DeepLeap's DELE-w0.5 ditches video-generation pipelines for robot manipulation
jiqizhixin · x · 2026-09-13
DeepLeap releases DELE-w0.5, arguing that video-generation-based world-action models waste enormous compute predicting high-dimensional visuals just to output low-dimensional actions. DELE-w0.5 abandons that pipeline entirely: it jointly learns robot actions without future visual trajectory prediction, instead learning goal-conditioned behavior reorganization — preserving the task goal and adapting actions when the environment deviates. The architecture flows from language instruction plus observation to VLM task planning to action execution.
More from Embodied
- Walden Robotics CTO: general-purpose robot manipulation is having its 'GPT moment' — adnothing · 2026-10-30
- Gym-Anything Team Stopped at 25 Simulations Due to Resource Limits, Seeks Collaborators — wellecks · 2026-10-30
- Gym-Anything Adds Audit Agents That Try to Break Simulations With Legal Robot Actions — wellecks · 2026-10-30
- NVIDIA researcher: train robot policies with privileged sim info, improving Pi 0.5 — wellecks · 2026-10-30
- CMU Auto-Builds 25 Robot Simulations with Agents; Frontier Model Solves 0 of 25 Test Tasks — wellecks · 2026-10-30
- Leveraging Simulator Privileged Information: Three Methods to Improve Robot Models — wellecks · 2026-10-30