T-WLA adds visual memory for fast touch, 6x decision rate over T-Rex
pliang279 · x · 2026-10-23
T-WLA from MIT/Harvard researchers introduces a visual memory mechanism connecting vision and touch for fast dexterous manipulation. On real hardware it achieves 6x T-Rex's decision rate (A6000 Pro), 30% lower finger-joint error, and 2x T-ReX's average success rate across 4 real-robot tasks.
More from Embodied
- OpenTac CAD auto-routes tactile sensor PCBs from a 3D model in 3 seconds — anand_bhattad · 2026-10-23
- Unitree's own robot conference matches top robotics gatherings in scale — micoolcho · 2026-10-23
- Robot Boxing Competition Expands to Europe Across Multiple EU Universities — lukas_m_ziegler · 2026-10-23
- Sergey Levine on Physical Intelligence: scaling robot foundation models from Pi Zero to Pi 0.7 — adityaag · 2026-10-23
- Carmack: robotics' 50Hz joint targets vs 60Hz video data is a skew worth fixing — ID_AA_Carmack · 2026-10-23
- Robot data startup Maxinsights raises $30M seed, launches tactile data glove MaxGlove — heyshrutimishra · 2026-10-23