{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QX4EMMWD57QKEA3ZLJ5ACDRG3V","short_pith_number":"pith:QX4EMMWD","schema_version":"1.0","canonical_sha256":"85f84632c3efe0a203795a7a010e26dd66135c2a3229c6fafeedb305115d3b5f","source":{"kind":"arxiv","id":"2605.13145","version":1},"attestation_state":"computed","paper":{"title":"Collaborating in Multi-Armed Bandits with Strategic Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"CAOS mechanism sustains collaboration as a Nash equilibrium in strategic multi-agent bandits via information sharing rules.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Idan Barnea, Ofir Schlisselberg, Yishay Mansour","submitted_at":"2026-05-13T08:10:36Z","abstract_excerpt":"We study collaborative learning in multi-agent Bayesian bandit problems, where strategic agents collectively solve the same bandit instance. While multiple agents can accelerate learning by sharing information, strategic agents might prefer to free-ride and avoid exploration. We consider a setting with persistent agents that participate in multiple time periods. This is in contrast to most previous works on incentives in multi-agent MAB, which assume short-lived agents, namely each agent has a single decision to make and optimizes their expected reward in that single decision. As in the multi-"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.13145","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T08:10:36Z","cross_cats_sorted":[],"title_canon_sha256":"66aea708a1a3d2139e2f755160c248da0d44433ba44e897b2bdcc478d308a435","abstract_canon_sha256":"8a811d4a5c6cec83e31e2fec7dd496b58db5cf34240362f4e11fc1b85a28f0a2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:08:57.317446Z","signature_b64":"rpRoTepCuAd9vLnKRWZf6gqeNRDhMd/W2uFky8zV4HZYE5M3eOrlY1Yc3ij7D30jArsQhIRB+NPwuDbX80k8Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"85f84632c3efe0a203795a7a010e26dd66135c2a3229c6fafeedb305115d3b5f","last_reissued_at":"2026-05-18T03:08:57.316635Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:08:57.316635Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Collaborating in Multi-Armed Bandits with Strategic Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"CAOS mechanism sustains collaboration as a Nash equilibrium in strategic multi-agent bandits via information sharing rules.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Idan Barnea, Ofir Schlisselberg, Yishay Mansour","submitted_at":"2026-05-13T08:10:36Z","abstract_excerpt":"We study collaborative learning in multi-agent Bayesian bandit problems, where strategic agents collectively solve the same bandit instance. While multiple agents can accelerate learning by sharing information, strategic agents might prefer to free-ride and avoid exploration. We consider a setting with persistent agents that participate in multiple time periods. This is in contrast to most previous works on incentives in multi-agent MAB, which assume short-lived agents, namely each agent has a single decision to make and optimizes their expected reward in that single decision. As in the multi-"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We propose CAOS, a mechanism that sustains collaboration as a Nash equilibrium while achieving strong regret guarantees.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The setting assumes persistent agents that participate across multiple time periods and that the only incentives come from information sharing rules, with agents playing according to Nash equilibrium.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"CAOS sustains collaboration as a Nash equilibrium among persistent strategic agents in Bayesian multi-armed bandits via information sharing, with strong regret guarantees.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"CAOS mechanism sustains collaboration as a Nash equilibrium in strategic multi-agent bandits via information sharing rules.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"8ab49aac42b27efe0e337864ffc6dfac47ad4289a78cd1b45941291acd8a95db"},"source":{"id":"2605.13145","kind":"arxiv","version":1},"verdict":{"id":"9431827e-d239-4418-bf67-7ba6f59a3a84","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:19:18.715703Z","strongest_claim":"We propose CAOS, a mechanism that sustains collaboration as a Nash equilibrium while achieving strong regret guarantees.","one_line_summary":"CAOS sustains collaboration as a Nash equilibrium among persistent strategic agents in Bayesian multi-armed bandits via information sharing, with strong regret guarantees.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The setting assumes persistent agents that participate across multiple time periods and that the only incentives come from information sharing rules, with agents playing according to Nash equilibrium.","pith_extraction_headline":"CAOS mechanism sustains collaboration as a Nash equilibrium in strategic multi-agent bandits via information sharing rules."},"references":{"count":32,"sample":[{"doi":"","year":2002,"title":"P. Auer, N. Cesa-Bianchi, Y. Freund, and R. E. Schapire. The nonstochastic multiarmed bandit problem.SIAM J. Comput., 2002","work_id":"2f51b82e-a01c-4b43-8c95-0161f6fd2752","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"K. Banihashem, N. Collina, and A. Slivkins. Bandit social learning with exploration episodes,","work_id":"69b04cb7-f640-4d16-92fb-0f51ec75307e","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"URLhttps://arxiv.org/abs/2602.05835","work_id":"72430df7-34c3-40b2-83d6-90464d7a1a59","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"Y. Bar-On and Y. Mansour. Individual regret in cooperative nonstochastic multi-armed bandits. Advances in Neural Information Processing Systems, 2019","work_id":"8c5d7494-ff87-4913-ae2b-8fb028e761d4","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"The Horizon Threshold in Cooperative Multi-Agent Reward-Free Exploration","work_id":"a92c1151-5181-4598-ad28-95b438219353","ref_index":5,"cited_arxiv_id":"2602.01453","is_internal_anchor":true}],"resolved_work":32,"snapshot_sha256":"692ab3246f3ecf69ea7d371b00633818e949cb250746b5aca758b4ef0a40270d","internal_anchors":2},"formal_canon":{"evidence_count":2,"snapshot_sha256":"f9d56f2a2cc4daa7cacaf5fed3d82dc3471bbf7de049891f5aca790d22892744"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.13145","created_at":"2026-05-18T03:08:57.316768+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.13145v1","created_at":"2026-05-18T03:08:57.316768+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13145","created_at":"2026-05-18T03:08:57.316768+00:00"},{"alias_kind":"pith_short_12","alias_value":"QX4EMMWD57QK","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"QX4EMMWD57QKEA3Z","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"QX4EMMWD","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V","json":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V.json","graph_json":"https://pith.science/api/pith-number/QX4EMMWD57QKEA3ZLJ5ACDRG3V/graph.json","events_json":"https://pith.science/api/pith-number/QX4EMMWD57QKEA3ZLJ5ACDRG3V/events.json","paper":"https://pith.science/paper/QX4EMMWD"},"agent_actions":{"view_html":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V","download_json":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V.json","view_paper":"https://pith.science/paper/QX4EMMWD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.13145&json=true","fetch_graph":"https://pith.science/api/pith-number/QX4EMMWD57QKEA3ZLJ5ACDRG3V/graph.json","fetch_events":"https://pith.science/api/pith-number/QX4EMMWD57QKEA3ZLJ5ACDRG3V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V/action/storage_attestation","attest_author":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V/action/author_attestation","sign_citation":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V/action/citation_signature","submit_replication":"https://pith.science/pith/QX4EMMWD57QKEA3ZLJ5ACDRG3V/action/replication_record"}},"created_at":"2026-05-18T03:08:57.316768+00:00","updated_at":"2026-05-18T03:08:57.316768+00:00"}