paper-with-me

홈 › Papers

Modeling Expert Interactions in Sparse Mixture of Experts via Graph Structures

2025-10-18 · Minh-Khoi Nguyen-Nhat, Rachel S. Y. Teo, Laziz Abdullaev, Maurice Mok, Viet-Hoang Tran, Tan Minh Nguyen arxiv

Sparse Mixture of Experts (SMoE) has emerged as a promising solution to achieving unparalleled scalability in deep learning by decoupling model parameter count from computational cost. By activating only a small subset of parameters per sample, SMoE enables significant growth in model capacity while maintaining efficiency. However, SMoE struggles to adapt to distributional shifts, leading to reduced robustness under data contamination. In this work, we introduce SymphonySMoE, a novel family of SMoE that introduces a social graph to model interactions among experts. This graph-based structure enhances the token routing process, addressing the robustness challenges that are inherent in conventional SMoE designs. SymphonySMoE is lightweight, modular, and integrates seamlessly with existing SMoE-based models such as the XMoE and the Generalist Language Model. We provide both theoretical analysis and empirical evidence demonstrating SymphonySMoE's advantages over baseline SMoE. Extensive experiments on language modeling and visual instruction tuning validate our method's effectiveness. We further highlight the scalability of SymphonySMoE to models with 4.2 and 7.4 billion parameters, showcasing its applicability in fine-tuning tasks for large-scale systems.

📄 PDF Abstract BibTeX arXiv:2510.16411

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

AdaEnsemble: Learning Adaptively Sparse Structured Ensemble Network for Click-Through Rate Prediction

2023-01-06 · YaChen Yan, Liubo Li

Learning feature interactions is crucial to success for large-scale CTR prediction in recommender systems and Ads ranking. Researchers and practitioners extensively proposed various neural network architectures for searc…

Click-Through Rate PredictionMixture-of-ExpertsRecommendation Systems

EvoMoE: An Evolutional Mixture-of-Experts Training Framework via Dense-To-Sparse Gate

2021-12-29 · Xiaonan Nie, Xupeng Miao, Shijie Cao, Lingxiao Ma 외

Mixture-of-experts (MoE) is becoming popular due to its success in improving the model quality, especially in Transformers. By routing tokens with a sparse gate to a few experts (i.e., a small pieces of the full model), …

Language ModelingLanguage ModellingMachine TranslationMasked Language Modeling+1

Doubly Sparse: Sparse Mixture of Sparse Experts for Efficient Softmax Inference

2019-01-30 · ICLR 2019 5 · Shun Liao, Ting Chen, Tian Lin, Denny Zhou 외

Computations for the softmax function are significantly expensive when the number of output classes is large. In this paper, we present a novel softmax inference speedup method, Doubly Sparse Softmax (DS-Softmax), that l…

image-classificationImage ClassificationLanguage ModelingLanguage Modelling+2

SoftMoE: Soft Differentiable Routing for Mixture-of-Experts in LLMs

2026-06-16 · Mikołaj Zasada, Łukasz Struski, Jacek Tabor, Marcin Kurdziel arxiv

Sparse Mixture-of-Experts (MoE) architectures enable scaling LLM parameters under a fixed inference budget by activating only a small subset of experts via top-$k$ routing. While this preserves causality and suits autore…

Sparse Mixture-of-Experts for Multi-Channel Imaging: Are All Channel Interactions Required?

2025-11-21 · Sukwon Yun, Heming Yao, Burkhard Hoeckendorf, David Richmond 외 arxiv

Vision Transformers ($\text{ViTs}$) have become the backbone of vision foundation models, yet their optimization for multi-channel domains - such as cell painting or satellite imagery - remains underexplored. A key chall…