Finding Fantastic Experts in MoEs: A Unified Study for Expert Dropping Strategies and Observations

• URL: http://arxiv.org/abs/2504.05586v2
• Date: Thu, 10 Apr 2025 02:32:14 GMT
• Title: Finding Fantastic Experts in MoEs: A Unified Study for Expert Dropping Strategies and Observations
• Authors: Ajay Jaiswal, Jianyu Wang, Yixiao Li, Pingzhi Li, Tianlong Chen, Zhangyang Wang, Chong Wang, Ruoming Pang, Xianzhi Du,
• Abstract summary: Sparsely activated Mixture-of-Experts (SMoE) has shown promise in scaling up the learning capacity of neural networks.<n>We propose MoE Experts Compression Suite (MC-Suite) to provide a benchmark for estimating expert importance from diverse perspectives.<n>We present an experimentally validated conjecture that, during expert dropping, SMoEs' instruction-following capabilities are predominantly hurt.
• Score: 86.90549830760513
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Sparsely activated Mixture-of-Experts (SMoE) has shown promise in scaling up the learning capacity of neural networks. However, vanilla SMoEs have issues such as expert redundancy and heavy memory requirements, making them inefficient and non-scalable, especially for resource-constrained scenarios. Expert-level sparsification of SMoEs involves pruning the least important experts to address these limitations. In this work, we aim to address three questions: (1) What is the best recipe to identify the least knowledgeable subset of experts that can be dropped with minimal impact on performance? (2) How should we perform expert dropping (one-shot or iterative), and what correction measures can we undertake to minimize its drastic impact on SMoE subnetwork capabilities? (3) What capabilities of full-SMoEs are severely impacted by the removal of the least dominant experts, and how can we recover them? Firstly, we propose MoE Experts Compression Suite (MC-Suite), which is a collection of some previously explored and multiple novel recipes to provide a comprehensive benchmark for estimating expert importance from diverse perspectives, as well as unveil numerous valuable insights for SMoE experts. Secondly, unlike prior works with a one-shot expert pruning approach, we explore the benefits of iterative pruning with the re-estimation of the MC-Suite criterion. Moreover, we introduce the benefits of task-agnostic fine-tuning as a correction mechanism during iterative expert dropping, which we term MoE Lottery Subnetworks. Lastly, we present an experimentally validated conjecture that, during expert dropping, SMoEs' instruction-following capabilities are predominantly hurt, which can be restored to a robust level subject to external augmentation of instruction-following capabilities using k-shot examples and supervised fine-tuning.

Related papers

• Domain-Specific Pruning of Large Mixture-of-Experts Models with Few-shot Demonstrations [48.890534958441016]
  We investigate domain specialization and expert redundancy in large-scale MoE models. We propose a simple yet effective pruning framework, EASY-EP, to identify and retain only the most relevant experts. Our method can achieve comparable performances and $2.99times$ throughput under the same memory budget with full DeepSeek-R1 with only half the experts.
  arXiv  Detail & Related papers  (2025-04-09T11:34:06Z)
• S2MoE: Robust Sparse Mixture of Experts via Stochastic Learning [34.20340688374905]
  Sparse Mixture of Experts (SMoE) enables efficient training of large language models by routing input tokens to a select number of experts.<n>Recent studies have focused on improving the router to mitigate this problem, but existing approaches face two key limitations.<n>We propose a novel approach called Sparse Mixture of Experts via Robust Learning (S2MoE), which is a mixture of experts designed to learn from both deterministic and nondeterministic inputs.
  arXiv  Detail & Related papers  (2025-03-29T08:14:27Z)
• Convergence Rates for Softmax Gating Mixture of Experts [78.3687645289918]
  Mixture of experts (MoE) has emerged as an effective framework to advance the efficiency and scalability of machine learning models.<n>Central to the success of MoE is an adaptive softmax gating mechanism which takes responsibility for determining the relevance of each expert to a given input and then dynamically assigning experts their respective weights.<n>We perform a convergence analysis of parameter estimation and expert estimation under the MoE equipped with the standard softmax gating or its variants, including a dense-to-sparse gating and a hierarchical softmax gating.
  arXiv  Detail & Related papers  (2025-03-05T06:11:24Z)
• Beyond Parameter Count: Implicit Bias in Soft Mixture of Experts [44.09546603624385]
  We introduce a notion of expert specialization for Soft MoE. We show that when there are many small experts, the architecture is implicitly biased in a fashion that allows us to efficiently approximate the specialized expert subset.
  arXiv  Detail & Related papers  (2024-09-02T00:39:00Z)
• HoME: Hierarchy of Multi-Gate Experts for Multi-Task Learning at Kuaishou [19.113649341888532]
  We present the practical problems and the lessons learned at short-video services from Kuaishou. In industry, a widely-used multi-task framework is the Mixture-of-Experts (MoE) paradigm.
  arXiv  Detail & Related papers  (2024-08-10T04:25:48Z)
• Generalization Error Analysis for Sparse Mixture-of-Experts: A Preliminary Study [65.11303133775857]
  Mixture-of-Experts (MoE) computation amalgamates predictions from several specialized sub-models (referred to as experts) Sparse MoE selectively engages only a limited number, or even just one expert, significantly reducing overhead while empirically preserving, and sometimes even enhancing, performance.
  arXiv  Detail & Related papers  (2024-03-26T05:48:02Z)
• 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)
• 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)

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.