MoE in reinforcement learning and robotics
Key facts. MoE in reinforcement learning and robotics: 151 papers on arXiv since 2018 (data through 7 Oct 2026). 9 of its 151 mapped papers (6%) cite “Mixtures of Experts Unlock Parameter Scaling for Deep RL” (2024, arXiv 2402.08609). 8 of its 151 mapped papers (5%) cite “DriveMoE: Mixture-of-Experts for Vision-Language-Action Model in End-to-End Autonomous Driving” (2025, arXiv 2505.16278). 2 new papers 17 Sep–7 Oct.
What’s new
17 Sep – 7 Oct 2026
2 new since 17 Sep
Foundations
Most cited by papers here
- Mixtures of Experts Unlock Parameter Scaling for Deep RL2024cited by 9 here
- DriveMoE: Mixture-of-Experts for Vision-Language-Action Model in End-to-End Autonomous Driving2025cited by 8 here
- MENTOR: Mixture-of-Experts Network with Task-Oriented Perturbation for Visual Reinforcement Learning2024cited by 7 here
- Group Sequence Policy Optimization2025cited by 6 here
- Stabilizing MoE Reinforcement Learning by Aligning Training and Inference Routers2025cited by 6 here
Lineage
The papers this direction builds on, oldest first.
Each paper on the line cites the one before.
- 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 42 here
- 2020GShard: Scaling Giant Models with Conditional Computation and Automatic Shardingcited by 31 here
- 2021Scaling Vision with Sparse Mixture of Expertscited by 17 here
- 2022ST-MoE: Designing Stable and Transferable Sparse Expert Modelscited by 9 here
- 2022Multimodal Contrastive Learning with LIMoE: the Language-Image Mixture of Expertscited by 7 here
- 2023From Sparse to Soft Mixtures of Expertscited by 7 here
- 2024A Survey on Mixture of Experts in Large Language Modelscited by 5 here
- 2025DriveMoE: Mixture-of-Experts for Vision-Language-Action Model in End-to-End Autonomous Drivingcited by 8 here
16 more it builds on, not in this chain
- 2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 41 here
- 2021GLaM: Efficient Scaling of Language Models with Mixture-of-Expertscited by 13 here
- 2022Mixture-of-Experts with Expert Choice Routingcited by 10 here
- 2024Mixtral of Expertscited by 15 here
- 2024DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Modelscited by 19 here
- 2024Mixtures of Experts Unlock Parameter Scaling for Deep RLcited by 9 here
- 2024LLaMA-MoE: Building Mixture-of-Experts from LLaMA with Continual Pre-trainingcited by 5 here
- 2024Mixture of Experts in a Mixture of RL settingscited by 5 here
- 2024MENTOR: Mixture-of-Experts Network with Task-Oriented Perturbation for Visual Reinforcement Learningcited by 7 here
- 2024Qwen2.5 Technical Reportcited by 5 here
- 2024DeepSeek-V3 Technical Reportcited by 15 here
- 2025A Comprehensive Survey of Mixture-of-Experts: Algorithms, Theory, and Applicationscited by 5 here
- 2025Qwen3 Technical Reportcited by 14 here
- 2025Group Sequence Policy Optimizationcited by 6 here
- 2025Stabilizing MoE Reinforcement Learning by Aligning Training and Inference Routerscited by 6 here
- 2025Qwen3-VL Technical Reportcited by 6 here
The 24 most cited of 200.
Browse
30 of 151, ranked 6–35
- 6Mixture of Experts in a Mixture of RL settings20245
- 7Kimi K2: Open Agentic Intelligence20254
- 8Probabilistic Mixture-of-Experts for Efficient Deep Reinforcement Learning20214
- 9MoE-Loco: Mixture of Experts for Multitask Locomotion20253
- 10MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention20253
- 11GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models20253
- 12ReLibra: Routing-Replay-Guided Load Balancing for MoE Training in Reinforcement Learning20262
- 13Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence20262
- 14ForceVLA: Enhancing VLA Models with a Force-aware MoE for Contact-rich Manipulation20252
- 15Mastering Massive Multi-Task Reinforcement Learning via Mixture-of-Expert Decision Transformer20252
All 30
- 16GEMINUS: Dual-aware Global and Scene-Adaptive Mixture-of-Experts for End-to-End Autonomous Driving20252
- 17Towards Stable and Effective Reinforcement Learning for Mixture-of-Experts20252
- 18Stabilizing Reinforcement Learning with LLMs: Formulation and Practices20252
- 19Nemotron 3 Nano: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning20252
- 20Acquiring Diverse Skills using Curriculum Reinforcement Learning with Mixture of Experts20242
- 21GeRM: A Generalist Robotic Model with Mixture-of-experts for Quadruped Robot20242
- 22MiniDrive: More Efficient Vision-Language Models with Multi-Level 2D Features as Text Tokens for Autonomous Driving20242
- 23MiMo-V2-Flash Technical Report20261
- 24Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation20261
- 25MoE-GRPO: Optimizing Mixture-of-Experts via Reinforcement Learning in Vision-Language Models20261
- 26Nemotron 3 Super: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning20261
- 27Harnessing Routing Foresight for Micro-step-level MoE load balancing in RL Post-training20261
- 28InsightDrive: Insight Scene Representation for End-to-End Autonomous Driving20251
- 29Mixture-of-Experts Meets In-Context Reinforcement Learning20251
- 30Ring-lite: Scalable Reasoning via C3PO-Stabilized Reinforcement Learning for LLMs20251
- 31MoSE: Skill-by-Skill Mixture-of-Experts Learning for Embodied Autonomous Machines20251
- 32CBDES MoE: Hierarchically Decoupled Mixture-of-Experts for Functional Modules in Autonomous Driving20251
- 33A Knowledge-Driven Diffusion Policy for End-to-End Autonomous Driving Based on Expert Routing20251
- 34LongScape: Advancing Long-Horizon Embodied World Models with Context-Aware MoE20251
- 35Sparse Diffusion Policy: A Sparse, Reusable, and Flexible Policy for Robot Learning20241