Search papers, labs, and topics across Lattice.
This paper introduces FedWeave, a novel framework that enhances federated learning for heterogeneous tasks by separating expert aggregation from router optimization. The approach addresses the challenges of cross-task interference and gradient conflicts by ensuring that experts receive pure updates while routers benefit from mixed-task observations. Experimental results demonstrate that FedWeave outperforms existing methods on a multi-task benchmark, achieving superior performance while maintaining efficient sparse inference with active experts.
Experts in federated learning can now achieve specialization without sacrificing performance across heterogeneous tasks, thanks to FedWeave's innovative asymmetric aggregation strategy.
Federated PEFT enables LLMs to collaboratively adapt to decentralized private data without sharing raw examples. However, task heterogeneity across clients can cause cross-task interference and gradient conflicts during aggregation. Federated MoE-LoRA addresses this challenge through specialized LoRA experts and conditional routing. Yet existing methods typically specialize at client granularity, implicitly assuming task-coherent clients. Our core insight is that experts need purity, namely pattern-coherent updates that preserve specialization, whereas routers need contrast, namely mixed-task observations that support expert comparison. We propose FedWeave, a framework that adopts asymmetric aggregation, separating expert aggregation from router optimization to meet these two requirements. FedWeave uses unsupervised prototype discovery to form local buckets and align them across clients, enabling prototype-level expert aggregation while retaining mixed-task client trajectories for router training. At inference, FedWeave performs sparse inference with one active expert while preserving nearly all soft-routing performance. Our theoretical analysis explains why asymmetric aggregation is advantageous: it controls expert convergence in stationarity through off-pattern contamination, identifies the consensus error induced by fragmented router trajectories, and bounds sparse-inference risk. On a heterogeneous multi-task benchmark with mainstream LLM backbones, FedWeave consistently outperforms strong baselines, while ablations verify the effectiveness of our design.
