{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Y7SBEASD34EDOHCTXHFDHACGMO","short_pith_number":"pith:Y7SBEASD","schema_version":"1.0","canonical_sha256":"c7e4120243df08371c53b9ca33804663a52fe78c4863e0dc08c08445c1c0f895","source":{"kind":"arxiv","id":"2504.09772","version":2},"attestation_state":"computed","paper":{"title":"Two Heads are Better Than One: Test-time Scaling of Multi-agent Collaborative Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Can Jin, Dimitris N. Metaxas, Hongwu Peng, Qixin Zhang, Tong Che, Yujin Tang","submitted_at":"2025-04-14T00:27:45Z","abstract_excerpt":"Multi-agent systems (MAS) built on large language models (LLMs) offer a promising path toward solving complex, real-world tasks that single-agent systems often struggle to manage. While recent advancements in test-time scaling (TTS) have significantly improved single-agent performance on challenging reasoning tasks, how to effectively scale collaboration and reasoning in MAS remains an open question. In this work, we introduce an adaptive multi-agent framework designed to enhance collaborative reasoning through both model-level training and system-level coordination. We construct M500, a high-"},"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":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2504.09772","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-04-14T00:27:45Z","cross_cats_sorted":[],"title_canon_sha256":"77dc0e942a589f3f3fbe7d50eb5b6580390f5fd64d6c2884e5daf693001fd086","abstract_canon_sha256":"c25a3e058b34273fdcf05f2ba08b8c1d9bb6dea6d7cd39643229104b27ad2a65"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:55:39.310755Z","signature_b64":"/A2vm249bxlT5mWANelLPBDwQqJSID73ws3N4sriRZDTb+EIO/nBIsLtgyGfcUMpWMt5xBjoUlWDJzcB2d23AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c7e4120243df08371c53b9ca33804663a52fe78c4863e0dc08c08445c1c0f895","last_reissued_at":"2026-07-05T11:55:39.310258Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:55:39.310258Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Two Heads are Better Than One: Test-time Scaling of Multi-agent Collaborative Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Can Jin, Dimitris N. Metaxas, Hongwu Peng, Qixin Zhang, Tong Che, Yujin Tang","submitted_at":"2025-04-14T00:27:45Z","abstract_excerpt":"Multi-agent systems (MAS) built on large language models (LLMs) offer a promising path toward solving complex, real-world tasks that single-agent systems often struggle to manage. While recent advancements in test-time scaling (TTS) have significantly improved single-agent performance on challenging reasoning tasks, how to effectively scale collaboration and reasoning in MAS remains an open question. In this work, we introduce an adaptive multi-agent framework designed to enhance collaborative reasoning through both model-level training and system-level coordination. We construct M500, a high-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.09772","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2504.09772/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"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":"2504.09772","created_at":"2026-07-05T11:55:39.310324+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.09772v2","created_at":"2026-07-05T11:55:39.310324+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.09772","created_at":"2026-07-05T11:55:39.310324+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y7SBEASD34ED","created_at":"2026-07-05T11:55:39.310324+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y7SBEASD34EDOHCT","created_at":"2026-07-05T11:55:39.310324+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y7SBEASD","created_at":"2026-07-05T11:55:39.310324+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2502.00709","citing_title":"RankFlow: A Multi-Role Collaborative Reranking Workflow Utilizing Large Language Models","ref_index":26,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO","json":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO.json","graph_json":"https://pith.science/api/pith-number/Y7SBEASD34EDOHCTXHFDHACGMO/graph.json","events_json":"https://pith.science/api/pith-number/Y7SBEASD34EDOHCTXHFDHACGMO/events.json","paper":"https://pith.science/paper/Y7SBEASD"},"agent_actions":{"view_html":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO","download_json":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO.json","view_paper":"https://pith.science/paper/Y7SBEASD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.09772&json=true","fetch_graph":"https://pith.science/api/pith-number/Y7SBEASD34EDOHCTXHFDHACGMO/graph.json","fetch_events":"https://pith.science/api/pith-number/Y7SBEASD34EDOHCTXHFDHACGMO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO/action/storage_attestation","attest_author":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO/action/author_attestation","sign_citation":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO/action/citation_signature","submit_replication":"https://pith.science/pith/Y7SBEASD34EDOHCTXHFDHACGMO/action/replication_record"}},"created_at":"2026-07-05T11:55:39.310324+00:00","updated_at":"2026-07-05T11:55:39.310324+00:00"}