Mixture-of-Experts Inference

Ask about this topic

Overview

Mixture-of-experts inference studies how to select and execute only the experts needed for each token or input, aiming to reduce computation without losing model quality. Methods include adaptive routing and expert skipping, quantization and pruning, predictive prefetching and caching, speculative decoding, and expert-parallel scheduling across GPUs or heterogeneous devices. Systems are evaluated in settings such as prefill versus token-by-token decoding, long-context serving, distributed GPU clusters, and memory-limited edge devices, using measures like latency, throughput, memory traffic, and task accuracy. A central tension is that activating fewer parameters does not necessarily make inference faster: moving expert weights and coordinating irregular workloads can dominate the computation, especially when memory or network bandwidth is limited.

Latest papers

CardsList

177 papers28 papers in the last four weeks, up 12% on the four weeks before. 0.2% of all new papers.

Apr 15, 2026

Awakening Dormant Experts:Counterfactual Routing to Mitigate MoE Hallucinations

Sparse Mixture-of-Experts (MoE) models have achieved remarkable scalability, yet they remain vulnerable to hallucinations, particularly when processing long-tail knowledge. We identify that this fragility stems from static Top-kk routing: routers tend to favor high-frequency patterns over rare factual associations. Consequently, specialist experts'' possessing critical long-tail knowledge are often assigned low gating scores and remain dormant'' -- under-prioritized for specific tokens despite their proven causal importance on other inputs. To address this, we propose Counterfactual Routing (CoR), a training-free inference framework designed to awaken these dormant experts. CoR integrates layer-wise perturbation analysis with the Counterfactual Expert Impact (CEI) metric to dynamically shift computational resources from syntax-dominant to knowledge-intensive layers while maintaining a constant total activation count, effectively retrieving causally decisive experts via virtual ablation. Extensive experiments on TruthfulQA, FACTOR, and TriviaQA demonstrate that CoR improves factual accuracy by 3.1% on average without increasing the inference budget, establishing a superior Pareto frontier compared to static scaling strategies.

Wentao Hu, Yanbo Zhai, Xiaohui Hu +6

Xi’an Jiaotong University · Xingchen AGI Lab, China Telecom Artificial Intelligence Technology (Beijing) Co., Ltd. · Beijing Foreign Studies University +1

LLM Hallucination MitigationMixture-of-Experts InferenceLLM RoutingMixture-of-Experts Models
AskPDF
Apr 11, 2026

SpecMoE: A Fast and Efficient Mixture-of-Experts Inference via Self-Assisted Speculative Decoding

The Mixture-of-Experts (MoE) architecture has emerged as a promising approach to mitigate the rising computational costs of large language models (LLMs) by selectively activating parameters. However, its high memory requirements and sub-optimal parameter efficiency pose significant challenges for efficient deployment. Although CPU-offloaded MoE inference systems have been proposed in the literature, they offer limited efficiency, particularly for large batch sizes. In this work, we propose SpecMoE, a memory-efficient MoE inference system based on our self-assisted speculative decoding algorithm. SpecMoE demonstrates the effectiveness of applying speculative decoding to MoE inference without requiring additional model training or fine-tuning. Our system improves inference throughput by up to 4.30×4.30\times, while significantly reducing bandwidth requirements of both memory and interconnect on memory-constrained systems.

Jehyeon Bang, Eunyeong Cho, Ranggi Hwang +2

KAIST · UNIST

Mixture-of-Experts InferenceLLM Inference AccelerationMemory-Efficient InferenceSpeculative Decoding
AskPDF
Apr 9, 2026

Alloc-MoE: Budget-Aware Expert Activation Allocation for Efficient Mixture-of-Experts Inference

Mixture-of-Experts (MoE) has become a dominant architecture for scaling large language models due to their sparse activation mechanism. However, the substantial number of expert activations creates a critical latency bottleneck during inference, especially in resource-constrained deployment scenarios. Existing approaches that reduce expert activations potentially lead to severe model performance degradation. In this work, we introduce the concept of activation budget as a constraint on the number of expert activations and propose Alloc-MoE, a unified framework that optimizes budget allocation coordinately at both the layer and token levels to minimize performance degradation. At the layer level, we introduce Alloc-L, which leverages sensitivity profiling and dynamic programming to determine the optimal allocation of expert activations across layers. At the token level, we propose Alloc-T, which dynamically redistributes activations based on routing scores, optimizing budget allocation without increasing latency. Extensive experiments across multiple MoE models demonstrate that Alloc-MoE maintains model performance under a constrained activation budget. Especially, Alloc-MoE achieves 1.15×1.15\times prefill and 1.34×1.34\times decode speedups on DeepSeek-V2-Lite at half of the original budget.

Baihui Liu, Kaiyuan Tian, Wei Wang +3

National Key Laboratory of Parallel and Distributed Computing, College of Computer Science and Technology, National University of Defense Technology

Mixture-of-Experts InferenceMixture of ExpertsEfficient Language Model Inference
AskPDF
Apr 8, 2026

MoBiE: Efficient Inference of Mixture of Binary Experts under Post-Training Quantization

Mixture-of-Experts (MoE) based large language models (LLMs) offer strong performance but suffer from high memory and computation costs. Weight binarization provides extreme efficiency, yet existing binary methods designed for dense LLMs struggle with MoE-specific issues, including cross-expert redundancy, task-agnostic importance estimation, and quantization-induced routing shifts. To this end, we propose MoBiE, the first binarization framework tailored for MoE-based LLMs. MoBiE is built on three core innovations: 1. using joint SVD decomposition to reduce cross-expert redundancy; 2. integrating global loss gradients into local Hessian metrics to enhance weight importance estimation; 3. introducing an error constraint guided by the input null space to mitigate routing distortion. Notably, MoBiE achieves these optimizations while incurring no additional storage overhead, striking a balance between efficiency and model performance. Extensive experiments demonstrate that MoBiE consistently outperforms state-of-the-art binary methods across multiple MoE-based LLMs and benchmarks. For example, on Qwen3-30B-A3B, MoBiE reduces perplexity by 52.2%\%, improves average zero-shot performance by 43.4%\%, achieves over 2 ×\times inference speedup, and further shortens quantization time. The code is available at https://github.com/Kishon-zzx/MoBiE.

Zhixiong Zhao, Zukang Xu, Zhixuan Chen +1

Houmo AI

Mixture-of-Experts InferenceMixture-of-Experts QuantizationLLM Inference AccelerationPost-Training Quantization
AskPDF
Apr 3, 2026

FluxMoE: Decoupling Expert Residency for High-Performance MoE Serving

Mixture-of-Experts (MoE) models have become mainstream for scaling language models to hundreds of billions of expert parameters. Despite sparse expert activation, existing inference engines keep all experts GPU-resident, crowding out the key-value cache in large-batch, long-output offline workloads. We present FluxMoE, which decouples experts from physical GPU residency and adapts their footprint to available memory through a new expert paging abstraction. FluxMoE combines PagedTensor for transparent remapping, a bandwidth-balanced hierarchy spanning losslessly compressed GPU memory and host DRAM, and a budget-aware residency planner. Unlike CPU-GPU co-inference and whole-layer offloading, FluxMoE streams weights on demand while keeping expert computation on GPUs. We implement FluxMoE atop vLLM and evaluate it on three MoE models. For GLM-4.5 on 8×\timesH20 GPUs, FluxMoE delivers up to 7.2×\times vLLM's throughput and 79.0% lower average Time-Per-Output-Token (TPOT), without measurable model-quality loss using lossless compression. For Mixtral-8×\times7B-Instruct on 2×\timesL40S GPUs, where weight-resident vLLM cannot fit, FluxMoE delivers 4.3×\times KTransformers's throughput and 29.1% lower average TPOT.

Qingxiu Liu, Yongchao He, Runhan Jiang +4

CUHK · Unaffiliated · SCITIX +1

Expert OffloadingGPU AccelerationMixture of ExpertsMixture-of-Experts InferenceMemory-Efficient InferenceLLM Inference Acceleration
AskPDF
Apr 2, 2026

DWDP: Distributed Weight Data Parallelism for High-Performance LLM Inference on NVL72

Large language model (LLM) inference increasingly depends on multi-GPU execution, yet existing inference parallelization strategies require layer-wise inter-rank synchronization, making end-to-end performance sensitive to workload imbalance. We present DWDP (Distributed Weight Data Parallelism), an inference parallelization strategy that preserves data-parallel execution while offloading MoE weights across peer GPUs and fetching missing experts on demand. By removing collective inter-rank synchronization, DWDP allows each GPU to progress independently. We further address the practical overheads of this design with two optimizations for split-weight management and asynchronous remote-weight prefetch. Implemented in TensorRT-LLM and evaluated with DeepSeek-R1 on GB200 NVL72, DWDP improves end-to-end output TPS/GPU by 8.8% at comparable TPS/user in the 20-100 TPS/user serving range under 8K input sequence length and 1K output sequence length.

Wanqian Li, Jintao Peng, Zongfei Jing +7

NVIDIA

LLM Inference AccelerationDistributed InferenceMixture-of-Experts Inference
AskPDF
Mar 29, 2026

Expert Streaming: Accelerating Low-Batch MoE Inference via Multi-chiplet Architecture and Dynamic Expert Trajectory Scheduling

Mixture-of-Experts is a promising approach for edge AI with low-batch inference. Yet, on-device deployments often face limited on-chip memory and severe workload imbalance; the prevalent use of offloading further incurs off-chip memory access bottlenecks. Moreover, MoE sparsity and dynamic gating shift distributed strategies toward much finer granularity and introduce runtime scheduling considerations. Recently, high die-to-die bandwidth chiplet interconnects have created new opportunities for multi-chiplet systems to address workload imbalance and offloading bottlenecks with fine-grained scheduling. In this paper, we propose Fully Sharded Expert Data Parallelism, a parallelization paradigm specifically architected for low-batch MoE inference on multi-chiplet accelerators. FSE-DP attains adaptive computation-communication overlap and balanced load by orchestrating fine-grained, complementary expert streams along dynamic trajectories across high-bandwidth D2D links. The attendant dataflow complexity is tamed by a minimal, hardware-amenable set of virtualization rules and a lightweight scheduling algorithm. Our approach achieves 1.22 to 2.00 times speedup over state-of-the-art baselines and saves up to 78.8 percent on-chip memory.

Songchen Ma, Hongyi Li, Weihao Zhang +8

AI Chip Center for Emerging Smart Systems, Hong Kong SAR, China · The Hong Kong University of Science and Technology, Hong Kong SAR, China · Shanghai UniVista Industrial Software Group Co., Ltd., Shanghai, China

Mixture-of-Experts InferenceEfficient Neural Network InferenceExpert ParallelismMemory-Efficient Inference
AskPDF
Mar 19, 2026

DyMoE: Dynamic Expert Orchestration with Mixed-Precision Quantization for Efficient MoE Inference on Edge

Despite the computational efficiency of MoE models, the excessive memory footprint and I/O overhead inherent in multi-expert architectures pose formidable challenges for real-time inference on resource-constrained edge platforms. While existing static methods struggle with a rigid latency-accuracy trade-off, we observe that expert importance is highly skewed and depth-dependent. Motivated by these insights, we propose DyMoE, a dynamic mixed-precision quantization framework designed for high-performance edge inference. Leveraging insights into expert importance skewness and depth-dependent sensitivity, DyMoE introduces: (1) importance-aware prioritization to dynamically quantize experts at runtime; (2) depth-adaptive scheduling to preserve semantic integrity in critical layers; and (3) look-ahead prefetching to overlap I/O stalls. Experimental results on commercial edge hardware show that DyMoE reduces Time-to-First-Token (TTFT) by 3.44x-22.7x and up to a 14.58x speedup in Time-Per-Output-Token (TPOT) compared to state-of-the-art offloading baselines, enabling real-time, accuracy-preserving MoE inference on resource-constrained edge devices.

Yuegui Huang, Zhiyuan Fang, Weiqi Luo +3

Sun Yat-sen University, Guangzhou, China · Tencent, Shenzhen, China

Mixture-of-Experts InferenceMixture-of-Experts QuantizationEdge InferenceMixed-Precision Quantization
AskPDF
Mar 16, 2026

Bridging Local and Global Knowledge: Cascaded Mixture-of-Experts Learning for Near-Shortest Path Routing

While deep learning models that leverage local features have demonstrated significant potential for near-optimal routing in dense Euclidean graphs, they struggle to generalize well in sparse networks where topological irregularities require broader structural awareness. To address this limitation, we train a Cascaded Mixture of Experts (Ca-MoE) to solve the all-pairs near-shortest path (APNSP) routing problem. Our Ca-MoE is a modular two-tier architecture that supports the decision-making for forwarder selection with lower-tier experts relying on local features and upper-tier experts relying on global features. It performs adaptive inference wherein the upper-tier experts are triggered only when the lower-tier ones do not suffice to achieve adequate decision quality. Computational efficiency is thus achieved by escalating model capacity only when necessitated by topological complexity, and parameter redundancy is avoided. Furthermore, we incorporate an online meta-learning strategy that facilitates independent expert fine-tuning and utilizes a stability-focused update mechanism to prevent catastrophic forgetting as new graph environments are encountered. Experimental evaluations demonstrate that Ca-MoE routing improves accuracy by up to 29.1% in sparse networks compared to single-expert baselines and maintains performance within 1%-6% of the theoretical upper bound across diverse graph densities.

Yung-Fu Chen, Anish Arora

The Ohio State University, Columbus, OH, USA

Mixture-of-Experts ModelsMixture-of-Experts InferenceAdaptive Model Routing
AskPDF
Mar 13, 2026

NCCL EP: Towards a Unified Expert Parallel Communication API for NCCL

Mixture-of-Experts (MoE) architectures have become essential for scaling large language models, driving the development of specialized device-initiated communication libraries such as DeepEP, Hybrid-EP, and others. These libraries demonstrate the performance benefits of GPU-initiated RDMA for MoE dispatch and combine operations. This paper presents NCCL EP (Expert Parallelism), a ground-up MoE communication library built entirely on NCCL's Device API. NCCL EP provides unified ncclEpDispatch and ncclEpCombine primitives with both C and Python interfaces, supporting Low-Latency (LL) mode for inference decoding and High-Throughput (HT) mode for training and inference prefill. LL targets small batch sizes (1-128 tokens) using direct all-to-all RDMA+NVLink mesh connectivity with double-buffered communication for overlapping dispatch and combine phases. HT targets large batches (4096+ tokens) using hierarchical communication that aggregates tokens within NVLink domains before inter-node RDMA transmission. Both modes leverage Device API for both intra- and inter-node communications, taking advantage of its topology awareness and optimized GPU-initiated implementation. We evaluate NCCL EP on an H100-based cluster across multi-node configurations, demonstrating competitive LL kernel performance and presenting end-to-end results with vLLM integration. By building MoE communication natively within NCCL, NCCL EP provides a supported path for expert parallelism on current and emerging NVIDIA platforms.

Amos Goldman, Nimrod Boker, Maayan Sheraizin +15

NVIDIA Corporation

Communication-Efficient Mixture-of-ExpertsExpert ParallelismMixture-of-Experts Inference
AskPDF
Mar 9, 2026

The qsqs Inequality: Quantifying the Double Penalty of Mixture-of-Experts at Inference

Mixture-of-Experts (MoE) models deliver high quality at low training FLOPs, but this efficiency often vanishes at inference. We identify a double penalty that structurally disadvantages MoE architectures during decoding: first, expert routing fragments microbatches and reduces weight reuse; second, massive resident expert pools reduce high-bandwidth memory (HBM) headroom for the KV cache. This phenomenon, formalized as reuse fragmentation, pushes feed-forward networks (FFNs) into a bandwidth-bound regime, especially at long context lengths. We introduce the qsqs inequality, a predictive criterion that identifies when MoE is structurally disadvantaged relative to a quality-matched dense model. This criterion unifies sparsity (ss), the fraction of parameters activated per token, and the quality-equivalence factor (qq), the size multiplier required for a dense model to match MoE performance. Our evaluation across frontier models including DeepSeek-V3, Qwen3-235B, Grok-1, and Switch-C demonstrates that this fragmentation is a general architectural phenomenon. For DeepSeek-V3 at 128k context, this results in a 4.5x throughput advantage for a quality-matched dense baseline. Crucially, massive architectures like Switch-C can become infeasible on cluster sizes where a quality-matched dense model remains viable. Our results suggest that training-time FLOP efficiency is an incomplete proxy for inference-time performance in long-context serving. They also indicate that MoE may be best viewed as a training-time optimization, with distillation into dense models as a possible path toward inference-efficient deployment.

Vignesh Adhinarayanan, Nuwan Jayasena

AMD Research and Advanced Development (RAD).

Mixture-of-Experts InferenceLong-Context Language Model InferenceMemory-Efficient Inference
AskPDF
Mar 9, 2026

LAR-MoE: Latent-Aligned Routing for Mixture of Experts in Robotic Imitation Learning

Imitation learning enables robots to acquire manipulation skills from demonstrations, yet deploying a policy across tasks with heterogeneous dynamics remains challenging, as models tend to average over distinct behavioral modes present in the demonstrations. Mixture-of-Experts (MoE) architectures address this by activating specialized subnetworks, but requires meaningful skill decompositions for expert routing. We introduce Latent-Aligned Routing for Mixture of Experts (LAR-MoE), a two-stage framework that decouples unsupervised skill discovery from policy learning. In pre-training, we learn a joint latent representation between observations and future actions through student-teacher co-training. In a post-training stage, the expert routing is regularized to follow the structure of the learned latent space, preventing expert collapse while maintaining parameter efficiency. We evaluate LAR-MoE in simulation and on hardware. On the LIBERO benchmark, our method achieves a 95.2% average success rate with 150M parameters. On a surgical bowel grasping and retraction task, LAR-MoE matches a supervised MoE baseline without requiring any phase annotations, and transfers zero-shot to ex vivo porcine tissue. Our findings suggest that latent-aligned routing provides a principled alternative to supervised skill decomposition, enabling structured expert specialization from unlabeled demonstrations.

Ariel Rodriguez, Chenpan Li, Lorenzo Mazza +5

Department of Translational Surgical Oncology, NCT/UCC Dresden, a partnership between DKFZ, MF UKD, TUD, HZDR Dresden, Germany. · Cluster of Excellence-CeTI, TUD, Germany. · Department of Visceral, Thoracic and Vascular Surgery, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD, Germany.

Robot Imitation LearningAgent Skill LearningRobotic ManipulationMixture-of-Experts Inference
AskPDF
Mar 9, 2026

FineRMoE: Dimension Expansion for Finer-Grained Expert with Its Upcycling Approach

As revealed by the scaling law of fine-grained MoE, model performance ceases to be improved once the granularity of the intermediate dimension exceeds the optimal threshold, limiting further gains from single-dimension fine-grained design. To address this bottleneck, we propose FineRMoE (FineR-Grained MoE), an architecture that extends fine-grained expert design to both intermediate and output dimensions, aiming to enhance expert specialization beyond the single-dimension limit. We further introduce a bi-level sparse forward computation paradigm and a specialized routing mechanism to govern the activation. In addition, to obviate the prohibitive cost of training FineRMoE from scratch, we devise a generalized upcycling method to build FineRMoE in a cost-effective manner. Extensive experiments demonstrate the superior performance achieved by FineRMoE across ten standard benchmarks. Compared with the strongest baseline, FineRMoE achieves 6 times higher parameter efficiency, 281 times lower prefill latency, and 136 timese higher decoding throughput during inference.

Ning Liao, Xiaoxing Wang, Xiaohan Qin +1

Shanghai Jiao Tong University

Sparse Mixture-of-ExpertsMixture-of-Experts ModelsMixture-of-Experts Inference
AskPDF
Mar 9, 2026

Speculating Experts Accelerates Inference for Mixture-of-Experts

Mixture-of-Experts (MoE) models have gained popularity as a means of scaling the capacity of large language models (LLMs) while maintaining sparse activations and reduced per-token compute. However, in memory-constrained inference settings, expert weights must be offloaded to CPU, creating a performance bottleneck from CPU-GPU transfers during decoding. We propose an expert prefetching scheme that leverages currently computed internal model representations to speculate future experts, enabling memory transfers to overlap with computation. Across multiple MoE architectures, we demonstrate that future experts can be reliably predicted by these internal representations. We also demonstrate that executing speculated experts generally maintains downstream task accuracy, thus preserving more effective compute-memory overlap by eliminating the need to re-fetch true router-selected experts. Integrated into an optimized inference engine, our approach achieves up to 14% reduction in time per output token (TPOT) over on-demand loading of experts from CPU memory. For MoEs where speculative execution alone yields suboptimal accuracy, we further examine lightweight estimators that improve expert prediction hit rates, thereby reducing performance degradation. Our code is released in open-source at https://github.com/axonn-ai/yalis/tree/offload_prefetch.

Vivan Madan, Prajwal Singhania, Abhinav Bhatele +2

University of Maryland · TogetherAI

Mixture-of-Experts InferenceLLM Inference AccelerationMixture-of-Experts Language Models
AskPDF
Mar 6, 2026

MoEless: Efficient MoE LLM Serving with Serverless Experts

Large Language Models (LLMs) increasingly adopt Mixture-of-Experts (MoE) architectures to scale efficiently under stringent resource constraints. However, MoE's sparse activation causes severe expert load imbalance, where a few experts become stragglers while others remain underutilized, leading to inflated inference latency and cost. Existing solutions assume static, serverful model deployments, limiting expert elasticity and often incurring costly expert swapping or degraded output quality. We present MoEless, an efficient serverless MoE serving framework that mitigates expert load imbalance via elastic expert execution. MoEless leverages lightweight, layer-aware predictors to estimate incoming expert load distributions and proactively identify stragglers. We design optimized scaling and placement strategies to improve function locality, GPU utilization, and cross-expert load balance. MoEless is prototyped on top of Megatron-LM and deployed on an eight-GPU testbed. Experiments with open-source MoE models and real-world workloads show that MoEless reduces inference latency by 43% and inference cost by 84% compared to state-of-the-art solutions.

Hanfei Yu, Bei Ouyang, Shwai He +2

University of Maryland College Park

LLM Inference EfficiencyExpert Load BalancingLLM ServingMixture-of-Experts Inference
AskPDF
Mar 3, 2026

Robust Heterogeneous Analog-Digital Computing for Mixture-of-Experts Models with Theoretical Generalization Guarantees

Sparse Mixture-of-Experts (MoE) models enable efficient scalability by activating only a small sub-set of experts per input, yet their massive parameter counts lead to substantial memory and energy inefficiency during inference. Analog in-memory computing (AIMC) offers a promising solution by eliminating frequent data movement between memory and compute units. However, mitigating hardware nonidealities of AIMC typically requires noise-aware retraining, which is infeasible for large MoE models. In this paper, we propose a retraining-free heterogeneous computation framework in which noise-sensitive experts, which are provably identifiable by their maximum neuron norm, are computed digitally while the majority of the experts are executed on AIMC hardware. We further assign densely activated modules, such as attention layers, to digital computation due to their high noise sensitivity despite comprising a small fraction of parameters. Extensive experiments on large MoE language models, including DeepSeekMoE and OLMoE, across multiple benchmark tasks validate the robustness of our approach in maintaining accuracy under analog nonidealities.

Mohammed Nowaz Rabbani Chowdhury, Hsinyu Tsai, Geoffrey W. Burr +3

Rensselaer Polytechnic Institute · IBM Research

Mixture-of-Experts InferenceHeterogeneous ComputingMixture-of-Experts Language ModelsSparse Mixture-of-ExpertsCompute-in-Memory
AskPDF
Mar 1, 2026

TriMoE: Augmenting GPU with AMX-Enabled CPU and DIMM-NDP for High-Throughput MoE Inference via Offloading

To deploy large Mixture-of-Experts (MoE) models cost-effectively, offloading-based single-GPU heterogeneous inference is crucial. While GPU-CPU architectures that offload cold experts are constrained by host memory bandwidth, emerging GPU-NDP architectures utilize DIMM-NDP to offload non-hot experts. However, non-hot experts are not a homogeneous memory-bound group: a significant subset of warm experts exists is severely penalized by high GPU I/O latency yet can saturate NDP compute throughput, exposing a critical compute gap. We present TriMoE, a novel GPU-CPU-NDP architecture that fills this gap by synergistically leveraging AMX-enabled CPU to precisely map hot, warm, and cold experts onto their optimal compute units. We further introduce a bottleneck-aware expert scheduling policy and a prediction-driven dynamic relayout/rebalancing scheme. Experiments demonstrate that TriMoE achieves up to 2.83x speedup over state-of-the-art solutions.

Yudong Pan, Yintao He, Tianhua Han +7

SKLP, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China · University of Chinese Academy of Sciences, Beijing, China

Mixture-of-Experts InferenceLLM Inference AccelerationComputation OffloadingHeterogeneous Computing
AskPDF
Feb 24, 2026

From Isolation to Integration: Building an Adaptive Expert Forest for Pre-Trained Model-based Class-Incremental Learning

Class-Incremental Learning (CIL) requires models to learn new classes without forgetting old ones. A common method is to freeze a pre-trained model and train a new, lightweight adapter for each task. While this prevents forgetting, it treats the learned knowledge as a simple, unstructured collection and fails to use the relationships between tasks. To this end, we propose the Semantic-guided Adaptive Expert Forest (SAEF), a new method that organizes adapters into a structured hierarchy for better knowledge sharing. SAEF first groups tasks into conceptual clusters based on their semantic relationships. Then, within each cluster, it builds a balanced expert tree by creating new adapters from merging the adapters of similar tasks. At inference time, SAEF finds and activates a set of relevant experts from the forest for any given input. The final prediction is made by combining the outputs of these activated experts, weighted by how confident each expert is. Experiments on several benchmark datasets show that SAEF achieves SOTA performance.

Ruiqi Liu, Boyu Diao, Hangda Liu +3

Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China · University of Chinese Academy of Sciences, Beijing, China

Adapter-Based Continual LearningClass-Incremental LearningIncremental LearningMixture-of-Experts Inference
AskPDF
Feb 23, 2026

A Replicate-and-Quantize Strategy for Plug-and-Play Load Balancing of Sparse Mixture-of-Experts LLMs

Sparse Mixture-of-Experts (SMoE) architectures are increasingly used to scale large language models efficiently, delivering strong accuracy under fixed compute budgets. However, SMoE models often suffer from severe load imbalance across experts, where a small subset of experts receives most tokens while others are underutilized. Prior work has focused mainly on training-time solutions such as routing regularization or auxiliary losses, leaving inference-time behavior, which is critical for deployment, less explored. We present a systematic analysis of expert routing during inference and identify three findings: (i) load imbalance persists and worsens with larger batch sizes, (ii) selection frequency does not reliably reflect expert importance, and (iii) overall expert workload and importance can be estimated using a small calibration set. These insights motivate inference-time mechanisms that rebalance workloads without retraining or router modification. We propose Replicate-and-Quantize (R&Q), a training-free and near-lossless framework for dynamic workload rebalancing. In each layer, heavy-hitter experts are replicated to increase parallel capacity, while less critical experts and replicas are quantized to remain within the original memory budget. We also introduce a Load-Imbalance Score (LIS) to measure routing skew by comparing heavy-hitter load to an equal allocation baseline. Experiments across representative SMoE models and benchmarks show up to 1.4x reduction in imbalance with accuracy maintained within +/-0.6%, enabling more predictable and efficient inference.

Zijie Liu, Jie Peng, Jinhao Duan +7

University of North Carolina at Chapel Hill · University of Minnesota Twin Cities · North Carolina State University +2

Expert Load BalancingMixture of ExpertsSparse Mixture-of-ExpertsMixture-of-Experts Inference
AskPDF
Feb 17, 2026

MoE-Spec: Expert Budgeting for Efficient Speculative Decoding

Speculative decoding accelerates Large Language Model (LLM) inference by verifying multiple drafted tokens in parallel. However, for Mixture-of-Experts (MoE) models, this parallelism introduces a severe bottleneck: large draft trees activate many unique experts, significantly increasing memory pressure and diminishing speedups from speculative decoding relative to autoregressive decoding. Prior methods reduce speculation depth when MoE verification becomes expensive. We propose MoE-Spec, a training-free verification-time expert budgeting method that decouples speculation depth from memory cost by enforcing a fixed expert capacity limit at each layer, loading only the experts that contribute most to verification and dropping the long tail of rarely used experts that drive bandwidth overhead. Experiments across multiple model scales and datasets show that this method yields 10--30% higher throughput than state-of-the-art speculative decoding baselines (EAGLE-3) at comparable quality, with flexibility to trade accuracy for further latency reductions through tighter budgets.

Bradley McDanel, Steven Li, Sruthikesh Surineni +1

Franklin and Marshall College · Meta Reality Labs

Speculative DecodingLLM Inference AccelerationMixture-of-Experts Inference
AskPDF
Feb 12, 2026

MoE-SpAc: Efficient MoE Inference Based on Speculative Activation Utility in Heterogeneous Edge Scenarios

Mixture-of-Experts (MoE) models enable scalable performance but face severe memory constraints on edge devices. Existing offloading strategies struggle with I/O bottlenecks due to the dynamic, low-information nature of autoregressive expert activation. In this paper, we propose to repurpose Speculative Decoding (SD) not merely as a compute accelerator, but as an informative lookahead sensor for memory management, supported by our theoretical and empirical analyses. Hence, we introduce MoE-SpAc, an MoE inference framework that integrates a Speculative Utility Estimator to track expert demand, a Heterogeneous Workload Balancer to dynamically partition computation via online integer optimization, and an Asynchronous Execution Engine to unify the prefetching and eviction in the same utility space. Extensive experiments on seven benchmarks demonstrate that MoE-SpAc achieves a 42% improvement in TPS over the SOTA SD-based baseline, and an average 4.04x speedup over all standard baselines. Code is available at https://github.com/lshAlgorithm/MoE-SpAc .

Shuhuai Li, Jianghao Lin, Dongdong Ge +1

Shanghai University, Shanghai, China · Shanghai Jiao Tong University, Shanghai, China · Standford University, California, United States

Mixture-of-Experts InferenceEdge InferenceMemory-Efficient InferenceSpeculative Decoding
AskPDF
Feb 5, 2026

OmniMoE: An Efficient MoE by Orchestrating Atomic Experts at Scale

Mixture-of-Experts (MoE) architectures are evolving towards finer granularity to improve parameter efficiency. However, existing MoE designs face an inherent trade-off between the granularity of expert specialization and hardware execution efficiency. We propose OmniMoE, a system-algorithm co-designed framework that pushes expert granularity to its logical extreme. OmniMoE introduces vector-level Atomic Experts, enabling scalable routing and execution within a single MoE layer, while retaining a shared dense MLP branch for general-purpose processing. Although this atomic design maximizes capacity, it poses severe challenges for routing complexity and memory access. To address these, OmniMoE adopts a system-algorithm co-design: (i) a Cartesian Product Router that decomposes the massive index space to reduce routing complexity from O(N) to O(sqrt(N)); and (ii) Expert-Centric Scheduling that inverts the execution order to turn scattered, memory-bound lookups into efficient dense matrix operations. Validated on seven benchmarks, OmniMoE (with 1.7B active parameters) achieves 50.9% zero-shot accuracy across seven benchmarks, outperforming coarse-grained (e.g., DeepSeekMoE) and fine-grained (e.g., PEER) baselines. Crucially, OmniMoE reduces inference latency from 73ms to 6.7ms (a 10.9-fold speedup) compared to PEER, demonstrating that massive-scale fine-grained MoE can be fast and accurate. Our code is open-sourced at https://github.com/flash-algo/omni-moe.

Jingze Shi, Zhangyang Peng, Yizhang Zhu +3

The Hong Kong University of Science and Technology (Guangzhou) · Beijing Academy of Artificial Intelligence

Mixture-of-Experts InferenceLLM Inference AccelerationExpert RoutingMixture-of-Experts Models
AskPDF
Jan 22, 2026

FlashMoE: Reducing SSD I/O Bottlenecks via ML-Based Cache Replacement for Mixture-of-Experts Inference on Edge Devices

Recently, Mixture-of-Experts (MoE) models have gained attention for efficiently scaling large language models. Although these models are extremely large, their sparse activation enables inference to be performed by accessing only a fraction of the model at a time. This property opens the possibility of on-device inference of MoE, which was previously considered infeasible for such large models. Consequently, various systems have been proposed to leverage this sparsity and enable efficient MoE inference for edge devices. However, previous MoE inference systems like Fiddler[8] or DAOP[13] rely on DRAM-based offloading and are not suitable for memory constrained on-device environments. As recent MoE models grow to hundreds of gigabytes, RAM-offloading solutions become impractical. To address this, we propose FlashMoE, a system that offloads inactive experts to SSD, enabling efficient MoE inference under limited RAM. FlashMoE incorporates a lightweight ML-based caching strategy that adaptively combines recency and frequency signals to maximize expert reuse, significantly reducing storage I/O. In addition, we built a user-grade desktop platform to demonstrate the practicality of FlashMoE. On this real hardware setup, FlashMoE improves cache hit rate by up to 51% over well-known offloading policies such as LRU and LFU, and achieves up to 2.6x speedup compared to existing MoE inference systems.

Byeongju Kim, Jungwan Lee, Donghyeon Han +2

Edge InferenceMixture-of-Experts InferenceCache ReplacementMemory-Efficient InferenceOn-Device Inference
AskPDF
Nov 19, 2025

Dynamic Expert Quantization for Scalable Mixture-of-Experts Inference

Mixture-of-Experts (MoE) has become a practical architecture for scaling LLM capacity while keeping per-token compute modest, but deploying MoE models on a single, memory-limited GPU remains difficult because expert weights dominate the HBM footprint. Existing expert offloading and prefetching systems reduce the resident set, yet they often pay expert-loading costs on the critical path when activation becomes dense. Post-training quantization (PTQ) lowers the footprint without transfers, but prevailing pipelines fix expert bit-widths offline and assume routing remains stable, even though MoE expert utilization is heavy-tailed and the hot set can shift across workloads. We present DynaExq, a runtime-aware mixed-precision serving system that treats single-GPU MoE inference under a hard HBM envelope as an online, budget-constrained precision allocation problem. The key insight is to keep the experts that dominate runtime traffic resident at higher precision, while maintaining a low-precision fallback for the remaining experts, so the system can reduce transfer volume and avoid the waiting latency that limits offloading and prefetching under dense activation. DynaExq estimates long-horizon expert hotness from router traces, selects a per-layer high-precision resident set via a budget-feasible top-nn rule, and applies promotions and demotions asynchronously through stable expert handles so the forward pass always executes on a fully materialized expert version. Across Qwen3-MoE-30B/80B and six benchmarks, DynaExq improves accuracy over static PTQ on Qwen3-80B (73.09% to 77.57%) under comparable device-memory budgets and achieves up to 2.73x higher throughput than offloading/prefetch baselines at batch size 32.

Kexin Chu, Dawei Xiang, Zixu Shen +3

University of Connecticut · University of California, Santa Cruz · Independent Researcher

Mixed-Precision QuantizationGPU AccelerationMixture-of-Experts InferenceLLM Inference AccelerationMixture-of-Experts QuantizationPost-Training Quantization+1
AskPDF
Nov 18, 2025

Specter: Fast MoE Inference via Quantization-Enabled Multi-Step Speculative Offloading

Mixture-of-experts (MoE) models often exceed a single accelerator's memory capacity. Offloading experts to host memory creates a severe PCIe I/O bottleneck because data-dependent routing keeps transfers on the critical path. We present Specter, an inference system that uses quantization to combine speculative decoding with expert offloading. Its lightweight quantized draft model runs efficiently on device while preserving high-fidelity expert activation patterns. This provides accurate multi-step lookahead over future expert usage, allowing the runtime to organize expert movement and execution ahead of verification and improve both communication and execution efficiency. Under single-request inference, Specter achieves an average 1.50x (up to 1.88x) speedup in time per output token (TPOT) over the state-of-the-art system. Specter makes memory-constrained, single-GPU MoE inference more practical on commodity hardware. Our code is available at https://github.com/wwwwwf/specter_asplos.

Wenfeng Wang, Xiaofeng Hou, Xinfeng Xia +5

Shanghai Jiao Tong University, Shanghai, China · Shanghai Jiao Tong University, China · Hong Kong University of Science and Technology, Hongkong, China

Mixture-of-Experts InferenceLLM Inference AccelerationExpert OffloadingMemory-Efficient Inference
AskPDF
Sep 30, 2025

Training Matryoshka Mixture-of-Experts for Elastic Inference-Time Expert Utilization

Mixture-of-Experts (MoE) has emerged as a promising paradigm for efficiently scaling large language models without a proportional increase in computational cost. However, the standard training strategy of Top-K router prevents MoE models from realizing their full potential for elastic inference. When the number of activated experts is altered at inference time, these models exhibit precipitous performance degradation. In this work, we introduce Matryoshka MoE (M-MoE), a training framework that instills a coarse-to-fine structure directly into the expert ensemble. By systematically varying the number of activated experts during training, M-MoE compels the model to learn a meaningful ranking: top-ranked experts collaborate to provide essential, coarse-grained capabilities, while subsequent experts add progressively finer-grained detail. We explore this principle at multiple granularities, identifying a layer-wise randomization strategy as the most effective. Our experiments demonstrate that a single M-MoE model achieves remarkable elasticity, with its performance at various expert counts closely matching that of an entire suite of specialist models, but at only a fraction of the total training cost. This flexibility not only unlocks elastic inference but also enables optimizing performance by allocating different computational budgets to different model layers. Our work paves the way for more practical and adaptable deployments of large-scale MoE models.

Yaoxiang Wang, Qingguo Hu, Yucheng Ding +6

Xiamen University · Shanghai Jiao Tong University · University of Science and Technology of China +1

Mixture-of-Experts Language ModelsMixture-of-Experts ModelsMixture-of-Experts Inference
AskPDF
Jan 28, 2025

3D-MoE: Towards Spatial Intelligence with Mixture-of-Experts for 3D Reasoning and Action Generation

Spatial intelligence, encompassing 3D perception and reasoning, is the essential next frontier of AI. Scaling current 3D vision-language models (VLMs) that rely on dense Transformers for spatial tasks incurs prohibitive computational costs. In this paper, we introduce 3D-MoE, a 3D VLM leveraging an efficient mixture-of-experts architecture with a modality- and spatial-context-aware probabilistic routing scheme, stably cultivated by a novel routing curriculum. To seamlessly extend 3D-MoE to embodied AI, we integrate a diffusion-based action head, Pose-DiT, transforming 3D-MoE into a 3D vision-language-action (VLA) model. By employing a rectified flow framework, Pose-DiT generates precise 6D pose actions in a single sampling step. Extensive experiments demonstrate that 3D-MoE achieves superior performance on diverse 3D vision-language benchmarks with drastically fewer activated parameters and yields higher success rates while enabling real-time inference for robot manipulation tasks.

Yueen Ma, Zenglin Xu, Irwin King

Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong SAR

3D Spatial ReasoningMixture-of-Experts InferenceRobotic ManipulationVision-Language-Action Models3D VLMs
AskPDF