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

arXiv: https://arxiv.org/abs/2603.08476 · HTML: https://mlsanity.com/paper/2603.08476/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data.
- Builds on 20 of its 23 references that are in the map.
- First posted on arXiv 9 Mar 2026.

## Placed in

- Offline RL and imitation learning from data / [Robot manipulation imitation learning](https://mlsanity.com/offlinerl/robot-manipulation-imitation-learning/)

## Builds on (most cited first, 20 of 20)

- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [Diffusion Policy: Visuomotor Policy Learning via Action Diffusion](https://mlsanity.com/paper/2303.04137/) (2023)
- [Mixtral of Experts](https://mlsanity.com/paper/2401.04088/) (2024)
- [π0.5: a Vision-Language-Action Model with Open-World Generalization](https://mlsanity.com/paper/2504.16054/) (2025)
- [Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware](https://mlsanity.com/paper/2304.13705/) (2023)
- [Octo: An Open-Source Generalist Robot Policy](https://mlsanity.com/paper/2405.12213/) (2024)
- [MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers](https://mlsanity.com/paper/2002.10957/) (2020)
- [LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning](https://mlsanity.com/paper/2306.03310/) (2023)
- [Scaling Vision with Sparse Mixture of Experts](https://mlsanity.com/paper/2106.05974/) (2021)
- [What Matters in Learning from Offline Human Demonstrations for Robot Manipulation](https://mlsanity.com/paper/2108.03298/) (2021)
- [SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics](https://mlsanity.com/paper/2506.01844/) (2025)
- [MoE-LLaVA: Mixture of Experts for Large Vision-Language Models](https://mlsanity.com/paper/2401.15947/) (2024)
- [EdgeNeXt: Efficiently Amalgamated CNN-Transformer Architecture for Mobile Vision Applications](https://mlsanity.com/paper/2206.10589/) (2022)
- [Mixtures of Experts Unlock Parameter Scaling for Deep RL](https://mlsanity.com/paper/2402.08609/) (2024)
- [Demonstration-Guided Reinforcement Learning with Learned Skills](https://mlsanity.com/paper/2107.10253/) (2021)
- [Skill-based Model-based Reinforcement Learning](https://mlsanity.com/paper/2207.07560/) (2022)
- [Supervised Mixture-of-Experts for Surgical Grasping and Retraction](https://mlsanity.com/paper/2601.21971/) (2026)
- [Mixture of Group Experts for Learning Invariant Representations](https://mlsanity.com/paper/2504.09265/) (2025)

## Built on by (most cited first, 1 of 1)

- [From Articulated Kinematics to Routed Visual Control for Action-Conditioned Surgical Video Generation](https://mlsanity.com/paper/2605.08712/) (2026)

## Abstract (arXiv)

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.