Semantically Structured Mixture-of-Experts for Compositional Robotic Manipulation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Deng, Chengyu, Chen, Guanqi, Chen, Yizhou, Liu, Zejia, Ruan, Zhiwen, Chen, Guanhua, Pan, Jia
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918518270722048
author Deng, Chengyu
Chen, Guanqi
Chen, Yizhou
Liu, Zejia
Ruan, Zhiwen
Chen, Guanhua
Pan, Jia
author_facet Deng, Chengyu
Chen, Guanqi
Chen, Yizhou
Liu, Zejia
Ruan, Zhiwen
Chen, Guanhua
Pan, Jia
contents Diffusion-based policies have established a new standard for precise robotic manipulation but face a critical scalability bottleneck: high-performance models are computationally expensive, while lightweight alternatives often fail to generalize across diverse multi-task environments. Mixture-of-Experts (MoE) architectures offer a promising path to efficiency by activating only a subset of parameters. However, existing MoE routing mechanisms typically rely on low-level noise or latent statistics, ignoring the compositional nature of manipulation tasks. This can fragment reusable behaviors across experts, limiting interpretability and transferability. We introduce Semantically Structured Mixture-of-Experts Diffusion Policy (SMoDP) for compositional robotic manipulation, a framework that grounds expert specialization in semantic task structure. SMoDP leverages a lightweight, inference-time skill predictor, supervised by offline annotations from Vision-Language Models (VLMs), to route action chunks to experts specialized for specific behavioral phases. To ensure robust assignment, we propose a dual contrastive alignment strategy that grounds multi-modal observations in language-defined skill semantics (Inter-modal) while enforcing routing consistency across visually distinct but functionally related behaviors (Intra-modal). Our approach outperforms representative diffusion and MoE-based baselines on multi-task benchmarks with significantly improved parameter efficiency and demonstrates effective compositional transfer to novel tasks through parameter-efficient fine-tuning. Project website: https://deng-cy20.github.io/SMoDP/
format Preprint
id arxiv_https___arxiv_org_abs_2605_23477
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Semantically Structured Mixture-of-Experts for Compositional Robotic Manipulation
Deng, Chengyu
Chen, Guanqi
Chen, Yizhou
Liu, Zejia
Ruan, Zhiwen
Chen, Guanhua
Pan, Jia
Robotics
Diffusion-based policies have established a new standard for precise robotic manipulation but face a critical scalability bottleneck: high-performance models are computationally expensive, while lightweight alternatives often fail to generalize across diverse multi-task environments. Mixture-of-Experts (MoE) architectures offer a promising path to efficiency by activating only a subset of parameters. However, existing MoE routing mechanisms typically rely on low-level noise or latent statistics, ignoring the compositional nature of manipulation tasks. This can fragment reusable behaviors across experts, limiting interpretability and transferability. We introduce Semantically Structured Mixture-of-Experts Diffusion Policy (SMoDP) for compositional robotic manipulation, a framework that grounds expert specialization in semantic task structure. SMoDP leverages a lightweight, inference-time skill predictor, supervised by offline annotations from Vision-Language Models (VLMs), to route action chunks to experts specialized for specific behavioral phases. To ensure robust assignment, we propose a dual contrastive alignment strategy that grounds multi-modal observations in language-defined skill semantics (Inter-modal) while enforcing routing consistency across visually distinct but functionally related behaviors (Intra-modal). Our approach outperforms representative diffusion and MoE-based baselines on multi-task benchmarks with significantly improved parameter efficiency and demonstrates effective compositional transfer to novel tasks through parameter-efficient fine-tuning. Project website: https://deng-cy20.github.io/SMoDP/
title Semantically Structured Mixture-of-Experts for Compositional Robotic Manipulation
topic Robotics
url https://arxiv.org/abs/2605.23477