Associate Professor, The Chinese University of Hong Kong
7 papers at NeurIPS 2025
HiFlow is a training-free and model-agnostic framework to unlock the resolution potential of pre-trained flow models.
Imagine360 creates high-quality, immersive 360 videos from perspective video anchors.
This paper introduces a simple and scalable semi-off-policy reinforcement learning method, i.e., SOPHIA, to enhance LVLMs’ ability to perform visual slow-thinking reasoning.
A memory framework for video world models that improves long-term consistency
Hierachical Balance Packing, Efficient SFT