FedTreeLoRA: Reconciling Statistical and Functional Heterogeneity in Federated LoRA Fine-Tuning
This addresses the challenge of reconciling generalization and personalization in privacy-preserving LLM fine-tuning for federated learning clients, representing an incremental improvement over existing personalized methods.
The paper tackled the problem of statistical and functional heterogeneity in federated LoRA fine-tuning for LLMs by proposing FedTreeLoRA, a framework using tree-structured aggregation for layer-wise alignment, which significantly outperformed state-of-the-art methods on NLU and NLG benchmarks.
Federated Learning (FL) with Low-Rank Adaptation (LoRA) has become a standard for privacy-preserving LLM fine-tuning. However, existing personalized methods predominantly operated under a restrictive Flat-Model Assumption: they addressed client-side \textit{statistical heterogeneity} but treated the model as a monolithic block, ignoring the \textit{functional heterogeneity} across LLM layers. We argue that these two statistical (horizontal) and functional (vertical) dimensions, are \textit{orthogonal in source yet coupled in interaction}, implying that the optimal depth of parameter sharing is functionally dependent on client similarity. To address this, we propose \textbf{FedTreeLoRA}, a framework employing tree-structured aggregation for fine-grained, layer-wise alignment. By dynamically constructing an aggregation hierarchy, FedTreeLoRA allows clients to share broad consensus on shallow `trunks' while progressively specializing on deep `branches'. Experiments on NLU and NLG benchmarks demonstrate that FedTreeLoRA significantly outperforms state-of-the-art methods by effectively reconciling generalization and personalization.