{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T08:48:01Z","timestamp":1782809281652,"version":"3.54.5"},"reference-count":0,"publisher":"ECMS","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,6,23]]},"abstract":"<jats:p>This paper presents an enhancement to Federated Learning (FL) for Artificial Neural Networks (ANNs), addressing limitations of conventional FL in noisy and heterogeneous environments with few participants. While a typical federated learning scenario involves a massive number of distributed nodes, this proposed method is specifically designed for situations where only a small number of participants or servers are available. Although this is not a typical FL case, it is highly critical in the context of cyber-security, where secure collaboration might be limited to just a few entities\u2014for example, one CERT team, one trusted third party, and one outer data source.\n\n\nFederated Learning enables distributed model training across these entities without data centralization, preserving privacy by aggregating local updates from clients that retain their sensitive, isolated data. The proposed approach introduces a strategy-selection-based aggregation mechanism inspired by Stackelberg game theory, in which the central server acts as the leader and the clients act as followers. By incorporating a game-theoretic interaction into the training process, the algorithm explicitly models the asymmetric power dynamics between the global and local models and deeply integrates client heterogeneity into the update mechanism.\n\n\nThe method is evaluated in the context of network intrusion detection, a domain where data sharing is often strictly restricted due to legal and security constraints. Experimental results demonstrate that the proposed algorithm enables effective cooperative learning among a strictly limited number of clients with highly incomplete and non-IID datasets, improving intrusion detection performance while strictly maintaining data confidentiality.<\/jats:p>","DOI":"10.7148\/2026-0338","type":"proceedings-article","created":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T08:36:46Z","timestamp":1782808606000},"page":"338-344","source":"Crossref","is-referenced-by-count":0,"title":["Modelling federated learning with strictly heterogeneous client data as a competitive stackelberg game for solving the intrusion detection problem"],"prefix":"10.7148","author":[{"given":"Agnieszka","family":"Jakobik","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"4144","published-online":{"date-parts":[[2026,6,23]]},"event":{"name":"40th ECMS International Conference on Modelling and Simulation"},"container-title":["ECMS 2026 Proceedings edited by Filippo Sanfilippo, Florenc Demrozi, Fabio Sgarbossa, Mohammad Poursina"],"original-title":[],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T08:36:56Z","timestamp":1782808616000},"score":1,"resource":{"primary":{"URL":"http:\/\/www.scs-europe.net\/dlib\/2026\/ecms2026acceptedpapers\/0338_secmos_ecms2026_0087.pdf"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,23]]},"references-count":0,"URL":"https:\/\/doi.org\/10.7148\/2026-0338","relation":{},"subject":[],"published":{"date-parts":[[2026,6,23]]}}}