Video Prediction Policy: A Generalist Robot Policy with Predictive Visual Representations
Original research
MODEL PROFILE · Related predictive policy
Video Prediction Policy
Uses predictive visual representations from video diffusion models for robot manipulation.
Research organisationTsinghua University / collaborators
Research approachRepresentation-only
Published year2024
BackboneVideo diffusion predictive features
Overview
Video Prediction Policy (VPP) uses predictive visual representations learned by video diffusion models as features for multi-task robot control. It is grouped with representation-based predictive control, not described as full video synthesis at inference.
Source record
Core research identity and broad method were source-checked as of 2026-11-03. Independent benchmarking, commercial availability, weights and licensing have not been assessed.