Jianhao Huang

Balancing image quality and transmission reliability, this framework achieves significant performance gains for solar-powered generative AI on satellites.

Token-level visual dependence is the key to preventing catastrophic forgetting in multimodal continual learning, enabling models to adapt without losing prior knowledge.

Smashed data bottlenecks in split learning? SL-FAC shrinks communication overhead by intelligently quantizing frequency components, boosting training efficiency.

By adapting AI model complexity to channel conditions, this work unlocks significant gains in edge inference throughput compared to fixed-complexity approaches.