Learning the Task or Learning the Verifier? Rethinking Criteria Exposure in Embodied Self-Improvement Paper • 2611.08318 • Published 12 days ago • 22
PhysLoop: Harnessing LLMs' Physics Awareness for Self-Improving Video Generation Models Paper • 2611.16748 • Published 5 days ago • 42
Rethinking World-Action Model for Compositional and In-Context Robotic Manipulation Paper • 2610.02368 • Published Oct 1 • 55
PyRUA-DataFarm: LLM Agent as a Cost-Efficient Teleoperator for Real-Ready Robot Data Paper • 2611.05320 • Published 13 days ago • 29
CausalAlign: Causal Alignment of Video Generation Models for Consistent Ego-Motion Paper • 2610.29915 • Published 27 days ago • 13
Where is Physical RSI capability from? An Observational Scaling Law of LLM-as-policy Paper • 2610.26848 • Published 29 days ago • 21
Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success Rate but 65% Fewer Tokens Paper • 2610.01939 • Published Oct 13 • 24
Systematically Exploring the Capabilities of GPT-6 Astra as Embodied Policies Paper • 2609.38537 • Published Sep 29 • 20
HumanTracker: Towards Comprehensive and Human-Aligned Motion Tracking Benchmark Paper • 2608.13555 • Published Aug 13 • 18
PhysisForcing: Physics Reinforced World Simulator for Robotic Manipulation Paper • 2606.28128 • Published Jun 26 • 51
ImageWAM: Do World Action Models Really Need Video Generation, or Just Image Editing? Paper • 2606.19531 • Published Jun 17 • 25
Humanoid-GPT: Scaling Data and Structure for Zero-Shot Motion Tracking Paper • 2606.03985 • Published Jun 2 • 42