FLEx: Personalized Federated Learning for Mixture-of-Experts LLMs via Expert Grafting

• URL: http://arxiv.org/abs/2506.00965v2
• Date: Tue, 07 Oct 2025 05:07:07 GMT
• Title: FLEx: Personalized Federated Learning for Mixture-of-Experts LLMs via Expert Grafting
• Authors: Fan Liu, Bikang Pan, Zhongyi Wang, Xi Yao, Xiaoying Tang, Jingya Wang, Ye Shi,
• Abstract summary: Federated instruction tuning of large language models (LLMs) is challenged by significant data heterogeneity across clients.<n>We propose FLEx, a novel framework that leverages pretrained MoE-based LLMs for efficient personalization.<n>For personalization, we introduce a novel expert grafting mechanism that leverages dynamic sparsity to construct a client-specific expert from selected components of pretrained experts.
• Score: 40.23842164423827
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Federated instruction tuning of large language models (LLMs) is challenged by significant data heterogeneity across clients, demanding robust personalization. The Mixture of Experts (MoE) architecture, where experts can specialize in distinct data patterns, presents a natural architectural solution to this challenge. The inherent sparsity of the MoE architecture, achieved by selectively activating experts, poses a significant challenge to its integration with federated learning (FL). Conventional FL frameworks, designed for dense models, naively aggregate all expert parameters irrespective of their local activation patterns. This naive approach not only undermines MoE's dynamic sparsity but also risks corrupting the world knowledge within pretrained experts. To address this, we propose FLEx (Federated LLMs with Personalized Experts), a novel framework that leverages pretrained MoE-based LLMs for efficient personalization. By aggregating only the shared non-expert parameters, FLEx significantly reduces communication overhead and preserves the world knowledge stored within the frozen pretrained experts. For personalization, we introduce a novel expert grafting mechanism that leverages dynamic sparsity to construct a client-specific expert from selected components of pretrained experts, tailored to local data. This grafted expert is then fine-tuned locally alongside the gating mechanism. This joint training enables the model to learn when to leverage the shared knowledge from frozen experts and when to employ the personalized one. Evaluations on diverse, non-IID instruction tuning datasets show that FLEx consistently outperforms federated baselines on average, while demonstrating strong knowledge preservation on the knowledge-driven benchmark MMLU. Our code is available at \href{https://anonymous.4open.science/r/FLEx-8F12}{\texttt{https://anonymous.4open.science/r/FLEx-8F12}}.

Related papers

• HFedMoE: Resource-aware Heterogeneous Federated Learning with Mixture-of-Experts [26.55877320740609]
  We propose HFedMoE, a heterogeneous MoE-based FL fine-tuning framework that customizes a subset of experts to each client.<n> HFedMoE identifies the expert importance based on its contributions to fine-tuning performance.<n>It then adaptively selects a subset of experts from an information bottleneck perspective to align with each client's computing budget.
  arXiv  Detail & Related papers  (2026-01-02T05:56:11Z)
• FLEX-MoE: Federated Mixture-of-Experts with Load-balanced Expert Assignment [38.27527504479237]
  Mixture-of-Experts (MoE) models enable scalable neural networks through conditional computation.<n>Our approach introduces client-expert fitness scores that quantify the expert suitability for local datasets through training feedback.<n>Our comprehensive experiments on three different datasets demonstrate the superior performance of the proposed FLEX-MoE.
  arXiv  Detail & Related papers  (2025-12-28T20:32:13Z)
• Dropping Experts, Recombining Neurons: Retraining-Free Pruning for Sparse Mixture-of-Experts LLMs [54.95810313530111]
  DERN is a task-agnostic and retraining-free framework for expert pruning and reconstruction.<n>It improves performance by more than 5% on commonsense reasoning and MMLU benchmarks under 50% expert sparsity.
  arXiv  Detail & Related papers  (2025-09-12T16:09:39Z)
• Distilling A Universal Expert from Clustered Federated Learning [23.801864975543122]
  Clustered Federated Learning (CFL) addresses the challenges posed by non-IID data by training multiple group- or cluster-specific expert models.<n>This paper introduces a novel FL framework that distills a universal expert model from the knowledge of multiple clusters.
  arXiv  Detail & Related papers  (2025-06-25T09:44:39Z)
• PM-MOE: Mixture of Experts on Private Model Parameters for Personalized Federated Learning [14.681194790227085]
  Federated learning (FL) has gained widespread attention for its privacy-preserving and collaborative learning capabilities.<n> Personalized federated learning addresses this issue by dividing the model into a globally shared part and a locally private part.<n>We propose PM-MoE architecture, which integrates a mixture of personalized modules and an energy-based personalized modules denoising.
  arXiv  Detail & Related papers  (2025-02-01T07:20:21Z)
• Personalized Federated Fine-Tuning for LLMs via Data-Driven Heterogeneous Model Architectures [15.645254436094055]
  Federated Learning (FL) enables collaborative fine-tuning of Large Language Models without accessing raw data.<n>We propose FedAMoLE, a lightweight personalized FL framework that enables data-driven heterogeneous model architectures.<n> Experiments show that FedAMoLE improves client-side performance by an average of 5.14% compared to existing approaches.
  arXiv  Detail & Related papers  (2024-11-28T13:20:38Z)
• Personalized Hierarchical Split Federated Learning in Wireless Networks [24.664469755746463]
  We propose a personalized hierarchical split federated learning (PHSFL) algorithm that is specially designed to achieve better personalization performance.<n>We first perform extensive theoretical analysis to understand the impact of model splitting and hierarchical model aggregations on the global model.<n>Once the global model is trained, we fine-tune each client to obtain the personalized models.
  arXiv  Detail & Related papers  (2024-11-09T02:41:53Z)
• MoE++: Accelerating Mixture-of-Experts Methods with Zero-Computation Experts [63.67734699877724]
  MoE++ is a general and heterogeneous MoE framework that integrates both Feed-Forward Network(FFN) and zero-computation experts. MoE++ achieves better performance while delivering 1.1-2.1x expert forward throughput compared to a vanilla MoE model of the same size.
  arXiv  Detail & Related papers  (2024-10-09T18:01:27Z)
• Upcycling Instruction Tuning from Dense to Mixture-of-Experts via Parameter Merging [36.0133566024214]
  Upcycling Instruction Tuning (UpIT) is a data-efficient approach for tuning a dense pre-trained model into a MoE instruction model. To ensure each specialized expert in the MoE model works as expected, we select a small amount of seed data that each expert excels to pre-optimize the router.
  arXiv  Detail & Related papers  (2024-10-02T14:48:22Z)
• FedMoE: Personalized Federated Learning via Heterogeneous Mixture of Experts [4.412721048192925]
  We present FedMoE, the efficient personalized Federated Learning framework to address data heterogeneity. FedMoE is composed of two fine-tuning stages. In the first stage, FedMoE simplifies the problem by conducting a search based on observed activation patterns. In the second stage, these submodels are distributed to clients for further training and returned for server aggregating.
  arXiv  Detail & Related papers  (2024-08-21T03:16:12Z)
• Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts [49.950419707905944]
  We present Self-MoE, an approach that transforms a monolithic LLM into a compositional, modular system of self-specialized experts. Our approach leverages self-specialization, which constructs expert modules using self-generated synthetic data. Our findings highlight the critical role of modularity, the applicability of Self-MoE to multiple base LLMs, and the potential of self-improvement in achieving efficient, scalable, and adaptable systems.
  arXiv  Detail & Related papers  (2024-06-17T19:06:54Z)
• Multi-Level Additive Modeling for Structured Non-IID Federated Learning [54.53672323071204]
  We train models organized in a multi-level structure, called Multi-level Additive Models (MAM)'', for better knowledge-sharing across heterogeneous clients. In federated MAM (FeMAM), each client is assigned to at most one model per level and its personalized prediction sums up the outputs of models assigned to it across all levels. Experiments show that FeMAM surpasses existing clustered FL and personalized FL methods in various non-IID settings.
  arXiv  Detail & Related papers  (2024-05-26T07:54:53Z)
• Uni-MoE: Scaling Unified Multimodal LLMs with Mixture of Experts [54.529880848937104]
  We develop a unified MLLM with the MoE architecture, named Uni-MoE, that can handle a wide array of modalities. Specifically, it features modality-specific encoders with connectors for a unified multimodal representation. We evaluate the instruction-tuned Uni-MoE on a comprehensive set of multimodal datasets.
  arXiv  Detail & Related papers  (2024-05-18T12:16:01Z)
• MAP: Model Aggregation and Personalization in Federated Learning with Incomplete Classes [49.22075916259368]
  In some real-world applications, data samples are usually distributed on local devices. In this paper, we focus on a special kind of Non-I.I.D. scene where clients own incomplete classes. Our proposed algorithm named MAP could simultaneously achieve the aggregation and personalization goals in FL.
  arXiv  Detail & Related papers  (2024-04-14T12:22:42Z)
• Merge, Then Compress: Demystify Efficient SMoE with Hints from Its Routing Policy [84.11508381847929]
  Sparsely activated Mixture-of-Experts (SMoE) has shown promise to scale up the learning capacity of neural networks. We propose M-SMoE, which leverages routing statistics to guide expert merging. Our MC-SMoE achieves up to 80% memory and a 20% FLOPs reduction, with virtually no loss in performance.
  arXiv  Detail & Related papers  (2023-10-02T16:51:32Z)
• FedJETs: Efficient Just-In-Time Personalization with Federated Mixture of Experts [48.78037006856208]
  FedJETs is a novel solution by using a Mixture-of-Experts (MoE) framework within a Federated Learning (FL) setup. Our method leverages the diversity of the clients to train specialized experts on different subsets of classes, and a gating function to route the input to the most relevant expert(s) Our approach can improve accuracy up to 18% in state of the art FL settings, while maintaining competitive zero-shot performance.
  arXiv  Detail & Related papers  (2023-06-14T15:47:52Z)
• MoEC: Mixture of Expert Clusters [93.63738535295866]
  Sparsely Mixture of Experts (MoE) has received great interest due to its promising scaling capability with affordable computational overhead. MoE converts dense layers into sparse experts, and utilizes a gated routing network to make experts conditionally activated. However, as the number of experts grows, MoE with outrageous parameters suffers from overfitting and sparse data allocation.
  arXiv  Detail & Related papers  (2022-07-19T06:09:55Z)
• Heterogeneous Ensemble Knowledge Transfer for Training Large Models in Federated Learning [22.310090483499035]
  Federated learning (FL) enables edge-devices to collaboratively learn a model without disclosing their private data to a central aggregating server. Most existing FL algorithms require models of identical architecture to be deployed across the clients and server. We propose a novel ensemble knowledge transfer method named Fed-ET in which small models are trained on clients, and used to train a larger model at the server.
  arXiv  Detail & Related papers  (2022-04-27T05:18:32Z)
• Efficient Split-Mix Federated Learning for On-Demand and In-Situ Customization [107.72786199113183]
  Federated learning (FL) provides a distributed learning framework for multiple participants to collaborate learning without sharing raw data. In this paper, we propose a novel Split-Mix FL strategy for heterogeneous participants that, once training is done, provides in-situ customization of model sizes and robustness.
  arXiv  Detail & Related papers  (2022-03-18T04:58:34Z)
• Federated Mutual Learning [65.46254760557073]
  Federated Mutual Leaning (FML) allows clients training a generalized model collaboratively and a personalized model independently. The experiments show that FML can achieve better performance than alternatives in typical Federated learning setting.
  arXiv  Detail & Related papers  (2020-06-27T09:35:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.