Metraipolis › archive › dc_75fb0f4dcd53

Selective and Hybrid-State Communication for Latent Multi-Agent Systems — arXiv:2610.37779

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.37779
License (as archived)
CC0-1.0
Archived
2026-10-31T06:06:38.506Z
SHA-256 of the archived text
29dca7fddd9708ff2046ee4f3a0942b114824a9972e4f5621ff81377a8ebb12e
Stable URL of this page
https://metraipolis.com/documento/dc_75fb0f4dcd53

How to cite

Selective and Hybrid-State Communication for Latent Multi-Agent Systems — arXiv:2610.37779 — arXiv (en), https://arxiv.org/abs/2610.37779 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_75fb0f4dcd53 (Metraipolis archive dc_75fb0f4dcd53, sha256 29dca7fddd9708ff2046ee4f3a0942b114824a9972e4f5621ff81377a8ebb12e)
BibTeX
@misc{dc75fb0f4dcd53,
  title = {Selective and Hybrid-State Communication for Latent Multi-Agent Systems — arXiv:2610.37779},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.37779},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_75fb0f4dcd53; sha256 29dca7fddd9708ff2046ee4f3a0942b114824a9972e4f5621ff81377a8ebb12e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.37779. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Efficient communication remains a major bottleneck in large language model multiagent systems. Latent communication reduces repeated natural-language exchange by transferring internal model states between agents, but these transfers remain costly and are typically designed around conventional KV-cache representations. We address this problem from two complementary perspectives. We introduce SelectKV, a boundary-aware method that selectively communicates KV states while protecting inherited context, and HyLaMAS, which extends latent collaboration to hybrid linear-attention models by communicating both KV caches and recurrent states through a shared workspace. Across six benchmarks, SelectKV reduces communicated KV positions and logical KV payload by 9.7-9.9%, with accuracy unchanged on two benchmarks, within one percentage point on three, and six points lower on HumanEval+. Across Qwen3.5 model scales, HyLaMAS reduces decoded output tokens by 53-80% and achieves 1.16-2.78x speedups relative to text-based multi-agent communication. Together, these independently evaluated methods represent two complementary approaches to efficient latent-state communication across conventional Transformer and hybrid model architectures.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.MA, nuovi lavori del giorno).
Categorie arXiv: Multiagent Systems (cs.MA) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.37779 — 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 2026-10-31T06:06:38.404Z dalla pagina https://arxiv.org/list/cs.MA/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).