Metraipolis › archive › dc_9ef61d179e8a
PFedMA: Who Should Join Next as Federated Multi-Agent Orchestration Evolves? — arXiv:2610.19458
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.19458
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T20:01:37.934Z
- SHA-256 of the archived text
0ddde432abe106cf88871ce4a520a512db5defd3656e6e7455a04031938bdda7- Stable URL of this page
https://metraipolis.com/documento/dc_9ef61d179e8a
How to cite
PFedMA: Who Should Join Next as Federated Multi-Agent Orchestration Evolves? — arXiv:2610.19458 — arXiv (en), https://arxiv.org/abs/2610.19458 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9ef61d179e8a (Metraipolis archive dc_9ef61d179e8a, sha256 0ddde432abe106cf88871ce4a520a512db5defd3656e6e7455a04031938bdda7)
BibTeX
@misc{dc9ef61d179e8a,
title = {PFedMA: Who Should Join Next as Federated Multi-Agent Orchestration Evolves? — arXiv:2610.19458},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.19458},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9ef61d179e8a; sha256 0ddde432abe106cf88871ce4a520a512db5defd3656e6e7455a04031938bdda7},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19458. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Federated Multi-Agent Systems (FedMAS) offer a promising paradigm for collaborative reasoning among LLM-based agents while keeping raw client data decentralized. However, reliable multi-round collaboration remains challenging. Agent selection must take into consideration evolving task requirements and the agent's historical reliability, while privacy rules must adapt to changing collaboration contexts and meet the client's personalized needs. To address these challenges, we propose PFedMA, a federated multi-agent framework for feedback driven multi-round collaboration under heterogeneous privacy requirements. Across rounds, PFedMA maintains historical capability and reliability states to guide agent selection, evaluates task responses before aggregation, and refines subsequent queries according to accumulated collaboration feedback. PFedMA enables bidirectional local to global and global to local private rule evolution across heterogeneous clients. Clients generalize locally updated privacy rules and selectively incorporate globally maintained rules, thereby facilitating cross-client privacy rule evolution while preserving personalized constraints. Extensive experiments across reasoning, privacy sensitive, and adversarial settings show that PFedMA improves task performance, reduces measured privacy leakage, and improves system robustness. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.19458 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-15T19:35:48.236Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).