{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/method/moe","slug":"moe","name":"MoE","full_name":"Mixture of Experts","full_name_withheld":false,"description_markdown":null,"description_state":"absent","introduced_year":null,"introduced_by":{"title":"Equipping Computational Pathology Systems with Artifact Processing Pipelines: A Showcase for Computation and Performance Trade-offs","paper":"/paper/equipping-computational-pathology-systems","first_author":"Neel Kanwal","n_authors":9,"url_abs":null,"archive_paper_url":"https://paperswithcode.com/paper/equipping-computational-pathology-systems"},"source":{"url":"https://arxiv.org/abs/2403.07743v3","title":"Equipping Computational Pathology Systems with Artifact Processing Pipelines: A Showcase for Computation and Performance Trade-offs","url_on_a_paper_host":true},"code_snippet_url":null,"code_snippet_url_on_a_code_host":false,"categories":[{"area":"General","area_id":"general","collection":"Ensembling","url":"/methods/category/ensembling","pwc_aliases":[]}],"n_papers_tagged":366,"archive_num_papers":366,"papers_newest_first":[{"paper":null,"slug":"mixture-of-experts-in-large-language-models","title":"Mixture of Experts in Large Language Models","date":"2025-07-15","arxiv_id":"2507.11181","n_code_links":0,"syntology":null},{"paper":"/paper/mofe-time-mixture-of-frequency-domain-experts","slug":"mofe-time-mixture-of-frequency-domain-experts","title":"MoFE-Time: Mixture of Frequency Domain Experts for Time-Series Forecasting Models","date":"2025-07-09","arxiv_id":"2507.06502","n_code_links":1,"syntology":null},{"paper":"/paper/growing-transformers-modular-composition-and","slug":"growing-transformers-modular-composition-and","title":"Growing Transformers: Modular Composition and Layer-wise Expansion on a Frozen Substrate","date":"2025-07-08","arxiv_id":"2507.07129","n_code_links":1,"syntology":null},{"paper":null,"slug":"speech-quality-assessment-model-based-on","title":"Speech Quality Assessment Model Based on Mixture of Experts: System-Level Performance Enhancement and Utterance-Level Challenge Analysis","date":"2025-07-08","arxiv_id":"2507.06116","n_code_links":0,"syntology":null},{"paper":"/paper/learning-robust-stereo-matching-in-the-wild","slug":"learning-robust-stereo-matching-in-the-wild","title":"Learning Robust Stereo Matching in the Wild with Selective Mixture-of-Experts","date":"2025-07-07","arxiv_id":"2507.04631","n_code_links":1,"syntology":null},{"paper":null,"slug":"sub-moe-efficient-mixture-of-expert-llms","title":"Sub-MoE: Efficient Mixture-of-Expert LLMs Compression via Subspace Expert Merging","date":"2025-06-29","arxiv_id":"2506.23266","n_code_links":0,"syntology":null},{"paper":"/paper/latent-prototype-routing-achieving-near","slug":"latent-prototype-routing-achieving-near","title":"Latent Prototype Routing: Achieving Near-Perfect Load Balancing in Mixture-of-Experts","date":"2025-06-26","arxiv_id":"2506.21328","n_code_links":1,"syntology":null},{"paper":null,"slug":"safex-analyzing-vulnerabilities-of-moe-based","title":"SAFEx: Analyzing Vulnerabilities of MoE-Based LLMs via Stable Safety-critical Expert Identification","date":"2025-06-20","arxiv_id":"2506.17368","n_code_links":0,"syntology":null},{"paper":null,"slug":"less-is-more-undertraining-experts-improves","title":"Less is More: Undertraining Experts Improves Model Upcycling","date":"2025-06-17","arxiv_id":"2506.14126","n_code_links":0,"syntology":null},{"paper":null,"slug":"lora-mixer-coordinate-modular-lora-experts","title":"LoRA-Mixer: Coordinate Modular LoRA Experts Through Serial Attention Routing","date":"2025-06-17","arxiv_id":"2507.00029","n_code_links":0,"syntology":null},{"paper":null,"slug":"ring-lite-scalable-reasoning-via-c3po","title":"Ring-lite: Scalable Reasoning via C3PO-Stabilized Reinforcement Learning for LLMs","date":"2025-06-17","arxiv_id":"2506.14731","n_code_links":0,"syntology":null},{"paper":null,"slug":"utility-driven-speculative-decoding-for","title":"Utility-Driven Speculative Decoding for Mixture-of-Experts","date":"2025-06-17","arxiv_id":"2506.20675","n_code_links":0,"syntology":null},{"paper":null,"slug":"eaquant-enhancing-post-training-quantization","title":"EAQuant: Enhancing Post-Training Quantization for MoE Models via Expert-Aware Optimization","date":"2025-06-16","arxiv_id":"2506.13329","n_code_links":0,"syntology":null},{"paper":null,"slug":"serving-large-language-models-on-huawei","title":"Serving Large Language Models on Huawei CloudMatrix384","date":"2025-06-15","arxiv_id":"2506.12708","n_code_links":0,"syntology":null},{"paper":"/paper/ming-omni-a-unified-multimodal-model-for","slug":"ming-omni-a-unified-multimodal-model-for","title":"Ming-Omni: A Unified Multimodal Model for Perception and Generation","date":"2025-06-11","arxiv_id":"2506.09344","n_code_links":1,"syntology":{"ran":12,"of":14,"n_ran_checked":11,"n_instrument":1,"unverified":2,"pointer_only":0,"phrase":"12 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 3 honoured, 0 violated, 8 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified","official":{"repos":["inclusionai/ming"],"state":"official (archive's flag): 12 ran","n_ran":12,"n_constructed":0,"n_ran_no_instrument_failure":11,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":null,"slug":"m2restore-mixture-of-experts-based-mamba-cnn","title":"M2Restore: Mixture-of-Experts-based Mamba-CNN Fusion Framework for All-in-One Image Restoration","date":"2025-06-09","arxiv_id":"2506.07814","n_code_links":0,"syntology":null},{"paper":null,"slug":"moe-gps-guidlines-for-prediction-strategy-for","title":"MoE-GPS: Guidlines for Prediction Strategy for Dynamic Expert Duplication in MoE Load Balancing","date":"2025-06-09","arxiv_id":"2506.07366","n_code_links":0,"syntology":null},{"paper":null,"slug":"smar-soft-modality-aware-routing-strategy-for","title":"SMAR: Soft Modality-Aware Routing Strategy for MoE-based Multimodal Large Language Models Preserving Language Capabilities","date":"2025-06-06","arxiv_id":"2506.06406","n_code_links":0,"syntology":null},{"paper":"/paper/flashdmoe-fast-distributed-moe-in-a-single","slug":"flashdmoe-fast-distributed-moe-in-a-single","title":"FlashDMoE: Fast Distributed MoE in a Single Kernel","date":"2025-06-05","arxiv_id":"2506.04667","n_code_links":2,"syntology":{"ran":3,"of":7,"n_ran_checked":3,"n_instrument":0,"unverified":4,"pointer_only":0,"phrase":"3 ran (of which 2 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified","official":{"repos":["osayamenja/aristos","osayamenja/kleos"],"state":"official (archive's flag): 3 ran","n_ran":3,"n_constructed":2,"n_ran_no_instrument_failure":3,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":null,"slug":"out-of-distribution-graph-models-merging","title":"Out-of-Distribution Graph Models Merging","date":"2025-06-04","arxiv_id":"2506.03674","n_code_links":0,"syntology":null},{"paper":null,"slug":"decoding-knowledge-attribution-in-mixture-of","title":"Decoding Knowledge Attribution in Mixture-of-Experts: A Framework of Basic-Refinement Collaboration and Efficiency Analysis","date":"2025-05-30","arxiv_id":"2505.24593","n_code_links":0,"syntology":null},{"paper":"/paper/mastering-massive-multi-task-reinforcement","slug":"mastering-massive-multi-task-reinforcement","title":"Mastering Massive Multi-Task Reinforcement Learning via Mixture-of-Expert Decision Transformer","date":"2025-05-30","arxiv_id":"2505.24378","n_code_links":1,"syntology":null},{"paper":null,"slug":"mixture-of-experts-for-personalized-and","title":"Mixture-of-Experts for Personalized and Semantic-Aware Next Location Prediction","date":"2025-05-30","arxiv_id":"2505.24597","n_code_links":0,"syntology":null},{"paper":null,"slug":"on-the-expressive-power-of-mixture-of-experts","title":"On the Expressive Power of Mixture-of-Experts for Structured Complex Tasks","date":"2025-05-30","arxiv_id":"2505.24205","n_code_links":0,"syntology":null},{"paper":null,"slug":"advancing-expert-specialization-for-better","title":"Advancing Expert Specialization for Better MoE","date":"2025-05-28","arxiv_id":"2505.22323","n_code_links":0,"syntology":null},{"paper":null,"slug":"evomoe-expert-evolution-in-mixture-of-experts","title":"EvoMoE: Expert Evolution in Mixture of Experts for Multimodal Large Language Models","date":"2025-05-28","arxiv_id":"2505.23830","n_code_links":0,"syntology":null},{"paper":"/paper/hidream-i1-a-high-efficient-image-generative","slug":"hidream-i1-a-high-efficient-image-generative","title":"HiDream-I1: A High-Efficient Image Generative Foundation Model with Sparse Diffusion Transformer","date":"2025-05-28","arxiv_id":"2505.22705","n_code_links":2,"syntology":{"ran":2,"of":3,"n_ran_checked":1,"n_instrument":1,"unverified":1,"pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified","official":{"repos":["hidream-ai/hidream-e1","hidream-ai/hidream-i1"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/flame-moe-a-transparent-end-to-end-research","slug":"flame-moe-a-transparent-end-to-end-research","title":"FLAME-MoE: A Transparent End-to-End Research Platform for Mixture-of-Experts Language Models","date":"2025-05-26","arxiv_id":"2505.20225","n_code_links":1,"syntology":null},{"paper":null,"slug":"moesd-unveil-speculative-decoding-s-potential","title":"MoESD: Unveil Speculative Decoding's Potential for Accelerating Sparse MoE","date":"2025-05-26","arxiv_id":"2505.19645","n_code_links":0,"syntology":null},{"paper":null,"slug":"mosaic-data-free-knowledge-distillation-via","title":"Mosaic: Data-Free Knowledge Distillation via Mixture-of-Experts for Heterogeneous Distributed Environments","date":"2025-05-26","arxiv_id":"2505.19699","n_code_links":0,"syntology":null},{"paper":"/paper/i2moe-interpretable-multimodal-interaction","slug":"i2moe-interpretable-multimodal-interaction","title":"I2MoE: Interpretable Multimodal Interaction-aware Mixture-of-Experts","date":"2025-05-25","arxiv_id":"2505.19190","n_code_links":1,"syntology":{"ran":4,"of":4,"n_ran_checked":4,"n_instrument":0,"unverified":0,"pointer_only":4,"phrase":"4 ran (of which 4 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; every one of the 4 samples that ran constructed an object rather than computing a result","official":{"repos":["raina-xin/i2moe"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":4,"n_ran_no_instrument_failure":4,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/guiding-the-experts-semantic-priors-for","slug":"guiding-the-experts-semantic-priors-for","title":"Guiding the Experts: Semantic Priors for Efficient and Focused MoE Routing","date":"2025-05-24","arxiv_id":"2505.18586","n_code_links":1,"syntology":null},{"paper":null,"slug":"on-minimax-estimation-of-parameters-in","title":"On Minimax Estimation of Parameters in Softmax-Contaminated Mixture of Experts","date":"2025-05-24","arxiv_id":"2505.18455","n_code_links":0,"syntology":null},{"paper":null,"slug":"evidencemoe-a-physics-guided-mixture-of","title":"EvidenceMoE: A Physics-Guided Mixture-of-Experts with Evidential Critics for Advancing Fluorescence Light Detection and Ranging in Scattering Media","date":"2025-05-23","arxiv_id":"2505.21532","n_code_links":0,"syntology":null},{"paper":"/paper/drivemoe-mixture-of-experts-for-vision","slug":"drivemoe-mixture-of-experts-for-vision","title":"DriveMoE: Mixture-of-Experts for Vision-Language-Action Model in End-to-End Autonomous Driving","date":"2025-05-22","arxiv_id":"2505.16278","n_code_links":0,"syntology":null},{"paper":"/paper/janusdna-a-powerful-bi-directional-hybrid-dna","slug":"janusdna-a-powerful-bi-directional-hybrid-dna","title":"JanusDNA: A Powerful Bi-directional Hybrid DNA Foundation Model","date":"2025-05-22","arxiv_id":"2505.17257","n_code_links":1,"syntology":{"ran":5,"of":5,"n_ran_checked":3,"n_instrument":2,"unverified":0,"pointer_only":2,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","official":{"repos":["qihao-duan/janusdna"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":null,"slug":"trimr-verifier-based-training-free-thinking","title":"TrimR: Verifier-based Training-Free Thinking Compression for Efficient Test-Time Scaling","date":"2025-05-22","arxiv_id":"2505.17155","n_code_links":0,"syntology":null},{"paper":null,"slug":"chinese-toxic-language-mitigation-via","title":"Chinese Toxic Language Mitigation via Sentiment Polarity Consistent Rewrites","date":"2025-05-21","arxiv_id":"2505.15297","n_code_links":0,"syntology":null},{"paper":null,"slug":"efficient-data-driven-mixture-of-expert","title":"Efficient Data Driven Mixture-of-Expert Extraction from Trained Networks","date":"2025-05-21","arxiv_id":"2505.15414","n_code_links":0,"syntology":null},{"paper":null,"slug":"hunyuan-turbos-advancing-large-language","title":"Hunyuan-TurboS: Advancing Large Language Models through Mamba-Transformer Synergy and Adaptive Chain-of-Thought","date":"2025-05-21","arxiv_id":"2505.15431","n_code_links":0,"syntology":null},{"paper":"/paper/not-all-models-suit-expert-offloading-on","slug":"not-all-models-suit-expert-offloading-on","title":"Not All Models Suit Expert Offloading: On Local Routing Consistency of Mixture-of-Expert Models","date":"2025-05-21","arxiv_id":"2505.16056","n_code_links":1,"syntology":null},{"paper":null,"slug":"efficientllm-efficiency-in-large-language","title":"EfficientLLM: Efficiency in Large Language Models","date":"2025-05-20","arxiv_id":"2505.13840","n_code_links":0,"syntology":null},{"paper":"/paper/local-mixtures-of-experts-essentially-free","slug":"local-mixtures-of-experts-essentially-free","title":"Local Mixtures of Experts: Essentially Free Test-Time Training via Model Merging","date":"2025-05-20","arxiv_id":"2505.14136","n_code_links":1,"syntology":null},{"paper":null,"slug":"thor-moe-hierarchical-task-guided-and-context","title":"THOR-MoE: Hierarchical Task-Guided and Context-Responsive Routing for Neural Machine Translation","date":"2025-05-20","arxiv_id":"2505.14173","n_code_links":0,"syntology":null},{"paper":null,"slug":"model-selection-for-gaussian-gated-gaussian","title":"Model Selection for Gaussian-gated Gaussian Mixture of Experts Using Dendrograms of Mixing Measures","date":"2025-05-19","arxiv_id":"2505.13052","n_code_links":0,"syntology":null},{"paper":"/paper/occult-optimizing-collaborative-communication","slug":"occult-optimizing-collaborative-communication","title":"Occult: Optimizing Collaborative Communication across Experts for Accelerated Parallel MoE Training and Inference","date":"2025-05-19","arxiv_id":"2505.13345","n_code_links":1,"syntology":null},{"paper":null,"slug":"transfertraj-a-vehicle-trajectory-learning","title":"TransferTraj: A Vehicle Trajectory Learning Model for Region and Task Transferability","date":"2025-05-19","arxiv_id":"2505.12672","n_code_links":0,"syntology":null},{"paper":null,"slug":"2505-10860","title":"On DeepSeekMoE: Statistical Benefits of Shared Experts and Normalized Sigmoid Gating","date":"2025-05-16","arxiv_id":"2505.10860","n_code_links":0,"syntology":null},{"paper":null,"slug":"2505-11415","title":"MoE-CAP: Benchmarking Cost, Accuracy and Performance of Sparse Mixture-of-Experts Systems","date":"2025-05-16","arxiv_id":"2505.11415","n_code_links":0,"syntology":null},{"paper":null,"slug":"2505-11432","title":"MegaScale-MoE: Large-Scale Communication-Efficient Training of Mixture-of-Experts Models in Production","date":"2025-05-16","arxiv_id":"2505.11432","n_code_links":0,"syntology":null},{"paper":"/paper/pt-moe-an-efficient-finetuning-framework-for","slug":"pt-moe-an-efficient-finetuning-framework-for","title":"PT-MoE: An Efficient Finetuning Framework for Integrating Mixture-of-Experts into Prompt Tuning","date":"2025-05-14","arxiv_id":"2505.09519","n_code_links":1,"syntology":null},{"paper":"/paper/qwen3-technical-report","slug":"qwen3-technical-report","title":"Qwen3 Technical Report","date":"2025-05-14","arxiv_id":"2505.09388","n_code_links":4,"syntology":null},{"paper":null,"slug":"umoe-unifying-attention-and-ffn-with-shared","title":"UMoE: Unifying Attention and FFN with Shared Experts","date":"2025-05-12","arxiv_id":"2505.07260","n_code_links":0,"syntology":null},{"paper":null,"slug":"floe-on-the-fly-moe-inference","title":"FloE: On-the-Fly MoE Inference on Memory-constrained GPU","date":"2025-05-09","arxiv_id":"2505.05950","n_code_links":0,"syntology":null},{"paper":"/paper/mxmoe-mixed-precision-quantization-for-moe","slug":"mxmoe-mixed-precision-quantization-for-moe","title":"MxMoE: Mixed-precision Quantization for MoE with Accuracy and Performance Co-Design","date":"2025-05-09","arxiv_id":"2505.05799","n_code_links":1,"syntology":null},{"paper":null,"slug":"pangu-ultra-moe-how-to-train-your-big-moe-on","title":"Pangu Ultra MoE: How to Train Your Big MoE on Ascend NPUs","date":"2025-05-07","arxiv_id":"2505.04519","n_code_links":0,"syntology":null},{"paper":null,"slug":"faster-moe-llm-inference-for-extremely-large","title":"Faster MoE LLM Inference for Extremely Large Models","date":"2025-05-06","arxiv_id":"2505.03531","n_code_links":0,"syntology":null},{"paper":null,"slug":"multimodal-deep-learning-empowered-beam","title":"Multimodal Deep Learning-Empowered Beam Prediction in Future THz ISAC Systems","date":"2025-05-05","arxiv_id":"2505.02381","n_code_links":0,"syntology":null},{"paper":null,"slug":"moxe-mixture-of-xlstm-experts-with-entropy","title":"MoxE: Mixture of xLSTM Experts with Entropy-Aware Routing for Efficient Language Modeling","date":"2025-05-01","arxiv_id":"2505.01459","n_code_links":0,"syntology":null},{"paper":null,"slug":"accelerating-mixture-of-experts-training-with","title":"Accelerating Mixture-of-Experts Training with Adaptive Expert Replication","date":"2025-04-28","arxiv_id":"2504.19925","n_code_links":0,"syntology":null},{"paper":null,"slug":"badmoe-backdooring-mixture-of-experts-llms","title":"BadMoE: Backdooring Mixture-of-Experts LLMs via Optimizing Routing Triggers and Infecting Dormant Experts","date":"2025-04-24","arxiv_id":"2504.18598","n_code_links":0,"syntology":null},{"paper":null,"slug":"moe-parallel-folding-heterogeneous","title":"MoE Parallel Folding: Heterogeneous Parallelism Mappings for Efficient Large-Scale MoE Model Training with Megatron Core","date":"2025-04-21","arxiv_id":"2504.14960","n_code_links":0,"syntology":null},{"paper":null,"slug":"d-2-moe-dual-routing-and-dynamic-scheduling","title":"D$^{2}$MoE: Dual Routing and Dynamic Scheduling for Efficient On-Device MoE-based LLM Serving","date":"2025-04-17","arxiv_id":"2504.15299","n_code_links":0,"syntology":null},{"paper":"/paper/dense-backpropagation-improves-training-for","slug":"dense-backpropagation-improves-training-for","title":"Dense Backpropagation Improves Training for Sparse Mixture-of-Experts","date":"2025-04-16","arxiv_id":"2504.12463","n_code_links":1,"syntology":{"ran":4,"of":6,"n_ran_checked":1,"n_instrument":3,"unverified":2,"pointer_only":6,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 3 where Syntology's instrument failed) · 2 unverified","official":{"repos":["vatsal0/default-moe"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":null,"slug":"unveiling-hidden-collaboration-within-mixture","title":"Unveiling Hidden Collaboration within Mixture-of-Experts in Large Language Models","date":"2025-04-16","arxiv_id":"2504.12359","n_code_links":0,"syntology":null},{"paper":null,"slug":"the-sword-of-damocles-in-vits-computational","title":"The Sword of Damocles in ViTs: Computational Redundancy Amplifies Adversarial Transferability","date":"2025-04-15","arxiv_id":"2504.10804","n_code_links":0,"syntology":null},{"paper":null,"slug":"moe-lens-towards-the-hardware-limit-of-high","title":"MoE-Lens: Towards the Hardware Limit of High-Throughput MoE LLM Serving Under Resource Constraints","date":"2025-04-12","arxiv_id":"2504.09345","n_code_links":0,"syntology":null},{"paper":"/paper/c3po-critical-layer-core-expert-collaborative","slug":"c3po-critical-layer-core-expert-collaborative","title":"C3PO: Critical-Layer, Core-Expert, Collaborative Pathway Optimization for Test-Time Expert Re-Mixing","date":"2025-04-10","arxiv_id":"2504.07964","n_code_links":1,"syntology":null},{"paper":"/paper/cluster-driven-expert-pruning-for-mixture-of","slug":"cluster-driven-expert-pruning-for-mixture-of","title":"Cluster-Driven Expert Pruning for Mixture-of-Experts Large Language Models","date":"2025-04-10","arxiv_id":"2504.07807","n_code_links":1,"syntology":null},{"paper":null,"slug":"holistic-capability-preservation-towards","title":"Holistic Capability Preservation: Towards Compact Yet Comprehensive Reasoning Models","date":"2025-04-09","arxiv_id":"2504.07158","n_code_links":0,"syntology":null},{"paper":"/paper/moediff-sr-mixture-of-experts-guided","slug":"moediff-sr-mixture-of-experts-guided","title":"MoEDiff-SR: Mixture of Experts-Guided Diffusion Model for Region-Adaptive MRI Super-Resolution","date":"2025-04-09","arxiv_id":"2504.07308","n_code_links":1,"syntology":null},{"paper":null,"slug":"finding-fantastic-experts-in-moes-a-unified","title":"Finding Fantastic Experts in MoEs: A Unified Study for Expert Dropping Strategies and Observations","date":"2025-04-08","arxiv_id":"2504.05586","n_code_links":0,"syntology":null},{"paper":"/paper/hybrimoe-hybrid-cpu-gpu-scheduling-and-cache","slug":"hybrimoe-hybrid-cpu-gpu-scheduling-and-cache","title":"HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference","date":"2025-04-08","arxiv_id":"2504.05897","n_code_links":1,"syntology":null},{"paper":null,"slug":"hetermoe-efficient-training-of-mixture-of","title":"HeterMoE: Efficient Training of Mixture-of-Experts Models on Heterogeneous GPUs","date":"2025-04-04","arxiv_id":"2504.03871","n_code_links":0,"syntology":null},{"paper":null,"slug":"megascale-infer-serving-mixture-of-experts-at","title":"MegaScale-Infer: Serving Mixture-of-Experts at Scale with Disaggregated Expert Parallelism","date":"2025-04-03","arxiv_id":"2504.02263","n_code_links":0,"syntology":null},{"paper":"/paper/milo-efficient-quantized-moe-inference-with","slug":"milo-efficient-quantized-moe-inference-with","title":"MiLo: Efficient Quantized MoE Inference with Mixture of Low-Rank Compensators","date":"2025-04-03","arxiv_id":"2504.02658","n_code_links":1,"syntology":null},{"paper":null,"slug":"advancing-moe-efficiency-a-collaboration","title":"Advancing MoE Efficiency: A Collaboration-Constrained Routing (C2R) Strategy for Better Expert Parallelism Design","date":"2025-04-02","arxiv_id":"2504.01337","n_code_links":0,"syntology":null},{"paper":null,"slug":"detecting-financial-fraud-with-hybrid-deep","title":"Detecting Financial Fraud with Hybrid Deep Learning: A Mix-of-Experts Approach to Sequential and Anomalous Patterns","date":"2025-04-01","arxiv_id":"2504.03750","n_code_links":0,"syntology":null},{"paper":null,"slug":"mixture-of-routers","title":"Mixture of Routers","date":"2025-03-30","arxiv_id":"2503.23362","n_code_links":0,"syntology":null},{"paper":null,"slug":"beyond-standard-moe-mixture-of-latent-experts","title":"Beyond Standard MoE: Mixture of Latent Experts for Resource-Efficient Language Models","date":"2025-03-29","arxiv_id":"2503.23100","n_code_links":0,"syntology":null},{"paper":null,"slug":"moqa-rethinking-moe-quantization-with-multi","title":"MoQa: Rethinking MoE Quantization with Multi-stage Data-model Distribution Awareness","date":"2025-03-27","arxiv_id":"2503.21135","n_code_links":0,"syntology":null},{"paper":"/paper/a-multi-scale-lithium-ion-battery-capacity","slug":"a-multi-scale-lithium-ion-battery-capacity","title":"A multi-scale lithium-ion battery capacity prediction using mixture of experts and patch-based MLP","date":"2025-03-26","arxiv_id":"2504.03706","n_code_links":1,"syntology":null},{"paper":null,"slug":"optimal-scaling-laws-for-efficiency-gains-in","title":"Optimal Scaling Laws for Efficiency Gains in a Theoretical Transformer-Augmented Sectional MoE Framework","date":"2025-03-26","arxiv_id":"2503.20750","n_code_links":0,"syntology":null},{"paper":null,"slug":"expertrag-efficient-rag-with-mixture-of","title":"ExpertRAG: Efficient RAG with Mixture of Experts -- Optimizing Context Retrieval for Adaptive LLM Responses","date":"2025-03-23","arxiv_id":"2504.08744","n_code_links":0,"syntology":null},{"paper":null,"slug":"expert-race-a-flexible-routing-strategy-for","title":"Expert Race: A Flexible Routing Strategy for Scaling Diffusion Transformer with Mixture of Experts","date":"2025-03-20","arxiv_id":"2503.16057","n_code_links":0,"syntology":null},{"paper":"/paper/mixture-of-lookup-experts","slug":"mixture-of-lookup-experts","title":"Mixture of Lookup Experts","date":"2025-03-20","arxiv_id":"2503.15798","n_code_links":1,"syntology":{"ran":2,"of":3,"n_ran_checked":0,"n_instrument":2,"unverified":1,"pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified","official":{"repos":["jieshibo/mole"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/diffmoe-dynamic-token-selection-for-scalable","slug":"diffmoe-dynamic-token-selection-for-scalable","title":"DiffMoE: Dynamic Token Selection for Scalable Diffusion Transformers","date":"2025-03-18","arxiv_id":"2503.14487","n_code_links":0,"syntology":null},{"paper":null,"slug":"frac-connections-fractional-extension-of","title":"Frac-Connections: Fractional Extension of Hyper-Connections","date":"2025-03-18","arxiv_id":"2503.14125","n_code_links":0,"syntology":null},{"paper":"/paper/reinforcement-learning-based-motion-imitation","slug":"reinforcement-learning-based-motion-imitation","title":"Reinforcement learning-based motion imitation for physiologically plausible musculoskeletal motor control","date":"2025-03-18","arxiv_id":"2503.14637","n_code_links":1,"syntology":null},{"paper":"/paper/samoyeds-accelerating-moe-models-with","slug":"samoyeds-accelerating-moe-models-with","title":"Samoyeds: Accelerating MoE Models with Structured Sparsity Leveraging Sparse Tensor Cores","date":"2025-03-13","arxiv_id":"2503.10725","n_code_links":1,"syntology":null},{"paper":null,"slug":"double-stage-feature-level-clustering-based","title":"Double-Stage Feature-Level Clustering-Based Mixture of Experts Framework","date":"2025-03-12","arxiv_id":"2503.09504","n_code_links":0,"syntology":null},{"paper":null,"slug":"moe-gen-high-throughput-moe-inference-on-a","title":"MoE-Gen: High-Throughput MoE Inference on a Single GPU with Module-Based Batching","date":"2025-03-12","arxiv_id":"2503.09716","n_code_links":0,"syntology":null},{"paper":null,"slug":"priority-aware-preemptive-scheduling-for","title":"Priority-Aware Preemptive Scheduling for Mixed-Priority Workloads in MoE Inference","date":"2025-03-12","arxiv_id":"2503.09304","n_code_links":0,"syntology":null},{"paper":null,"slug":"accelerating-moe-model-inference-with-expert","title":"Accelerating MoE Model Inference with Expert Sharding","date":"2025-03-11","arxiv_id":"2503.08467","n_code_links":0,"syntology":null},{"paper":null,"slug":"moe-loco-mixture-of-experts-for-multitask","title":"MoE-Loco: Mixture of Experts for Multitask Locomotion","date":"2025-03-11","arxiv_id":"2503.08564","n_code_links":0,"syntology":null},{"paper":"/paper/a-comprehensive-survey-of-mixture-of-experts","slug":"a-comprehensive-survey-of-mixture-of-experts","title":"A Comprehensive Survey of Mixture-of-Experts: Algorithms, Theory, and Applications","date":"2025-03-10","arxiv_id":"2503.07137","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":0,"n_instrument":1,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["deepseek-ai/DeepEP"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/resmoe-space-efficient-compression-of-mixture","slug":"resmoe-space-efficient-compression-of-mixture","title":"ResMoE: Space-efficient Compression of Mixture of Experts LLMs via Residual Restoration","date":"2025-03-10","arxiv_id":"2503.06881","n_code_links":1,"syntology":{"ran":10,"of":15,"n_ran_checked":7,"n_instrument":3,"unverified":5,"pointer_only":1,"phrase":"10 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 1 violated, 6 with no contract checked; 3 where Syntology's instrument failed) · 5 unverified","official":{"repos":["idea-isail-lab-uiuc/resmoe"],"state":"official (archive's flag): 10 ran","n_ran":10,"n_constructed":0,"n_ran_no_instrument_failure":7,"n_unverified":5,"ran_from_kinds":["official"]}}},{"paper":null,"slug":"mofe-mixture-of-frozen-experts-architecture","title":"MoFE: Mixture of Frozen Experts Architecture","date":"2025-03-09","arxiv_id":"2503.06491","n_code_links":0,"syntology":null},{"paper":null,"slug":"genieblue-integrating-both-linguistic-and","title":"GenieBlue: Integrating both Linguistic and Multimodal Capabilities for Large Language Models on Mobile Devices","date":"2025-03-08","arxiv_id":"2503.06019","n_code_links":0,"syntology":null},{"paper":null,"slug":"capacity-aware-inference-mitigating-the","title":"Capacity-Aware Inference: Mitigating the Straggler Effect in Mixture of Experts","date":"2025-03-07","arxiv_id":"2503.05066","n_code_links":0,"syntology":null}],"papers_shown":100,"papers_order":"date (newest first), then slug","paper_counts":{"archive_papers_tagged":366,"with_a_code_link":142,"where_syntology_ran_a_sample":64,"not_listed_spam_title":0,"listed":366,"listed_where_code_ran":64,"where_syntology_ran_a_sample_split":{"with_a_run_with_no_instrument_failure":48,"every_run_a_failure_of_syntologys_instrument":16,"listed_with_a_run_with_no_instrument_failure":48,"listed_every_run_a_failure_of_syntologys_instrument":16,"filter":{"states":["a run with no instrument failure","any run, instrument failures included"],"default":"a run with no instrument failure","note":"on the 'only where code ran' pages the default hides, in the browser, the rows where every run was a failure of Syntology's instrument; the second state shows them again. Rows are hidden, never re-ordered; these twins list every row"}},"definition":"distinct papers the archive tags; 'where Syntology ran a sample' counts papers with at least one harvested sample that ran, which is not a correctness claim"},"papers_where_code_ran":64,"papers_where_code_ran_first_page":"/method/moe/papers/ran/1","papers_where_code_ran_caption":"We ran code from the paper's repository; we did not isolate this method inside it.","pages":[{"order":"archive","page":2,"of":4,"url":"/method/moe/papers/2","json":"/method/moe/papers/2.json","rows":100},{"order":"archive","page":3,"of":4,"url":"/method/moe/papers/3","json":"/method/moe/papers/3.json","rows":100},{"order":"archive","page":4,"of":4,"url":"/method/moe/papers/4","json":"/method/moe/papers/4.json","rows":66},{"order":"ran","page":1,"of":1,"url":"/method/moe/papers/ran/1","json":"/method/moe/papers/ran/1.json","rows":64}],"tasks":[{"task":"/task/mixture-of-experts","name":"Mixture-of-Experts","papers":312},{"task":null,"name":"GPU","papers":38},{"task":"/task/language-modelling","name":"Language Modelling","papers":34},{"task":"/task/computational-efficiency","name":"Computational Efficiency","papers":30},{"task":"/task/language-modeling","name":"Language Modeling","papers":29},{"task":"/task/large-language-model","name":"Large Language Model","papers":16},{"task":"/task/quantization","name":"Quantization","papers":14},{"task":null,"name":"CPU","papers":13},{"task":"/task/multi-task-learning","name":"Multi-Task Learning","papers":13},{"task":"/task/image-classification","name":"Image Classification","papers":12},{"task":"/task/parameter-efficient-fine-tuning","name":"parameter-efficient fine-tuning","papers":12},{"task":"/task/scheduling","name":"Scheduling","papers":11},{"task":"/task/mmlu","name":"MMLU","papers":10},{"task":"/task/question-answering","name":"Question Answering","papers":10},{"task":"/task/decoder","name":"Decoder","papers":9},{"task":"/task/diversity","name":"Diversity","papers":9},{"task":"/task/image-classification","name":"image-classification","papers":9},{"task":"/task/continual-learning","name":"Continual Learning","papers":7},{"task":"/task/math","name":"Math","papers":7},{"task":"/task/retrieval","name":"Retrieval","papers":7}],"tasks_shown":20,"n_tasks":254,"usage_by_year":[{"year":"2022","papers":1},{"year":"2024","papers":206},{"year":"2025","papers":159}],"row_source":"methods_table","archive":{"source":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","archive_url":"https://paperswithcode.com/method/moe"},"syntology_read_at":"2026-10-05T00:36:11+00:00","syntology":{"source":"Syntology's have check for this name (the free coverage lookup), read from Syntology's graph","have_url":"https://syntology.ai/api/graph/have?x=MoE","read_at":"2026-10-05T00:59:16+00:00","query":"MoE","kind":"method","have":"declared_code","tier":"repo_known","verified_implementation":false,"level":null,"level_meaning":"the served verification level, per answer; never summed across methods","resolved_to":{"method":"MoE","origin_arxiv_id":"2505.24449"},"resolution":"exact","resolved_name_differs":false,"absence_asserted":false,"absence_unchecked":null,"no_code_is_about_this_method":false,"subject":"different_paper","archive_paper_arxiv_id":"2403.07743","repositories":{"listed":1,"in_graph":1,"list_truncated":false,"strongest_link":"official"},"retrieve_with":null,"name_competition":{"verdict":"outranked","may_be_wrong_sense":true,"bearers_found":2,"served":{"arxiv_id":"2505.24449","citations":6},"leading_other_bearer":{"arxiv_id":"1701.06538","title":"Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer","citations":5455,"named_by":["acronym_expansion"],"bearer_name":"Sparsely-Gated Mixture-of-Experts layer (MoE)","ask_with":"have('1701.06538')"},"n_other_bearers_listed":1,"note":"'MoE' is ALSO the name of 'Sparsely-Gated Mixture-of-Experts layer (MoE)', a method in 'Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer' (arXiv:1701.06538, 5,455 citations), which is cited FAR more than the paper this answer is about (6). This answer may be about the wrong MoE. If you meant the better-known one, ask again with that entry of `also_named`."},"name_competition_text":"The same name also belongs to “Sparsely-Gated Mixture-of-Experts layer (MoE)”, a method in the paper “Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer” (arXiv:1701.06538, 5,455 citations), which the lookup finds cited far more than the paper this answer is about (6 citations), so this answer may be about a different sense of the name.","near_matches":null,"code_nearby":[{"name":"Soft MoE","relation":"shares_name_tokens","relation_means":"shares name tokens with this name; a different method","level":3,"retrieve_with":"get_reference_implementation(name='Soft MoE')","origin_arxiv_id":"2308.00951"},{"name":"TIME-MOE","relation":"shares_name_tokens","relation_means":"shares name tokens with this name; a different method","level":3,"retrieve_with":"get_reference_implementation(name='TIME-MOE')","origin_arxiv_id":"2409.16040"}],"code_nearby_caption":"Code Syntology holds under other names, as the lookup lists them (code_nearby); not a complete list, and none of them is this method.","origin_paper":{"arxiv_id":"2505.24449","page":"/paper/when-large-multimodal-models-confront","syntology_url":"https://syntology.ai/paper/2505.24449","in_graph_layer":true,"samples":6,"ran":4,"n_ran_checked":0,"n_instrument":4,"unverified":2,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 2 unverified","instrument_means":"failures of Syntology's instrument, not of the code"},"line":"about another paper, arXiv:2505.24449: repository known, no verified implementation (declared_code; repositories: 1 listed of 1 in the graph, strongest link: official). About another paper: this answer is about arXiv:2505.24449, while the archive credits arXiv:2403.07743.","level_legend":{"state":"quoted","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","card_sha256":"479785a84658dd682a11b10237b3620d651d7f87799f5dc7add970aaa83f6d39","note":"Syntology's served verification levels, quoted verbatim from its MCP server card; a row is split where the card separates clauses with a double hyphen, and the separator is not reproduced","levels":[{"level":"V0","quoted":["imports"]},{"level":"V1","quoted":["executes against a synthesized fixture and returns a value of the recorded type"]},{"level":"V2","quoted":["two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs","neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)"]},{"level":"V3","quoted":["paper-derived property tests pass on both of those same two generations."]}],"no_rung":"No rung on this ladder compares the served code against the authors' implementation, so treat every level as evidence about runnability and internal agreement, not about matching the paper's results."}},"syntology_have_state":"loaded","record_sha256":"0b9556463dea26487ce6e90051bd90015a980375c0f93e01258e10c1e851ff5b","record_changed_at":"2026-10-05","record_changed_at_basis":"changed"}