Jiahua Dong

Ego-centric video scaling saturates agent modeling long before object interaction fidelity, revealing a fundamental bottleneck in current world model training paradigms.

Projector-level forgetting in MLLMs can be mitigated, leading to significant performance gains in continual instruction tuning without excessive parameter growth.

LEEVLA reveals that effectively guiding attention to task-critical evidence can dramatically enhance performance in vision-language-action tasks.