment, and robustness under compounding errors. A growing body of work argues that these limitations stem not only from insufficient action prediction capacity ( Ye et al. , 2026b ; Dang et al. , 2026 ) , but also from the lack of explicit predict”
  • Authors: bohan-hou; gen-li; jindou-jia; et al.

ERR-PVC-READ-0008

  • Claim: DQAF 不用成功/失败二值标签代替数据质量，而是联合子任务进度、动作平滑度、停顿和运动学极限生成 episode 级评估与给操作者的纠正反馈。
  • Stance: support | Confidence: direct
  • Paper: 2605.26349 Closing the Loop in Teleoperation: Episode-Level Data Quality Assessment and Feedback for High-Quality Demonstration Collection
  • Locator: Abstract (full-text section)
  • Evidence: 摘要明确列出了质量信号、结构化评估和可执行的自然语言反馈。
  • Quote: “Abstract Industrial automation is at a pivotal moment, as Physical AI is driving a transition from rigid, hand-engineered automation systems toward more flexible and adaptive systems. This shift has created a growing demand for large-scale, real-world robot demonstration data, making teleoperation an increasingly important mechanism for data collection. However, high-quality teleoperated demonstrations remain difficult to obtain in practice, as novice operators often produce episodes that are ta”
  • Authors: gokul-narayanan; yash-shahapurkar; melih-erdogan; et al.

ERR-PVC-READ-0011

  • Claim: 失败恢复可以把认知层(failure mode/recovery stage 判断与 reward 选择)与控制层(residual 纠正)分开;VLM 的失败分类错误由此成为与感知误差独立的认知误差来源。
  • Stance: support | Confidence: direct
  • Paper: 2606.09630 ReCoVLA: VLM-Guided Reward Compilation for Failure Recovery in Vision-Language-Action Policies
  • Locator: 1 Introduction
  • Evidence: ReCoVLA 用外部 VLM 只推断 failure type、recovery stage、active entities、confidence 和 reward mask,不直接生成动作;确定性 reward compiler 做实体 grounding 与 stage gates,residual policy 在冻结 VLA latents 上学纠正。Limitations 明确列出 VLM failure-classification mistakes 与 perception errors、sim-to-real mismatch 并列为失败来源。
  • Quote: “Instead, it produces a structured recovery descriptor containing the failure type, recovery stage, active entities, confidence, and reward mask.”
  • Authors: haodi-hu; chung-ta-huang; jing-liu; et al.

ERR-PVC-READ-0012

  • Claim: ProbeAct 的探针实验表明，扰动下 VLA 视觉骨干仍保留目标物空间表示，而失败集中在动作头回落到记忆的训练轨迹。
  • Stance: support | Confidence: direct
  • Paper: 2606.09740 ProbeAct: Probe-Guided Training-Free Failure Recovery in Vision-Language-Action Models
  • Locator: 1 Introduction
  • Evidence: 引言将该失败定义为 latent perception 与 motor execution 的解耦，并把瓶颈定位到过拟合的 action head。
  • Quote: “Critically, this memory trap manifests as a stark decoupling between latent perception and motor execution [ 28 , 8 , 21 ] . As corroborated by our own probing experiments, the VLA’s visual backbone successfully processes the perturbed scene, maintaining accurate spatial representations of the target object. The failure bottleneck lies exclusively within the action head: constrained by an overfitted mapping, the network collapses to the nominal trajectory of its training distribution. We systema”
  • Authors: fan-zhang; seongbin-park; baharan-mirzasoleiman; et al.

ERR-PVC-READ-0013

  • Claim: 对依赖历史地图的导航，感知重建本身可以正确，但地形物理变化仍会使原路线失效；物理可行世界模型通过介入前的 what-if 修改场景暴露这类长时程规划失败。
  • Stance: support | Confidence: direct
  • Paper: 2607.00673 Path Planning in Physically Viable World Models
  • Locator: Abstract (full-text section)
  • Evidence: 摘要对比了原始重建环境与物理修改场景下的路线可行性，并报告后者能揭示前者不可见的失败。
  • Quote: “Abstract Robots deployed in unstructured outdoor environments often plan from scene reconstructions collected before deployment because operators cannot remap large or remote sites before every mission. As a result, robots must make long-horizon planning decisions using stale maps that assume the terrain remains unchanged, even though physical changes to the environment may render previously feasible routes unsafe or unreachable at execution time. We present a physically viable world model for e”
  • Authors: su-ann-low; cheng-hsi-hsiao; xingjian-li; et al.

ERR-PVC-READ-0014

  • Claim: 数据质量必须经闭环评估定义；对机器人世界模型来说，短期视觉真实感不如长程、动作忠实的 rollout 一致性更能决定其作为策略评估器的可靠性。
  • Stance: support | Confidence: direct
  • Paper: 2607.02642 GigaWorld-1: A Roadmap to Build World Models for Robot Policy Evaluation
  • Locator: Abstract (full-text section)
  • Evidence: 论文指出真实机器人策略评估受硬件和人工监督限制，是基础模型迭代瓶颈；WMBench 用真实 teleoperation 数据和匹配 policy rollouts 构造评估，并分析 7 个视频世界模型、4 种动作表示和 324,000 余次模拟 rollout。其结论强调 evaluator 质量由长程 action-faithful rollout consistency、可迁移物理先验、动作编码、记忆和评估导向 post-training 共同决定。
  • Quote: “Using WMBench, we analyze 7 video world models, 4 action representation schemes, and over 324,000 simulated policy rollouts paired with real robot executions, further enriching our analysis with large-scale community submissions from the CVPR 2026 GigaBrain Challenge, curated synthetic trajectories, and a training videos spanning more than 12,000 hours. Our experiments deliver three core insights: evaluator quality is dominated by long-horizon, action-faithful rollout consistency rather than sho”
  • Authors: gigaworld-team; angyuan-ma; boyuan-wang; et al.

ERR-PVC-READ-0015

  • Claim: Lift3D-VLA 指出，纯 2D VLA 难以保真地表达可达性、遮挡、接触和随时间演化的几何约束，而现有 2D‑3D 转换又会损失几何保真度。
  • Stance: support | Confidence: direct
  • Paper: 2607.06564 Lift3D-VLA: Lifting VLA Models to 3D Geometry and Dynamics-Aware Manipulation
  • Locator: I Introduction
  • Evidence: 引言将操作需求归结为显式 3D 结构与时间一致性，并说明纯 2D 管线及有损的跨模态变换会削弱这些约束。
  • Quote: “Despite this progress, robotic manipulation fundamentally requires spatial reasoning in the physical world [ 57 , 82 , 12 , 13 , 59 ] : the robot must infer 3D structure, reason about geometric relationships (e.g., reachability, occlusion, and contact), and plan actions that remain temporally consistent as the geometry evolves. Purely 2D VLA pipelines often struggle to reliably capture these geometric constraints, particularly in cluttered or dynamic environments. A natural direction is to expli”
  • Authors: jiaming-liu; qingpo-wuwu; nuowei-han; et al.

ERR-PVC-READ-0005

  • Claim: ActionReasoning假设感知已由视觉算法可靠提供，将 LLM 的任务收窄为 3D 动作推理；作者认为这种解耦可降低端到端训练的数据需求。
  • Stance: conditional | Confidence: direct
  • Paper: 2602.21161 ActionReasoning: Robot Action Reasoning in 3D Space with LLM for Robotic Brick Stacking
  • Locator: II-B LLM/VLM Based Robotic Operation
  • Evidence: 相关工作段明确提出解耦视觉部件，让 LLM 在已知感知状态上做 3D 物理与动作推理。
  • Quote: “To address this, we argue for decoupling the visual component: assuming the robot already has sufficiently accurate perceptual information via computer vision algorithms, and the LLMs are asked to focus only on action reasoning. This approach significantly reduces data requirements while also leveraging the wealth of existing research in computer vision. For example, the ReKep series [ 11 , 16 ] use vision-language models to identify keypoints for robotic manipulation, which significantly reduce”
  • Authors: guangming-wang; qizhen-ying; yixiong-jing; et al.

ERR-PVC-READ-0006

  • Claim: τ0-WM 把真实机器人遥操作、UMI 式交互、人类第一视角视频与 rollout/失败轨迹合并训练，并用按模态的监督掩码处理各数据源的标注缺失。
  • Stance: conditional | Confidence: direct
  • Paper: 2606.01027 $τ_0$-WM: A Unified Video-Action World Model for Robotic Manipulation
  • Locator: Abstract (full-text section)
  • Evidence: 摘要直接列出四类交互数据和 modality-specific supervision masks。
  • Quote: “Abstract Robotic manipulation requires models that generate executable actions while anticipating and evaluating their future consequences before physical execution. We present -World Model ( -WM), a unified video-action world model that integrates policy learning, video prediction, and action evaluation within a single future-predictive framework. Built on a shared video diffusion backbone, -WM provides two complementary interfaces. First, a video action model jointly predicts future visual lat”
  • Authors: pengfei-zhou; shengcong-chen; di-chen; et al.

ERR-PVC-READ-0002

  • Claim: ERVLA 的大规模实验表明，有效的具身 CoT 需要落到末端执行器运动或图像空间轨迹等动作相关表示，而将显式 CoT 作为自回归动作前缀会在推理时累积误差。
  • Stance: conditional | Confidence: direct
  • Paper: 2606.03784 Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation
  • Locator: Abstract (full-text section)
  • Evidence: 摘要同时给出了动作相关 grounding 的有效性与 autoregressive action prefix 的 compounding-error 限制。
  • Quote: “Abstract Embodied chain-of-thought (CoT) aims to bridge linguistic reasoning with robotic control, yet its effective form and integration remain underexplored. In this paper, we revisit embodied CoT for robotic control at an unprecedented scale. We curate the largest embodied CoT corpus to date, comprising 978,743 trajectories, 226.3M samples, and 2592.5 hours of data. Through extensive experiments, we show that effective CoT must ground high-level semantic understaning in concrete linguistic ac”
  • Authors: nan-sun; yuan-zhang; yongkun-yang; et al.

ERR-PVC-READ-0003

  • Claim: HapTile 的数据质量控制在机器人控制环中同步全部模态，检查空轨迹、损坏轨迹和时间戳缺口，并验证动作—状态一致性。
  • Stance: conditional | Confidence: direct
  • Paper: 2606.04825 HapTile: A Haptic-Informed Vision-Tactile-Language-Action Dataset for Contact-Rich Imitation Learning
  • Locator: 3.2 Synchronization and Data Quality Control
  • Evidence: 数据质量段明确记录了控制环同步、时间戳缺口检查、损坏轨迹剔除和 action-state consistency 检查。
  • Quote: “All data modalities are synchronized through the robot control loop. For policy learning, actions are converted to a unified 7D end-effector delta representation (1) where are translational deltas, are rotational deltas, and is the gripper command. This decouples learning from the exact robot configuration, enabling cross-embodiment by focusing the policy on local contact adjustment from tactile feedback. Several quality checks are applied to every collected trajectory. Empty or corrupted trajec”
  • Authors: amirhosein-alian; yongqiang-zhao; shiyi-gu; et al.

ERR-PVC-READ-0004

  • Claim: 如果目标包含扰动恢复，数据集必须显式收集nominal demonstrations和recovery interaction data；否则模型很难学习接触丢失后的再建立。
  • Stance: conditional | Confidence: direct
  • Paper: 2606.11184 TacForeSight: Force-Guided Tactile World Model for Contact-Rich Manipulation
  • Locator: IV-B 2 Perturbation-Aware Evaluation
  • Evidence: TacForeSight在perturbation-aware设置中额外收集恢复示教，让外部扰动发生在执行中并要求示教者重新建立稳定接触；完整模型在三个扰动任务上报告90%、85%、85%平均完成/成功分数。
  • Quote: “Policies in this setting are trained using both nominal demonstrations and recovery interaction data.”
  • Authors: yujie-zang; yuhang-zheng; xian-nie; et al.

ERR-PVC-READ-0007

  • Claim: A recorded robot action is not a universal supervision signal: the same command can produce different motions across controllers, embodiments, hardware units, and deployment-time dynamics.
  • Stance: limit | Confidence: direct
  • Paper: 2606.24049 SPACE: Enabling Learning from Cross-Robot Data Toward Generalist Policies
  • Locator: 3.2 Inconsistency of Control Commands across Robots
  • Evidence: SPACE predicts Cartesian state deltas as a shared end-effector-space representation and uses an action adapter to convert them into robot-specific control commands, improving cross-robot and dynamics-shift robustness.
  • Quote: “Recent work has scaled robot learning by training policies on data from multiple embodiments [ 27 , 23 , 32 ] , often using the Cartesian delta action space [ 23 , 32 ] since it is less dependent on robot-specific kinematics and invariant to base-frame translation [ 18 , 14 ] . In practice, this is typically realized by predicting Cartesian delta control commands that are fed to the underlying robot controller [ 23 , 32 ] . Figure 2: Different robots (e.g., UR5 vs. Franka Research 3) require dif”
  • Authors: haeone-lee

ERR-PVC-READ-0001

  • Claim: TACO 用触觉感知世界模型联合预测未来视频和力序列，再将真实失败附近状态转成带纠正动作的想象片段，用于接触丰富任务的 VLA 后训练。
  • Stance: limit | Confidence: direct
  • Paper: 2607.02840 TACO: TActile World Model as a Self-COrrector forScalable VLA Post-Training
  • Locator: 5 Conclusion and Limitations
  • Evidence: 结论的 Recognize–Imagine–Label 回路明确连接了真实失败、视频—力联合想象与纠正动作标注。
  • Quote: “We presented TACO, a tactile-aware world-model-driven framework for scalable VLA post-training in contact-rich manipulation. Following a Recognize–Imagine–Label loop, TACO converts real-world failures into imagined corrections without repeated human intervention: a tactile-aware world model jointly denoises future video and force sequences, while a unified progress-action model recognizes failure-adjacent states and labels imagined segments with corrective actions. To incorporate this supervisio”
  • Authors: shengbang-liu; yueru-jia; yuyang-yan; et al.

References

Interactive Wiki

进入交互阅读

静态页采用默认版本；v0、v0、v0可在同一交互 Wiki 中阅读，不建立相互竞争的索引页。

用 AI 继续研究 v0 v0 v0

Method and limits

研究方法、证据边界与引用说明

本页由当前结算的证据层自动派生，不改写已结算 run。论文只有在完整非 OCR 全文、结构化论文笔记和 claim-support audit 通过后，才进入接受证据。

页面结论是对所列论文与研究窗口的综合，不等同于无限外推。引用本专题时，请同时保留本页 canonical 地址、访问日期，并在需要时回溯原论文。