mlsanityData to 14 Oct 2026

Resource-Aware Federated Mixture-of-Experts with Adaptive Pruning for Onboard Learning in LEO Satellite Constellations

arXiv 2609.31932 ↗First posted 25 Sep 2026cs.LG cs.CVcited by 0

New in Mixture-of-experts; no direction yet. Mixture-of-experts

Abstract (arXiv)

Low-Earth-orbit (LEO) satellites are increasingly expected to perform onboard learning for applications such as disaster response and environmental monitoring. However, conventional federated learning (FL) is ill-suited to onboard satellite learning, as it assumes computational, memory, and communication resources beyond the capabilities of resource-constrained LEO platforms, often necessitating the transmission of raw imagery to ground stations. We present COSMIC-FL, a resource-aware FL framework for efficient onboard learning in LEO satellite constellations. COSMIC-FL introduces two complementary Mixture-of-Experts (MoE) architectures: a Sliced design that shares backbone representations while activating task-specific channel subsets, and a Modular design that employs lightweight gating to route inputs to physically separated expert networks. A semantic class-to-expert mapping enables each satellite to train, update, and communicate only the expert paths relevant to its local data. To further improve efficiency, COSMIC-FL integrates staged optimization with three structured pruning strategies: server-side pruning, client-side fixed-ratio pruning with mean-vote aggregation, and adaptive client-side per-layer pruning based on aggregated importance and a MAD-based gap criterion. Combined with semantic expert routing, these techniques jointly adapt computation and model sparsity to both data semantics and layer importance, yielding a favourable accuracy--efficiency trade-off for heterogeneous space platforms. Experiments on six image classification benchmarks under highly non-i.i.d. settings show that COSMIC-FL maintains competitive accuracy while reducing communication, computation, and energy consumption by up to 80% over SOTA FL methods. We further validate COSMIC-FL on an NVIDIA Jetson AGX Orin, confirming its efficiency gains under realistic embedded deployment constraints.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 16 of its 31 references that are in the map. First posted on arXiv 25 Sep 2026.

Builds on

16 of 31 references.

References · 16

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,669
  3. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
  4. 2018Federated Optimization in Heterogeneous Networkscited by 2,987
  5. 2015Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Codingcited by 2,329
  6. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,496
  7. 2019On the Convergence of FedAvg on Non-IID Datacited by 1,229
  8. 2017Remote Sensing Image Scene Classification: Benchmark and State of the Artcited by 797
  9. 2021Scaling Vision with Sparse Mixture of Expertscited by 747
  10. 2016AID: A Benchmark Dataset for Performance Evaluation of Aerial Scene Classificationcited by 347
  11. 2020HeteroFL: Computation and Communication Efficient Federated Learning for Heterogeneous Clientscited by 337
  12. 2018Expanding the Reach of Federated Learning by Reducing Client Resource Requirementscited by 163
4 more
  1. 2022Federated Learning in Satellite Constellationscited by 21
  2. 2024Mixture of Experts in Image Classification: What's the Sweet Spot?cited by 11
  3. 2023Optimizing Federated Learning in LEO Satellite Constellations via Intra-Plane Model Propagation and Sink Satellite Schedulingcited by 10
  4. 2023Secure and Efficient Federated Learning in LEO Constellations using Decentralized Key Generation and On-Orbit Model Aggregationcited by 6

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2023One-Shot Federated Learning for LEO Constellations that Reduces Convergence Time from Days to 90 Minutes
  2. 2026FractalNet-Based Heterogeneous Federated Learning for Orbital Edge Intelligence in Satellite Mega-Constellations: A Wildfire Case Study
  3. 2026On-Orbit Space AI: Federated, Multi-Agent, and Collaborative Algorithms for Satellite Constellations
  4. 2025A Semi-Supervised Federated Learning Framework with Hierarchical Clustering Aggregation for Heterogeneous Satellite Networks
  5. 2026SpaceMoE: Realizing Distributed Mixture-of-Experts Inference over Space Networks
  6. 2025OptiVote: Non-Coherent FSO Over-the-Air Majority Vote for Communication-Efficient Distributed Federated Learning in Space Data Centers
  7. 2026MLFCIL: A Multi-Level Forgetting Mitigation Framework for Federated Class-Incremental Learning in LEO Satellites
  8. 2024Towards Satellite Non-IID Imagery: A Spectral Clustering-Assisted Federated Learning Approach