Today for AI
HOT RADAR
agentHEAT 6.8°

AI CURATED HOT STORY · 11/13/2026

Tsinghua Survey Reframes Embodied AI as Autonomous Agent Systems

1 reports archived1 independent sourcesupdated 11/13/2026, 19:12:00
Synthesis & Latest Updates
1 Sources Cross-Validated

A Tsinghua team has released a survey that reframes embodied AI through an agent-centric lens, unifying end-to-end policies, multimodal foundation models and long-horizon autonomous deployment into one system picture. It places model capability, compute infrastructure, safety, data and evaluation in a single framework and maps the path from skill learning toward closed-loop cognition.

LATEST/A Tsinghua team has released a survey that reframes embodied AI through an agent-centric lens, unifying end-to-end policies, multimodal foundation models and long-horizon autonomous deployment into one system picture. It places model capability, compute infrastructure, safety, data and evaluation in a single framework and maps the path from skill learning toward closed-loop cognition.

Hourly heat curve

11 fully observed hours
Now
6.8
Peak
8.510:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. 新智元 (微信公众号)T2·72 pts

    Tsinghua Survey Reframes Embodied AI as Autonomous Agent Systems

    • Reframes embodied AI research around an agent perspective, stressing closed-loop organization of perception, decision and action over long-horizon tasks
    • Integrates model capability, compute infrastructure, safety, data and evaluation into one system picture
    • Traces the evolution from feedback control and imitation/RL skill acquisition to multimodal foundation model-driven generalist robot policies