Work

Hypergraph-Aided Dynamic Model Splitting in 6G-Enabled Collaborative Edge Computing with Meta Reinforcement Learning