Metraipolis › archive › dc_f0420eb6d306

FARM: Fundamental Agentic Reward Model For Multi-task Wireless Network Optimization — arXiv:2610.02947

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.02947
License (as archived)
CC0-1.0
Archived
2026-10-06T11:41:57.115Z
SHA-256 of the archived text
afa438532af287686f44ce1c12362a6c907ef0f3431e74689ecf9b4b0b02aacc
Stable URL of this page
https://metraipolis.com/documento/dc_f0420eb6d306

How to cite

FARM: Fundamental Agentic Reward Model For Multi-task Wireless Network Optimization — arXiv:2610.02947 — arXiv (en), https://arxiv.org/abs/2610.02947 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f0420eb6d306 (Metraipolis archive dc_f0420eb6d306, sha256 afa438532af287686f44ce1c12362a6c907ef0f3431e74689ecf9b4b0b02aacc)
BibTeX
@misc{dcf0420eb6d306,
  title = {FARM: Fundamental Agentic Reward Model For Multi-task Wireless Network Optimization — arXiv:2610.02947},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.02947},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f0420eb6d306; sha256 afa438532af287686f44ce1c12362a6c907ef0f3431e74689ecf9b4b0b02aacc},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.02947. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Future wireless networks require learning agents to adapt across heterogeneous channel conditions, traffic patterns, quality-of-service (QoS) requirements, objectives, and operational constraints. Reusing decision knowledge across such tasks is challenging because conventional multi-task and transfer reinforcement learning methods primarily share or transfer policies, coupling transferable knowledge with task-dependent action mappings. This paper proposes FARM (Fundamental Agentic Reward Model for Multi-task Wireless Network Optimization), a reward-space transfer framework that shifts cross-task knowledge reuse from policy space to trajectory-level decision evaluation. FARM introduces an Agentic Reward Model (ARM) that learns a task-conditioned reward prior from heterogeneous source-task trajectories and provides auxiliary guidance for task-specific policy optimization. In Stage I, ARM jointly models task conditions, temporal trajectory dependencies, and objective-dependent reward structures while each source task retains its own controller. In Stage II, the learned reward prior is frozen and reused to guide the adaptation of a target-specific controller for previously unseen tasks, without transferring source-task policies. Experiments on heterogeneous multi-access edge computing (MEC) tasks show that FARM achieves a mean late-stage gain of 29.8% over Single-task SAC on unseen Rate-Latency targets, compared with 16.1% for CRA Transfer, and reaches a 46.2% gain on the moderate-OOD FAR-M case. Further analysis shows that both Mamba and Transformer trajectory encoders support Reward-Space Transfer, while Mamba provides improved robustness as longer history dependencies are introduced.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-10-05 in eess.SY).
Categorie arXiv: Systems and Control (eess.SY).
Pagina del lavoro: https://arxiv.org/abs/2610.02947 — 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 2026-10-06T10:05:37.594Z dalla pagina https://arxiv.org/catchup/eess.SY/2026-10-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).