Robbyant open-sources LingBot-Touch, a JEPA-based tactile foundation model for dexterous manipulation
pmttyji · reddit · 2026-11-10
Robbyant released LingBot-Touch, a JEPA-based tactile representation foundation model that bridges heterogeneous tactile signals and dexterous robotic manipulation. It uses modality-specific tokenizers and a shared Transformer, learning contact structure and temporal dynamics via latent-space prediction.
It covers three sensing modalities:
- Magnetic: three-axis tactile measurements
- Piezoresistive: scalar measurements placed in the Z channel, X/Y zeroed
- Visuotactile: decoded scalar deformation/intensity in the Z channel; raw RGB is not an input
Paper (arXiv 2611.12806), code, and model weights are all open-sourced on GitHub and Hugging Face.
Related event: Ant's Robbyant Open-Sources Tactile Foundation Model LingBot-Touch(20 posts)→
More from Embodied
- SimEX: agents bootstrap robot skills in simulation, then fold towels in the real world — chris_j_paxton · 2026-11-10
- Amazon Ring unveils $250 autonomous home security drone that patrols and streams live video — Polymarket · 2026-11-10
- NVIDIA-led study: unblocking VLM gradients in VLA training lifts manipulation progress from 10.3% to 32.9% — stepjamUK · 2026-11-10
- Google DeepMind launches Gemini Robotics 2.5 with 96% accuracy sorting Mercedes-Benz parts — GoogleDeepMind · 2026-11-10
- SlashVibe is a physical desk controller for managing multiple running AI coding agents — Scobleizer · 2026-11-10
- Path Robotics raises $300M Series E at $2.45B valuation to expand physical AI — Rewkang · 2026-11-10