Zhi Jing

LM-X: Explainable Action Modeling with Progress, Event, and Uncertainty Prediction for Generalist Robot Manipulation

One Token Per Frame: Reconsidering Visual Bandwidth in World Models for VLA Policy

OneWM-VLA is introduced, which represents each retained camera view with one predictive token per future step, which incorporates visual prediction into a pretrained VLA policy while keeping the generative sequence compact.

RoboCousin: Build Your Own Simulation Playground for Robust Bimanual Robotic Manipulation

Simulation and real-robot experiments show that the automatically generated interaction annotations are comparable to curated annotations, generated assets provide effective sim-to-real supervision, and tabletop cousins can improve transfer beyond training on a single reconstructed scene.

LM-X: Explainable Vision--Language--Action Modeling via Progress, Event, and Uncertainty Prediction

LM-X is introduced, which organizes prediction across task, event, and motor scales without claiming anatomical correspondence and shows that explicit multi-timescale predictive state can strengthen control while exposing interpretable internal estimates.