English
Related papers

Related papers: MoE-Spec: Expert Budgeting for Efficient Speculati…

200 papers

Mixture of Experts (MoE) has become a mainstream architecture for building Large Language Models (LLMs) by reducing per-token computation while enabling model scaling. It can be viewed as partitioning a large Feed-Forward Network (FFN) at…

Machine Learning · Computer Science 2025-08-27 Weilin Cai , Le Qin , Shwai He , Junwei Cui , Ang Li , Jiayi Huang

Large Language Models (LLMs) achieve strong performance across many tasks but suffer from high inference latency due to autoregressive decoding. The issue is exacerbated in Large Reasoning Models (LRMs), which generate lengthy chains of…

Computation and Language · Computer Science 2026-02-05 Ximing Dong , Shaowei Wang , Dayi Lin , Boyuan Chen , Ahmed E. Hassan

LLMs have low GPU efficiency and high latency due to autoregressive decoding. Speculative decoding (SD) mitigates this using a small draft model to speculatively generate multiple tokens, which are then verified in parallel by a target…

Computation and Language · Computer Science 2026-04-21 Sungkyun Kim , Jaemin Kim , Dogyung Yoon , Jiho Shin , Junyeol Lee , Jiwon Seo

Mixture-of-Experts (MoE) models enable efficient scaling of large language models (LLMs) by activating only a subset of experts per input. However, we observe that the commonly used auxiliary load balancing loss often leads to expert…

Computation and Language · Computer Science 2026-01-27 Hongcan Guo , Haolang Lu , Guoshun Nan , Bolun Chu , Jialin Zhuang , Yuan Yang , Wenhao Che , Xinye Cao , Sicong Leng , Qimei Cui , Xudong Jiang

Speculative decoding accelerates LLM inference by drafting a tree of candidate continuations and verifying it in one target forward. Existing drafters fall into two camps with opposite weaknesses. Autoregressive drafters such as EAGLE-3…

Computation and Language · Computer Science 2026-05-22 Weijie Shi , Qiang Xu , Fan Deng , Yaguang Wu , Jiarun Liu , Yehong Xu , Hao Chen , Jia Zhu , Jiajie Xu , Xiangjun Huang , Jian Yang , Xiaofang Zhou

Mixture of Experts layers (MoEs) enable efficient scaling of language models through conditional computation. This paper presents a detailed empirical study of how autoregressive MoE language models scale in comparison with dense models in…

The growing gap between the increasing complexity of large language models (LLMs) and the limited computational budgets of edge devices poses a key challenge for efficient on-device inference, despite gradual improvements in hardware…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-11-06 Xiangchen Li , Dimitrios Spatharakis , Saeid Ghafouri , Jiakun Fan , Hans Vandierendonck , Deepu John , Bo Ji , Dimitrios Nikolopoulos

Sparse Mixture-of-Experts (SMoE) language models achieve strong capability at low per-token compute, yet deployment remains constrained by memory footprint and throughput because the full expert pool must still be stored and served.…

Machine Learning · Computer Science 2026-04-14 Zongfang Liu , Shengkun Tang , Boyang Sun , Zhiqiang Shen , Xin Yuan

Mixture-of-Experts (MoE) large language models (LLMs), which leverage dynamic routing and sparse activation to enhance efficiency and scalability, have achieved higher performance while reducing computational costs. However, these models…

Machine Learning · Computer Science 2025-05-08 Xing Hu , Zhixuan Chen , Dawei Yang , Zukang Xu , Chen Xu , Zhihang Yuan , Sifan Zhou , Jiangyong Yu

With the widespread adoption of Large Language Models (LLMs), many deep learning practitioners are looking for strategies of running these models more efficiently. One such strategy is to use sparse Mixture-of-Experts (MoE) - a type of…

Machine Learning · Computer Science 2023-12-29 Artyom Eliseev , Denis Mazur

The generation quality of large language models (LLMs) is often improved by utilizing inference-time sequence-level scaling methods (e.g., Chain-of-Thought). We introduce hyper-parallel scaling, a complementary framework that improves…

Artificial Intelligence · Computer Science 2025-10-15 Soheil Zibakhsh , Mohammad Samragh , Kumari Nishu , Lauren Hannah , Arnav Kundu , Minsik Cho

The proliferation of large language models (LLMs) has led to the adoption of Mixture-of-Experts (MoE) architectures that dynamically leverage specialized subnetworks for improved efficiency and performance. Despite their benefits, MoE…

Computation and Language · Computer Science 2024-10-28 Ruisi Cai , Yeonju Ro , Geon-Woo Kim , Peihao Wang , Babak Ehteshami Bejnordi , Aditya Akella , Zhangyang Wang

Sparse Mixture-of-Experts (MoE) models offer a powerful way to scale model size without increasing compute, as per-token FLOPs depend only on k active experts rather than the total pool of E experts. Yet, this asymmetry creates an MoE…

Machine Learning · Computer Science 2026-05-15 Linghao Jin , Chufan Shi , Huijuan Wang , Nuan Wen , Zhengzhong Liu , Eric Xing , Xuezhe Ma

Vision-Language Models (VLMs) enable powerful multimodal reasoning but suffer from slow autoregressive inference, limiting their deployment in real-time applications. We introduce Spec-LLaVA, a system that applies speculative decoding to…

Computation and Language · Computer Science 2025-09-16 Mingxiao Huo , Jiayi Zhang , Hewei Wang , Jinfeng Xu , Zheyu Chen , Huilin Tai , Yijun Chen

Mixture of experts (MoE) is a popular technique to improve capacity of Large Language Models (LLMs) with conditionally-activated parallel experts. However, serving MoE models on memory-constrained devices is challenging due to the large…

Artificial Intelligence · Computer Science 2024-05-30 Rui Kong , Yuanchun Li , Qingtian Feng , Weijun Wang , Xiaozhou Ye , Ye Ouyang , Linghe Kong , Yunxin Liu

Mixture-of-Experts (MoE) activates only a subset of experts during inference, allowing the model to maintain low inference FLOPs and latency even as the parameter count scales up. However, since MoE dynamically selects the experts, all the…

Machine Learning · Computer Science 2025-05-27 Shibo Jie , Yehui Tang , Kai Han , Yitong Li , Duyu Tang , Zhi-Hong Deng , Yunhe Wang

Speculative decoding has been shown as an effective way to accelerate Large Language Model (LLM) inference by using a Small Speculative Model (SSM) to generate candidate tokens in a so-called speculation phase, which are subsequently…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-03-21 Fahao Chen , Peng Li , Tom H. Luan , Zhou Su , Jing Deng

As large language models gain widespread adoption, running them efficiently becomes crucial. Recent works on LLM inference use speculative decoding to achieve extreme speedups. However, most of these works implicitly design their algorithms…

Computation and Language · Computer Science 2024-12-03 Ruslan Svirschevski , Avner May , Zhuoming Chen , Beidi Chen , Zhihao Jia , Max Ryabinin

Mixture-of-Experts (MoE) architectures have become the dominant choice for scaling Large Language Models (LLMs), activating only a subset of parameters per token. While MoE architectures are primarily adopted for computational efficiency,…

Computation and Language · Computer Science 2026-05-19 Jeremy Herbst , Stefan Wermter , Jae Hee Lee

Speculative decoding is a powerful way to accelerate autoregressive large language models (LLMs), but directly porting it to vision-language models (VLMs) faces unique systems constraints: the prefill stage is dominated by visual tokens…

Computer Vision and Pattern Recognition · Computer Science 2025-09-23 Haiduo Huang , Fuwei Yang , Zhenhua Liu , Xuanwu Yin , Dong Li , Pengju Ren , Emad Barsoum