Touch-R1 Becomes First Touch Reasoning Model Accepted at NeurIPS 2026

1 reports archived1 independent sourcesupdated 11/9/2026, 16:45:00

Synthesis & Latest Updates

1 Sources Cross-Validated

A joint team from Daimon Robotics, Xiamen University, Nanjing University, Great Bay University and Fudan University introduced Touch-R1, the first model to apply reinforcement learning to cross-sensor tactile reasoning, enabling MLLMs to override visual misjudgments with touch evidence. The work ships with the TouchReason-1M dataset and TouchReason-Bench benchmark, and has been accepted at NeurIPS 2026.

LATEST/A joint team from Daimon Robotics, Xiamen University, Nanjing University, Great Bay University and Fudan University introduced Touch-R1, the first model to apply reinforcement learning to cross-sensor tactile reasoning, enabling MLLMs to override visual misjudgments with touch evidence. The work ships with the TouchReason-1M dataset and TouchReason-Bench benchmark, and has been accepted at NeurIPS 2026.

Hourly heat curve

10 fully observed hours

Now

9.8

Peak

11.917:00

24h change

–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first

机器之心 (微信公众号)T2Nov 9, 16:45·78 pts
Touch-R1 Becomes First Touch Reasoning Model Accepted at NeurIPS 2026
- First to apply reinforcement learning to cross-sensor tactile reasoning, correcting misjudgments when vision and touch conflict
- Releases the TouchReason-1M dataset and TouchReason-Bench evaluation benchmark
- Accepted at NeurIPS 2026 with code, model and data publicly available
View Related Reports & Analysis Original Source