Researchers from GreatX Robot and the Chinese University of Hong Kong's MMLab have introduced ACE-Ego, a novel embodied operation VLA model. This model utilizes a "human-centric" training paradigm, efficiently combining large-scale first-person human videos with multi-robot data. ACE-Ego has achieved state-of-the-art results on two major embodied intelligence benchmarks, RoboCasa GR1 TableTop and RoboTwin 2.0, demonstrating strong generalization capabilities in complex retail scenarios. AI
IMPACT Sets new SOTA on embodied intelligence benchmarks, potentially accelerating the development and deployment of robots in complex real-world tasks.
RANK_REASON The cluster describes the release and technical details of a new embodied operation VLA model, including its training methodology and benchmark performance. [lever_c_demoted from research: ic=1 ai=1.0]
- ACE-Ego
- CUHK MMLab
- Gaode ABot-M0
- GreatX Robot
- JD JoyAI-RA
- NVIDIA GR00T
- PI π₀.₅
- RoboCasa GR1 TableTop
- RoboTwin 2.0
- Tencent Hunyuan-VLA
- Chinese University of Hong Kong
- Xiaopeng DIAL
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
