Predicts future egocentric video from actions for human and robot embodiments. It represents movement with camera poses and manipulation with projected skeletons, trained on a unified collection of human and robot experience.
Uses natural language to describe actions and visual changes, letting one predictor learn from simulation, robot trajectories, images, and videos. It supports cross-domain planning and can initialize robot policies.
AI Dataset2026
Provides continuous human whole-body manipulation demonstrations for robot policy training, including locomotion and coordinated two-hand actions. Motion is reconstructed from egocentric observations, with reported whole-body joint error of about 3 cm.
Connects scene understanding, long-horizon visual planning, and robot control by alternating language reasoning with video prediction. An agentic variant learns when generated visual evidence is worth using.
AI Agent Papers2026
Coordinates world creation, trajectory exploration, and Harness evolution to improve multimodal agents. It organizes tasks by shared capabilities rather than surface categories, reporting mean gains of 34 points on training benchmarks and 26 on unseen benchmarks.
Rewinds world-model trajectories to test alternative robot actions and turn outcome comparisons into action-level training signals. Real-robot data continually corrects the model; across three physical tasks, success rates beat the strongest baseline by 10.67–11.85 percentage points.
Conditions robot action generation on predicted visual and motion features, capturing physical dynamics without generating future video during control. Reports results across four simulation benchmarks.