混合专家论文趋势 2023–2026

主要结论

• 混合专家论文从 2023 年的 236 篇增长到 2026 年的 326 篇，三年间扩大到约 1.4 倍。
• NeurIPS 是发混合专家论文最多的会议，2023–2026 年共 354 篇，占四个会议总量的 39%。
• 交叉方向里涨得最快的是多模态大模型：同时属于两个主题的论文从 2023 年的 4 篇增至 2026 年的 79 篇。
• 在「优化与缩放」的 5 个主题里，它的 2026 年论文数排第 4；同组增长最快的缩放定律是 4.1 倍，混合专家是 1.4 倍。

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

• NeurIPS 2026A Biconvex Formulation for Stable Transport of Mixture Models with a Unique Solution
• ICML 2026$\phi$-Balancing for Mixture-of-Experts Training
• ICLR 2026A Federated Generalized Expectation-Maximization Algorithm for Mixture Models with an Unknown Number of Components
• CVPR 2026Divide Conquer and Aggregate Asymmetric Experts for Class-Imbalanced Semi-Supervised Medical
• NeurIPS 2025Advancing Expert Specialization for Better MoE
• ICML 2025$\texttt{I$^2$MoE}$: Interpretable Multimodal Interaction-aware Mixture-of-Experts
• ICLR 2025A Sanity Check for AI-generated Image Detection
• CVPR 2025Complexity Experts are Task-Discriminative Learners for Any Image Restoration
• NeurIPS 2024$\texttt{Model-GLUE}$: Democratized LLM Scaling for A Large Model Zoo in the Wild
• ICML 2024$\texttt{MoE-RBench}$: Towards Building Reliable Language Models with Sparse Mixture-of-Experts
• ICLR 2024Fusing Models with Complementary Expertise
• CVPR 2024MoDE: CLIP Data Experts via Clustering
• NeurIPS 2023[Re] Numerical influence of ReLU'(0) on backpropagation
• ICML 2023A Distribution Optimization Framework for Confidence Bounds of Risk Measures
• ICLR 2023Knowledge-in-Context: Towards Knowledgeable Semi-Parametric Language Models
• CVPR 2023Balanced Product of Calibrated Experts for Long-Tailed Recognition
• NeurIPS 2026AdERA: Adaptive Exponent Reuse for Lossless Allgather in Sharded MoE Training
• ICML 2026Attention Sink Forges Native MoE in Attention Layers: Sink-Aware Training to Address Head Collapse
• ICLR 2026Abstracting Robot Manipulation Skills via Mixture-of-Experts Diffusion Policies
• CVPR 2026ERMoE Eigen-Reparameterized Mixture-of-Experts for Stable Routing and Interpretable Specialization

全部混合专家论文

共 223 篇，第 2 / 5 页

51. ICLR 2025Mixture of Attentions For Speculative Decoding
52. ICLR 2025Mixture of Experts Made Personalized: Federated Prompt Learning for Vision-Language Models
53. ICLR 2025Mixture of In-Context Prompters for Tabular PFNs
54. ICLR 2025Mixture of Parrots: Experts improve memorization more than reasoning
55. ICLR 2025Mixture-of-Agents Enhances Large Language Model Capabilities
56. ICLR 2025MoE++: Accelerating Mixture-of-Experts Methods with Zero-Computation Experts
57. ICLR 2025MoLEx: Mixture of Layer Experts for Fine-tuning with Sparse Upcycling
58. ICLR 2025Monet: Mixture of Monosemantic Experts for Transformers
59. ICLR 2025More Experts Than Galaxies: Conditionally-Overlapping Experts with Biologically-Inspired Fixed Routing
60. ICLR 2025MoS: Unleashing Parameter Efficiency of Low-Rank Adaptation with Mixture of Shards
61. ICLR 2025Multiple Heads are Better than One: Mixture of Modality Knowledge Experts for Entity Representation Learning
62. ICLR 2025NetMoE: Accelerating MoE Training through Dynamic Sample Placement
63. ICLR 2025No Need to Talk: Asynchronous Mixture of Language Models
64. ICLR 2025OLMoE: Open Mixture-of-Experts Language Models
65. ICLR 2025pMoE: Prompting Diverse Experts Together Wins More in Visual Adaptation
66. ICLR 2025Provable Uncertainty Decomposition via Higher-Order Calibration
67. ICLR 2025RegMix: Data Mixture as Regression for Language Model Pre-training
68. ICLR 2025ReMoE: Fully Differentiable Mixture-of-Experts with ReLU Routing
69. ICLR 2025Revisiting Prefix-tuning: Statistical Benefits of Reparameterization among Prompts
70. ICLR 2025Routing Experts: Learning to Route Dynamic Experts in Existing Multi-modal Large Language Models
71. ICLR 2025Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts
72. ICLR 2025Solving Token Gradient Conflict in Mixture-of-Experts for Large Vision-Language Model
73. ICLR 2025Statistical Advantages of Perturbing Cosine Router in Mixture of Experts
74. ICLR 2025TC-MoE: Augmenting Mixture of Experts with Ternary Expert Choice
75. ICLR 2025Theory on Mixture-of-Experts in Continual Learning
76. ICLR 2025Tight Clusters Make Specialized Experts
77. ICLR 2025Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts
78. ICLR 2025Transformers Provably Learn Two-Mixture of Linear Classification via Gradient Flow
79. ICLR 2025Warm Diffusion: Recipe for Blur-Noise Mixture Diffusion Models
80. ICLR 2025Your Mixture-of-Experts LLM Is Secretly an Embedding Model for Free
81. ICML 2025$\texttt{I$^2$MoE}$: Interpretable Multimodal Interaction-aware Mixture-of-Experts
82. ICML 2025A Generic Family of Graphical Models: Diversity, Efficiency, and Heterogeneity
83. ICML 2025A Stronger Mixture of Low-Rank Experts for Fine-Tuning Foundation Models
84. ICML 2025Addressing Imbalanced Domain-Incremental Learning through Dual-Balance Collaborative Experts
85. ICML 2025Autonomy-of-Experts Models
86. ICML 2025Bayesian Inference for Correlated Human Experts and Classifiers
87. ICML 2025Chameleon: A Flexible Data-mixing Framework for Language Model Pretraining and Finetuning
88. ICML 2025Cooperation of Experts: Fusing Heterogeneous Information with Large Margin
89. ICML 2025DEFAME: Dynamic Evidence-based FAct-checking with Multimodal Experts
90. ICML 2025Delta Decompression for MoE-based LLMs Compression
91. ICML 2025Diff-MoE: Diffusion Transformer with Time-Aware and Space-Adaptive Experts
92. ICML 2025Domain2Vec: Vectorizing Datasets to Find the Optimal Data Mixture without Training
93. ICML 2025Dynamic Mixture of Curriculum LoRA Experts for Continual Multimodal Instruction Tuning
94. ICML 2025Expert Race: A Flexible Routing Strategy for Scaling Diffusion Transformer with Mixture of Experts
95. ICML 2025FlexiReID: Adaptive Mixture of Expert for Multi-Modal Person Re-Identification
96. ICML 2025FloE: On-the-Fly MoE Inference on Memory-constrained GPU
97. ICML 2025Gaussian Mixture Flow Matching Models
98. ICML 2025Graph Neural Network Generalization With Gaussian Mixture Model Based Augmentation
99. ICML 2025Joint MoE Scaling Laws: Mixture of Experts Can Be Memory Efficient
100. ICML 2025Learning Mixtures of Experts with EM: A Mirror Descent Perspective

常见问题

2023–2026 年四大会议一共有多少篇混合专家论文？

共 904 篇，按年份分别是 2023 年 236 篇、2024 年 119 篇、2025 年 223 篇、2026 年 326 篇。

NeurIPS 2026 有多少篇混合专家论文？

NeurIPS 2026 有 75 篇混合专家论文，上一届是 86 篇。

ICLR 2026 有多少篇混合专家论文？

ICLR 2026 有 73 篇混合专家论文，上一届是 59 篇。

ICML 2026 有多少篇混合专家论文？

ICML 2026 有 132 篇混合专家论文，上一届是 57 篇。

CVPR 2026 有多少篇混合专家论文？

CVPR 2026 有 46 篇混合专家论文，上一届是 21 篇。

哪个会议的混合专家论文最多？

NeurIPS，2023–2026 年共 354 篇。