From 'Understanding the World' to 'Simulating the Next Step': How World Models and VLAs Give Robots a True 'Embodied Brain'?
从“看懂世界”到“预演下一步”：世界模型与 VLA，如何让机器人拥有真正的“具身大脑”？
This report examines the convergence of VLA models and world models in robotics, highlighting deployments in pharmacy and industrial settings. It details specific technical achievements, including a model achieving a 2.6x improvement in success rates on unfamiliar scenes with 30,000 hours of training data, and discusses industry standards being developed to evaluate these technologies.
The article reports on the parallel development and initial deployment of Vision-Language-Action (VLA) models and world models in Shanghai, citing specific performance metrics such as a success rate increase from 17.1% to 44.1% for a model named GE-Act 2.0 and data thresholds around 10 million hours for significant capability emergence.
Source: Chinese robotics — Robot learning research · Read original article ↗
English translation unavailable. Choose the original article to read it.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
action generation time
104
View original evidence
推演需快速：在RTX 5090上，GE-Act 2.0生成一个连续动作块约需104毫秒；拟真要精确：有的模型能生成机器人端水杯的逼真画面，杯中之水却纹丝不动；数据要丰富：据大晓机器人首席科学家陶大程与开悟世界模型研发负责人王飞判断，能力明显涌现或需“接近千万小时级”的数据。
Open source S22
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
推演需快速：在RTX 5090上，GE-Act 2.0生成一个连续动作块约需104毫秒；拟真要精确：有的模型能生成机器人端水杯的逼真画面，杯中之水却纹丝不动；数据要丰富：据大晓机器人首席科学家陶大程与开悟世界模型研发负责人王飞判断，能力明显涌现或需“接近千万小时级”的数据。
Open source S22
这款“大脑”在预训练阶段适配了17个主流机器人品牌、20种机器人构型，让同一个“大脑”可以跨本体、跨任务运行。
Open source S7
Source：Chinese robotics — Robot learning research · news.qq.com