1 paper
Emilio Paolini, Andrea Pinto, Flavio Esposito +1
Federated LLM fine-tuning enables large models to be adapted using private and geographically distributed data at the network edge, creating recurring and deadline-sensitive commun…