← Search

Chengming Hu

2 accepted papers

2026

Fedfit: Federated dynamic pruning via Fisher Information scoring

ICML 2026poster

Cross-device Federated Learning (FL) is frequently bottlenecked by the prohibitive computational and communication costs of training deep neural networks on resource-constrained edge hardware. While federated dynamic pruning aims to alleviate these costs by adjusting sparse topologies during trainin…

Cited by 0SourceScholar
2024

Less or More From Teacher: Exploiting Trilateral Geometry For Knowledge Distillation

ICLR 2024poster

Knowledge distillation aims to train a compact student network using soft supervision from a larger teacher network and hard supervision from ground truths. However, determining an optimal knowledge fusion ratio that balances these supervisory signals remains challenging. Prior methods generally res…

Cited by 3SourcePDFScholar