PrismWAM and RoboDoctor Release
Overview
Professor Li Yiming's team from ClearMind Intelligence has released two Physical AI research results: PrismWAM enhances robot motion generation capabilities by decoupling semantic and physical state modeling; RoboDoctor enables diagnosis and self-improvement through execution feedback. Both works explore Scaling Physics from the model and system levels. PrismWAM achieved first place in RoboTwin 2.0 evaluations, while RoboDoctor scored 57.1 in RoboDojo tests, ranking first among current comparisons.
From Machine Heart — Robotics on WeChat
Event evidence and corrections
1 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Reported quantity · Reported success rate: 83.67 percent · Basis not reported
“在没有动作数据预训练的情况下，PrismWAM 在两类场景上的 平均成功率达到 83.67%，取得官方榜单第一名 （截至 2026 年 10 月 23 日）。”
Exact source · revision 1jiqizhixin
Reported quantity · pretraining hours: 2300 hours · Basis not reported
“在论文设定中，PrismWAM 使用预训练视频模型作为初始化，并针对深度编码器进行了约 2,300 小时深度数据的预训练；但没有进行额外的动作数据预训练。”
Exact source · revision 1jiqizhixin
Reported quantity · tasks covered: 50 other · Basis not reported
“RoboTwin 2.0：在复杂操作与场景迁移中检验模型能力 RoboTwin 是双臂机器人操作领域的权威评测基准，涵盖 50 项双臂操作任务，并在背景、光照、桌面杂物、桌面高度与语言指令五个维度上进行随机化，以测试模型泛化能力的上限。”
Exact source · revision 1jiqizhixin
Evidence dependency changes (2)
- 2026-10-30T20:19:50.786Z · evidence updated · source revision 1. Evidence extraction was updated; current source attributions are shown above.
- 2026-10-30T20:19:28.722Z · evidence updated · source revision 1. Evidence extraction was updated; current source attributions are shown above.
Report timeline
Follow attributed reports and material updates.
- Machine Heart — Robotics on WeChatSignalPrismWAM and RoboDoctor: Exploring the Scaling Path of Physical AI
Professor Li Yiming's team from ClearMind Intelligence has released two Physical AI research results: PrismWAM enhances robot motion generation capabilities by decoupling semantic and physical state modeling; RoboDoctor enables diagnosis and self-improvement through execution feedback. Both works explore Scaling Physics from the model and system levels. PrismWAM achieved first place in RoboTwin 2.0 evaluations, while RoboDoctor scored 57.1 in RoboDojo tests, ranking first among current comparisons.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).