{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:P4ZSUWND67TYO3S2KOG2PT3H2Q","short_pith_number":"pith:P4ZSUWND","schema_version":"1.0","canonical_sha256":"7f332a59a3f7e7876e5a538da7cf67d43884cf2827535c7106b681c86a991292","source":{"kind":"arxiv","id":"2505.15062","version":5},"attestation_state":"computed","paper":{"title":"SAKE: Structured Agentic Knowledge Extrapolation for Complex LLM Reasoning via Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Alejandro Ribeiro, Bowen Jiang, Dan Roth, Ignacio Houine, Jiashu He, Jinxuan Fan","submitted_at":"2025-05-21T03:30:55Z","abstract_excerpt":"Knowledge extrapolation is the process of inferring novel information by combining and extending existing knowledge that is explicitly available. It is essential for solving complex questions in specialized domains where retrieving comprehensive external knowledge is impractical. We propose SAKE (Structured Agentic Knowledge Extrapolation), a RL powered agentic framework that trains LLMs to autonomously retrieve and extrapolate structured knowledge through tool-augmented reinforcement learning. SAKE defines two exte nal KG tools: entity group construction and cross-group triplet retrieval. The"},"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":"2505.15062","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-21T03:30:55Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"49a765fa56e9f009aca07868b1dd7887061fdabce165c6344bd3c685e3ddb65c","abstract_canon_sha256":"77e71941c06afd8f27960aeaa5f6195280f3a48bbcc4aea2b24a39dd28a6374f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-11T01:19:57.559101Z","signature_b64":"SyZVkcgU+3x1EzHfFYkBNIgEmkH8D6Gb/bQMQWHnlQSMKVygXF4DcP5ooX7+MuQdopgy9xU0J8fOYtPZJIvNBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7f332a59a3f7e7876e5a538da7cf67d43884cf2827535c7106b681c86a991292","last_reissued_at":"2026-08-11T01:19:57.556503Z","signature_status":"signed_v1","first_computed_at":"2026-08-11T01:19:57.556503Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SAKE: Structured Agentic Knowledge Extrapolation for Complex LLM Reasoning via Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Alejandro Ribeiro, Bowen Jiang, Dan Roth, Ignacio Houine, Jiashu He, Jinxuan Fan","submitted_at":"2025-05-21T03:30:55Z","abstract_excerpt":"Knowledge extrapolation is the process of inferring novel information by combining and extending existing knowledge that is explicitly available. It is essential for solving complex questions in specialized domains where retrieving comprehensive external knowledge is impractical. We propose SAKE (Structured Agentic Knowledge Extrapolation), a RL powered agentic framework that trains LLMs to autonomously retrieve and extrapolate structured knowledge through tool-augmented reinforcement learning. SAKE defines two exte nal KG tools: entity group construction and cross-group triplet retrieval. The"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.15062","kind":"arxiv","version":5},"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/2505.15062/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":"2505.15062","created_at":"2026-08-11T01:19:57.557275+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.15062v5","created_at":"2026-08-11T01:19:57.557275+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.15062","created_at":"2026-08-11T01:19:57.557275+00:00"},{"alias_kind":"pith_short_12","alias_value":"P4ZSUWND67TY","created_at":"2026-08-11T01:19:57.557275+00:00"},{"alias_kind":"pith_short_16","alias_value":"P4ZSUWND67TYO3S2","created_at":"2026-08-11T01:19:57.557275+00:00"},{"alias_kind":"pith_short_8","alias_value":"P4ZSUWND","created_at":"2026-08-11T01:19:57.557275+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.17900","citing_title":"Leveraging Large Language Model for Intelligent Log Processing and Autonomous Debugging in Cloud AI Platforms","ref_index":24,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q","json":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q.json","graph_json":"https://pith.science/api/pith-number/P4ZSUWND67TYO3S2KOG2PT3H2Q/graph.json","events_json":"https://pith.science/api/pith-number/P4ZSUWND67TYO3S2KOG2PT3H2Q/events.json","paper":"https://pith.science/paper/P4ZSUWND"},"agent_actions":{"view_html":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q","download_json":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q.json","view_paper":"https://pith.science/paper/P4ZSUWND","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.15062&json=true","fetch_graph":"https://pith.science/api/pith-number/P4ZSUWND67TYO3S2KOG2PT3H2Q/graph.json","fetch_events":"https://pith.science/api/pith-number/P4ZSUWND67TYO3S2KOG2PT3H2Q/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q/action/storage_attestation","attest_author":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q/action/author_attestation","sign_citation":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q/action/citation_signature","submit_replication":"https://pith.science/pith/P4ZSUWND67TYO3S2KOG2PT3H2Q/action/replication_record"}},"created_at":"2026-08-11T01:19:57.557275+00:00","updated_at":"2026-08-11T01:19:57.557275+00:00"}