Stanford's Transformer Transformer Auto-Designs Robot Morphologies

Stanford researchers unveiled Transformer Transformer, a co-design method that generates mobile manipulation robot morphologies conditioned on diverse motion datasets such as UMI. The paper has been accepted to CoRL 2026, with project website, video, and paper released.

2026-09-05 ~ 2026-09-06 · 3 related posts

• Transformer Transformer: design a full robot from one manipulation demo, 73% tracking error cut — SongShuran · 2026-09-05
• Transformer Transformer robot co-optimization paper accepted at CoRL 2026 — SongShuran · 2026-09-06
• Transformer Transformer: feed an UMI dataset, get an optimized mobile manipulator — SongShuran · 2026-09-06