{"paper":{"title":"Decentralized Ranking Aggregation via Gossip: Convergence and Robustness","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Autonomous agents reach a global ranking consensus from distributed preferences using only local random gossip interactions, without any central coordinator.","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Anna van Elst, Igor Colin, Kerrian Le Caillec, Stephan Cl\\'emen\\c{c}on","submitted_at":"2026-02-26T10:37:23Z","abstract_excerpt":"The concept of ranking aggregation plays a central role in preference analysis, and numerous algorithms for calculating median rankings, often originating in social choice theory, have been documented in the literature, offering theoretical guarantees in a centralized setting, \\textit{i.e.}, when all the ranking data to be aggregated can be brought together in a single computing unit. For many technologies (\\textit{e.g.} peer-to-peer networks, IoT, multi-agent systems), extending the ability to calculate consensus rankings with guarantees of convergence and resilience to potential contaminatio"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The approach proposed and analyzed here relies on the robustness guarantees offered by random gossip communication, which allows autonomous agents to compute a global ranking consensus using local interactions only, without coordination or a central authority.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That standard random gossip convergence properties for scalar statistics extend directly to the non-convex, discrete setting of ranking aggregation while preserving robustness guarantees against node corruption.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A gossip protocol lets network agents reach consensus on collective rankings using only local exchanges, with proven convergence and resilience to bad nodes.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Autonomous agents reach a global ranking consensus from distributed preferences using only local random gossip interactions, without any central coordinator.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"a660dbe22dfc508a9c808bdfa9013474a4610f5d58923e627d5853d69b1e65da"},"source":{"id":"2602.22847","kind":"arxiv","version":2},"verdict":{"id":"e3dce3a2-4c51-4fe8-8f7b-e801da5f07e5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T18:58:29.745360Z","strongest_claim":"The approach proposed and analyzed here relies on the robustness guarantees offered by random gossip communication, which allows autonomous agents to compute a global ranking consensus using local interactions only, without coordination or a central authority.","one_line_summary":"A gossip protocol lets network agents reach consensus on collective rankings using only local exchanges, with proven convergence and resilience to bad nodes.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That standard random gossip convergence properties for scalar statistics extend directly to the non-convex, discrete setting of ranking aggregation while preserving robustness guarantees against node corruption.","pith_extraction_headline":"Autonomous agents reach a global ranking consensus from distributed preferences using only local random gossip interactions, without any central coordinator."},"references":{"count":29,"sample":[{"doi":"","year":2023,"title":"Leonidas Akritidis, Miltiadis Alamaniotis, and Panayiotis Bozanis, ‘Flagr: A flexible high-performance library for rank aggregation’,Soft- wareX,21, (2023)","work_id":"c309202c-844a-4a7f-a0ac-934645b185ce","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1951,"title":"Kenneth J Arrow, ‘Social choice and individual values.’,Cowles Mono- graph No. 12, (1951)","work_id":"a2896952-4611-4e74-b975-72475ef4cb14","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Lyes Badis, Mourad Amad, Djamil A¨ıssani, and Sofiane Abbar, ‘P2pcf: A collaborative filtering based recommender system for peer to peer social networks’,Journal of High Speed Networks,27(1), 13–31, (","work_id":"22fa235d-ef98-4093-bb40-8a369b752ba1","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2010,"title":"Linas Baltrunas, Tadas Makcinskas, and Francesco Ricci, ‘Group rec- ommendations with rank aggregation and collaborative filtering’, inPro- ceedings of the 4th ACM conference on Recommender systems, (","work_id":"bca1a688-0c36-48d6-9260-d21351d3f5be","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Jean-Charles de Borda, ‘M´emoire sur les ´elections au scrutin’,Histoire de l’Acad´emie Royale des Sciences, (1781)","work_id":"a4463113-d04a-452a-b095-d60e021ee695","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":29,"snapshot_sha256":"66dc6ed6d30f6969580b457208c52dddd53e125806dd83b17610b36cf3b96443","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"6c57dac7da44ca9c6751c899a6bdb08efae6fe2ef0641add6d35e6a03ecc86bb"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}