LLM picks its own rules to combine models in federated learning

Original: FedReason: Large Language Models as Weight-Space Meta-Aggregators for Heterogeneous Federated Learning

Source: arxiv.org ↗