Plans Are Not Executions: Adapting Frozen World-Model Planners to Controller-Induced Execution Shift

Follow a connection

Related by topics and source metadata.

1. 1

   Cross-Embodiment Causal Transfer from Human Videos to Robot Manipulation

   paper

   Embodied robotics and world modelsXingyu Zhang, Yingchao Li, Peizheng Guo, Jie Hu, Changwen Zheng, Wenwen QiangOctober 15, 2026

   Scaling robot manipulation models requires large amounts of interaction data, yet collecting diverse real-robot experience remains costly. Human manipulation videos, in contrast, provide abundant and diverse task experience that could potentially benefit robot learning. However, differences in morphology, action spaces, and control mechanisms make human motions unsuitable as direct supervision for robot actions, raising a scientific question: what knowledge in human videos should be transferred across embodiments to robot manipulation? To address this question, we analyze subtask transfer thro

   Original source ↗Frontier AI
2. 2

   VeriVLA: Adaptive Test-Time Verification for Flow-Based VLAs with Latent World Models

   paper

   Embodied robotics and world modelsXueli Sun, Yuxiang Lu, Yantong Liu, Xianzhe Fan, Xi Chen, Zhe Liu, Hengshuang ZhaoOctober 15, 2026

   Flow-based Vision-Language-Action (VLA) policies are inherently stochastic: different noise seeds can produce action chunks of varying quality from the same observation. At test time, however, the policy typically commits to a single sample without an explicit estimate of its quality. In fine-grained manipulation, executing a suboptimal action can irreversibly disrupt task progress and lead to failure. To exploit policy stochasticity, we introduce a training-free two-axis trigger that uses two initial samples to measure within-chunk temporal variation and cross-sample variance, detecting the c

   Original source ↗Frontier AI
3. 3

   VP-EVT: Active Visibility Preservation for Embodied Visual Tracking

   paper

   Embodied robotics and world modelsYanci Wen, Tianle Zeng, Jianwei Peng, Weixi Situ, Hanjing Ye, Hong ZhangOctober 15, 2026

   Embodied Visual Tracking (EVT) is an essential capability for intelligent robots in human-robot collaboration, requiring an agent to follow a designated target person in dynamic environments relying solely on egocentric visual perception. However, persistent pedestrian occlusions in dynamic scenes pose critical challenges to existing EVT methods. Lacking proactive visibility preservation, reactive tracking methods frequently experience prolonged target loss under severe occlusions. Therefore, we propose VP-EVT, a framework that treats target visibility as an explicit objective of motion contro

   Original source ↗Frontier AI
4. 4

   GeoState: Learning View-Decoupled Geometry States for 3D World Modeling

   paper

   Embodied robotics and world modelsHongchang Liu, Jian Liu, Yujuan Ding, Liang Peng, Xun Jiang, Fei Ma, Xing Xu, Yi BinOctober 15, 2026

   Recent 3D world models increasingly use features from geometry foundation models (GFMs) as predictive world states, benefiting from their rich geometric priors. However, the view dependence of these features entangles scene dynamics with camera-induced variation, making the resulting state evolution harder to predict. We propose GeoState, a view-decoupled geometry representation for predictive 3D world modeling that separates scene-state evolution from camera-dependent observation. Specifically, a scene tokenizer aggregates multi-view GFM features into a view-decoupled scene state, which a geo

   Original source ↗Frontier AI
5. 5

   T2-z: Test-Time Optimization of Environment Latents for Robotic World Models

   paper

   Embodied robotics and world modelsChenyuan Zhou, Yining Hong, Huang Huang, Yunfan Jiang, Wenlong Huang, David Held, Vitor Campagnolo Guizilini, Yue Wang, et al.October 15, 2026

   Predicting how the world responds to an action requires more than what can be seen: robots must account for environmental factors that are neither directly observable nor explicitly specified, from physical properties such as friction and mass to human-defined environmental rules. We introduce T2-Z for test-time optimization of environment latents, which augments a robotic world model with a learnable environment latent to condition world predictions. This compact latent captures underlying environmental variation without supervision of what environmental factors exist or what values they take

   Original source ↗Frontier AI
6. 6

   Push the Limits: Human‐Like Multimodal Extreme Embodied Intelligence for Autonomous Racing

   paper

   Embodied robot learning — abstract watchYanxin Zhou, Baichuan Lou, Wenhui Huang, Binbin Hu, Haohan Yang, Haochen Liu, Xiangkun He, Chee Kiong Ong, et al.October 22, 2026

   Autonomous racing serves as a platform for testing embodied intelligence under extreme control conditions. It requires systems to operate at the limits of speed, actuation, and decision‐making. Inspired by human multimodality, which is the ability to integrate sensory inputs, learn adaptively, and transfer knowledge across domains for holistic situational awareness and optimal performance, we present Apex, an embodied intelligent agent that is implemented on an autonomous super‐formula racing vehicle on the Abu Dhabi Yas Marina F1 circuit. Apex integrates physics‐informed neural network‐enhanc

   Original source ↗Frontier AI
7. 7

   Neostatic Bioframe BF-07 - Feasible Architecture for Embodied Intelligence - 2026 Audit

   paper

   Embodied robot learning — abstract watchRussell SaundersOctober 22, 2026

   Neostatic Bioframe BF-07 is a feasible modular architecture for embodied AI with 8 integrated subsystems: vessel architecture, Neostatic Brain Jelly (NBJ) field computer, sensory ASP skin, somatic emotional engine, distributed energy harvesting, self-healing nanobot mesh with metabolism-as-filtering, locomotion, and feasibility roadmap. 2026 Audit Updates:- VT MADE Lab GaIn-PDMS printable self-healing composite membrane- Zwitterionic bimodal elastomer P(BA-SB) 272% strain / 95 MV/m breakdown / 1520 W/kg (Nature Communications 2026 DOI 10.1038/s41467-026-72611-5)- Tesla US Pat App US 2026/03102

   Original source ↗Frontier AISoftware engineering
8. 8

   Bridging cognition, reasoning and action for polymorphic embodied agents towards human-heterogeneous multi-robot collaborative assembly with vision-language model and digital twin

   paper

   Embodied robot learning — abstract watchChangchun Liu, Dunbing Tang, Haihua Zhu, Qixiang Cai, Liping Wang, Qingwei NieOctober 22, 2026

   With the advent of Industry 5.0, human-robot collaborative assembly has become a core production paradigm that combines human cognitive flexibility with robotic operational efficiency. Single-robot systems cannot meet the diverse requirements of complex assembly tasks. Therefore, collaboration between human operators and heterogeneous multi-robot teams is required for flexible and efficient manufacturing. However, current collaborative systems are limited by fragmented perception-reasoning-execution pipelines. Poor generalization of general-purpose models in industrial environments is also obs

   Original source ↗Frontier AI
9. 9

   Gradient-interlayer–based flexible pressure sensors for high-performance robotic perception

   paper

   Embodied robot learning — abstract watchMingyuan Sun, G L, Shuai Dong, Erlong Wang, Shi‐Yang Tang, Xuanhan Chen, Yipu Guo, Xiaoqin Sun, et al.October 23, 2026

   Tactile perception systems based on flexible pressure sensors are indispensable for humanoid robots to perform dexterous manipulation, compliant interaction, and operation in visually challenging environments. Although strategies relying on surface or spatial microstructure engineering have improved sensor performance, the long-standing trade-off among sensitivity, sensing range, and linearity continues to hinder their practical deployment. Inspired by the gradient distribution of mechanoreceptors in human fingertip skin, we report a gradient-interlayer–based flexible iontronic pressure sensor

   Original source ↗Frontier AISoftware engineering
10. 10

    Education Attack

    paper

    Embodied robot learning — abstract watchJincheng ZhangOctober 23, 2026

    Artificial intelligence (AI), generative artificial intelligence (GenAI), large language models (LLMs), autonomous agents, robotics, and embodied intelligence are transforming the capabilities available to students, teachers, and educational institutions. These technologies can improve access to information, personalize instruction, automate routine tasks, and support educational creativity. However, they also introduce a broader class of risks that cannot be fully characterized by conventional cybersecurity, academic integrity, or educational technology frameworks. Students may obtain answers

    Original source ↗Frontier AI
11. 11

    EBM-Nav: energy-based model for zero-shot visual navigation via implicit scene representation

    paper

    Embodied robot learning — abstract watchMing Tang, Juan Wan, Xuan DangOctober 23, 2026

    Abstract Zero-shot visual navigation enables agents to generalize to unseen environments without task-specific training and is core to service robotics, warehouse logistics, and autonomous exploration. Existing approaches are observed to fail to achieve genuine zero-shot generalization: explicit-map methods rely on accurate reconstruction and are sensitive to sensor noise; supervised policies memorize training-scene patterns and degrade under illumination, furniture, and layout shifts; implicit representations offer continuous, differentiable encoding but remain confined to passive reconstruct

    Original source ↗Frontier AI
12. 12

    THE NATURE OF THE SOUL (Volume 1) An Interdisciplinary Postdoctoral Inquiry into Consciousness, Selfhood, Mind, Matter, and Personal Survival

    paper

    Panpsychism and analytic idealismMajid Karimi BaghmalekiOctober 13, 2026

    GENERAL PREFACE THE NATURE OF THE SOUL An Interdisciplinary Philosophical Inquiry into Consciousness, Personhood, Identity, Agency, Meaning, and the Architecture of Becoming The question of the soul is among the oldest and most persistent questions in human intellectual history. Long before consciousness became an object of neuroscience, before psychology emerged as an independent discipline, before computation provided a language for information and artificial intelligence, human beings asked what makes a living person more than a collection of physical parts. What is the principle by which a

    Original source ↗Consciousness & philosophyFrontier AIScience & psychologySoftware engineering