mlsanityData to 14 Oct 2026

RoboFL: Federated Expert Assembly for World Action Models

arXiv 2609.34968 ↗First posted 28 Sep 2026cs.RO cs.AIcited by 0

On the map: World models → World models for language agents also fits: Embodied and robot world models

Abstract (arXiv)

Vision-language-action and world-action models are increasingly popular, yet remain bottlenecked by physical interaction data that is scarce, institutionally siloed, and task-heterogeneous. A natural federated solution is to let each client adapt a shared foundation model through parameter-efficient fine-tuning, avoiding the exchange of full-model updates. However, federating these adapters is nontrivial, as naive aggregation can entangle incompatible updates, while incorporating MoE-style routing into federated aggregation may dilute specialization and destabilize expert selection. We present RoboFL, which instantiates MoSAIC (Mixture of Slotted Adapters) for federated world-action learning. MoSAIC directly installs locally trained LoRA adapters as the expert branches of a server MoE. Server-side routers learn token assignments over these prior-informed branches while jointly refining routing and expert parameters. Foresight-to-Action Routing Distillation (FARD) aligns routing across the model's three paths, while Path-Consensus Expert Aggregation (PCEA) converts complete expert updates into a compact global adapter for personalized redistribution. Experiments on RoboTwin 2.0, RLBench, and a real-world Franka robot arm show the superiority of RoboFL with structured expert assembly, as it outperforms centralized PEFT InternVLA-A1 by 12.23% on the Franka arm, while reducing per-round client communication by up to 86.81% relative to MoE-based federated VLA baselines.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in World models for language agents, one of 11 directions in World models. Builds on 44 of its 47 references that are in the map. First posted on arXiv 28 Sep 2026.

Builds on

44 of 47 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025Cosmos World Foundation Model Platform for Physical AI7 Jan

Other references · 43

  1. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  2. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,669
  3. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  4. 2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 3,073
  5. 2019Advances and Open Problems in Federated Learningcited by 3,046
  6. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 2,887
  7. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 2,407
  8. 2023PaLM-E: An Embodied Multimodal Language Modelcited by 2,071
  9. 2024Octo: An Open-Source Generalist Robot Policycited by 1,484
  10. 2017Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semanticscited by 1,358
  11. 2025GR00T N1: An Open Foundation Model for Generalist Humanoid Robotscited by 1,225
  12. 2020Adaptive Federated Optimizationcited by 1,046
31 more
  1. 2019On the Variance of the Adaptive Learning Rate and Beyondcited by 772
  2. 2019RLBench: The Robot Learning Benchmark & Learning Environmentcited by 689
  3. 2020Ensemble Distillation for Robust Model Fusion in Federated Learningcited by 682
  4. 2025FAST: Efficient Action Tokenization for Vision-Language-Action Modelscited by 665
  5. 2025SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Modelcited by 519
  6. 2026World Action Models are Zero-shot Policiescited by 465
  7. 2021Data-Free Knowledge Distillation for Heterogeneous Federated Learningcited by 383
  8. 2020Three Approaches for Personalization with Applications to Federated Learningcited by 325
  9. 2025Motus: A Unified Latent Action World Modelcited by 251
  10. 2024Mixture of LoRA Expertscited by 181
  11. 2020FedBE: Making Bayesian Model Ensemble Applicable to Federated Learningcited by 176
  12. 2024RoboMIND: Benchmark on Multi-embodiment Intelligence Normative Data for Robot Manipulationcited by 175
  13. 2024Mixture-of-Transformers: A Sparse and Scalable Architecture for Multi-Modal Foundation Modelscited by 163
  14. 2024FLoRA: Federated Fine-Tuning Large Language Models with Heterogeneous Low-Rank Adaptationscited by 146
  15. 2026GigaWorld-Policy: An Efficient Action-Centered World--Action Modelcited by 139
  16. 2025RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twinscited by 121
  17. 2025ForceVLA: Enhancing VLA Models with a Force-aware MoE for Contact-rich Manipulationcited by 91
  18. 2024Selective Aggregation for Low-Rank Adaptation in Federated Learningcited by 80
  19. 2025MoLe-VLA: Dynamic Layer-skipping Vision Language Action Model via Mixture-of-Layers for Efficient Robot Manipulationcited by 70
  20. 2026InternVLA-A1: Unifying Understanding, Generation and Action for Robotic Manipulationcited by 50
  21. 2024FedMoE: Personalized Federated Learning via Heterogeneous Mixture of Expertscited by 27
  22. 2025Robobench: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models as Embodied Braincited by 25
  23. 2024Towards Robust and Efficient Federated Low-Rank Adaptation with Heterogeneous Clientscited by 22
  24. 2025FedVLA: Federated Vision-Language-Action Learning with Dual Gating Mixture-of-Experts for Robotic Manipulationcited by 18
  25. 2024Mixture of Experts Made Personalized: Federated Prompt Learning for Vision-Language Modelscited by 17
  26. 2024FedFisher: Leveraging Fisher Information for One-Shot Federated Learningcited by 15
  27. 2025Expertise need not monopolize: Action-Specialized Mixture of Experts for Vision-Language-Action Learningcited by 12
  28. 2024T-REX: Mixture-of-Rank-One-Experts with Semantic-aware Intuition for Multi-task Large Language Model Finetuningcited by 8
  29. 2026ω-0: A Latent Predictive World Action Model for Concurrent Humanoid Loco-Manipulationcited by 6
  30. 2025FLAME: A Federated Learning Benchmark for Robotic Manipulationcited by 4
  31. 2026ForgeVLA: Federated Vision-Language-Action Learning without Language Annotationscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026FedWorld: Scope-Aware Federation of Agent World Models
  2. 2026UniW(A)M: A Unified Mixture-of-Experts Framework for World and World-Action Models
  3. 2026An Empirical Study and Open Testbed for Federated Fine-Tuning of Vision-Language-Action Models
  4. 2026LeapBot-JAC: Joint Action Chains for World Action Models
  5. 2026PFAdapter: Hierarchical LoRA Decomposition for Personalized Federated MLLMs
  6. 2026Federated Subspace Guided Vision-Language-Action Policy Distillation for Non-IID Multi-Robot Manipulation
  7. 2025MoIRA: Modular Instruction Routing Architecture for Multi-Task Robotics
  8. 2026Test-Time Mixture of World Models for Embodied Agents in Dynamic Environments