Hui Chen

MaskMoE improves token learning in MoE models by enhancing infrequent token performance.

Research reveals a unique shock in laser-created plasmas, offering insight into space phenomena.

GazeGaussian is changing how we understand gaze estimation for better digital experiences.

BD-RIS technology enhances localization accuracy in modern communication systems.

New methods improve flaw detection in industrial products using advanced models.

PrefixKV optimizes large vision-language models for better performance and less resource use.

Learn how VTC-CLS improves multimodal AI models by managing visual data effectively.

Discover how GraphAvatar is shaping realistic digital experiences.

A groundbreaking method for identifying both known and unknown objects in real-time.