{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:UO7YVABAGTPH7QMDREGJAU4C4G","short_pith_number":"pith:UO7YVABA","canonical_record":{"source":{"id":"2605.15905","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-05-15T12:42:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"096ce218679d892b9946064eadc53fa1b36c70b4d128ea7ce0491a96a2d2219f","abstract_canon_sha256":"75abb2fa56e8777109f9d014175ae7b44cdb8dfa017c39ee58eff302a3b4e0e5"},"schema_version":"1.0"},"canonical_sha256":"a3bf8a802034de7fc183890c905382e181484e42937c725251860ec30397b01a","source":{"kind":"arxiv","id":"2605.15905","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15905","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15905v1","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15905","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"pith_short_12","alias_value":"UO7YVABAGTPH","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"pith_short_16","alias_value":"UO7YVABAGTPH7QMD","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"pith_short_8","alias_value":"UO7YVABA","created_at":"2026-05-20T00:01:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:UO7YVABAGTPH7QMDREGJAU4C4G","target":"record","payload":{"canonical_record":{"source":{"id":"2605.15905","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-05-15T12:42:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"096ce218679d892b9946064eadc53fa1b36c70b4d128ea7ce0491a96a2d2219f","abstract_canon_sha256":"75abb2fa56e8777109f9d014175ae7b44cdb8dfa017c39ee58eff302a3b4e0e5"},"schema_version":"1.0"},"canonical_sha256":"a3bf8a802034de7fc183890c905382e181484e42937c725251860ec30397b01a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:24.792019Z","signature_b64":"bolm7nAEXnDR7D/T4fR86LHIMoX/NCUipP8hxvTY+wna92CChrG+I7TAqvvnkZJN83MK6Wv4flEmFd0X2m2VBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a3bf8a802034de7fc183890c905382e181484e42937c725251860ec30397b01a","last_reissued_at":"2026-05-20T00:01:24.791252Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:24.791252Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.15905","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:01:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"gH3pjdfPyY1PFI+XKzePn7Tmh/MdjOokk94LbfAJMMEQjCOT4Q2y6yJImD/qE/kQAnghjd5KkpsZG9iB2lvYDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-24T17:33:02.625922Z"},"content_sha256":"a2378093edafddcdd0ea5a406fb8dc9c1d22a03901b1fd1a2c14ec300664a1d8","schema_version":"1.0","event_id":"sha256:a2378093edafddcdd0ea5a406fb8dc9c1d22a03901b1fd1a2c14ec300664a1d8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:UO7YVABAGTPH7QMDREGJAU4C4G","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Generative Long-term User Interest Modeling for Click-Through Rate Prediction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A generative module produces multiple target-independent interest distributions to capture diverse long-term user behaviors for click-through rate prediction.","cross_cats":["cs.AI"],"primary_cat":"cs.IR","authors_text":"Bo Zhang, Hao Fang, Huimu Ye, Jiangli Shao, Kaifu Zheng, Shu Han, Xingxing Wang, Zhiwei Liu","submitted_at":"2026-05-15T12:42:54Z","abstract_excerpt":"Modeling long-term user interests with massive historical user behaviors enhances click-through rate (CTR) prediction performance in advertising and recommendation systems. Typically, a two-stage framework is widely adopted, where a general search unit (GSU) first retrieves top-$k$ relevant behaviors towards the target item, and an exact search unit (ESU) generates interest features via tailored attention. However, current target-centered GSU would ignore other latent user interests, leading to incomplete and biased interest features. Additionally, the matching-based retrieval process in GSUs "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Based on the generation process, GenLI improves the diversity of user interests and avoids complex matching-based behavioral retrieval, achieving a better balance between accuracy and efficiency for CTR prediction.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The interest generation module produces distributions that faithfully represent multiple latent aspects of real user interests even when generated without any information about the target item, and the simple lookup operation in the behavior retrieval module selects behaviors that remain sufficiently relevant for downstream interest fusion.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"GenLI generates diverse target-independent interest distributions via an IGM, retrieves behaviors with O(1) lookup in BRM, and fuses via IFM gating to balance accuracy and efficiency in CTR prediction.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A generative module produces multiple target-independent interest distributions to capture diverse long-term user behaviors for click-through rate prediction.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"12cd562837d8d82a6da849fc10d4900c822dae4fd0651b874dbff06b00a7c17e"},"source":{"id":"2605.15905","kind":"arxiv","version":1},"verdict":{"id":"74e2230b-405c-4b74-8ec5-e6227644a1df","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T22:16:37.944334Z","strongest_claim":"Based on the generation process, GenLI improves the diversity of user interests and avoids complex matching-based behavioral retrieval, achieving a better balance between accuracy and efficiency for CTR prediction.","one_line_summary":"GenLI generates diverse target-independent interest distributions via an IGM, retrieves behaviors with O(1) lookup in BRM, and fuses via IFM gating to balance accuracy and efficiency in CTR prediction.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The interest generation module produces distributions that faithfully represent multiple latent aspects of real user interests even when generated without any information about the target item, and the simple lookup operation in the behavior retrieval module selects behaviors that remain sufficiently relevant for downstream interest fusion.","pith_extraction_headline":"A generative module produces multiple target-independent interest distributions to capture diverse long-term user behaviors for click-through rate prediction."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15905/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T22:31:19.642418Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T22:21:50.250465Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:46.561950Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T17:01:55.768308Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"8ee1c5ca738e42f12545d15a2504379104f07889f1b970ce65b23fa49745bbba"},"references":{"count":43,"sample":[{"doi":"","year":2022,"title":"Yue Cao, Xiaojiang Zhou, Jiaqi Feng, Peihao Huang, Yao Xiao, Dayao Chen, and Sheng Chen. 2022. Sampling Is All You Need on Modeling Long-Term User Behaviors for CTR Prediction. InProceedings of the 31","work_id":"d6e08fd4-02fb-4ece-a4e6-807dbeb82234","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Jianxin Chang, Chenbin Zhang, Zhiyi Fu, Xiaoxue Zang, Lin Guan, Jing Lu, Yiqun Hui, Dewei Leng, Yanan Niu, Yang Song, and Kun Gai. 2023. TWIN: TWo-stage Interest Network for Lifelong User Behavior Mod","work_id":"0d204978-f60e-4b73-be87-e66a462f23bf","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"Junxuan Chen, Baigui Sun, Hao Li, Hongtao Lu, and Xian-Sheng Hua. 2016. Deep CTR Prediction in Display Advertising. InProceedings of the 24th ACM International Conference on Multimedia(Amsterdam, The ","work_id":"f70ed98a-f905-4015-a004-548ddb12bf32","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Qiwei Chen, Changhua Pei, Shanshan Lv, Chao Li, Junfeng Ge, and Wenwu Ou","work_id":"ab15ab62-5b73-44e7-94a8-efec83833a8a","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"End-to-end user behavior retrieval in click-through rateprediction model","work_id":"01c79691-2a73-4ce2-9d21-c68702877449","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":43,"snapshot_sha256":"903eb963ff2ba7e6c01d4a224ee7701daafb2c37e33bfe24f9979d87f6bc47d3","internal_anchors":3},"formal_canon":{"evidence_count":2,"snapshot_sha256":"9eab1a71ebb3fc19e0602d580a3e7c7386b75b7180d70c768f1fdb014bd7e3b6"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"74e2230b-405c-4b74-8ec5-e6227644a1df"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:01:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uWh3q+A4fPnXP0zCcLBOQlfJuOdjDpva5ZU4cCGTkee3C79cL+DNMeDg+/s42XSrQFKnu4DLzU8mNKzQwQdbDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-24T17:33:02.627163Z"},"content_sha256":"15da0052180e8afc5d36f643bfee9b9f93500014f7f7b7ee93c3bd80069c3046","schema_version":"1.0","event_id":"sha256:15da0052180e8afc5d36f643bfee9b9f93500014f7f7b7ee93c3bd80069c3046"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/UO7YVABAGTPH7QMDREGJAU4C4G/bundle.json","state_url":"https://pith.science/pith/UO7YVABAGTPH7QMDREGJAU4C4G/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/UO7YVABAGTPH7QMDREGJAU4C4G/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-24T17:33:02Z","links":{"resolver":"https://pith.science/pith/UO7YVABAGTPH7QMDREGJAU4C4G","bundle":"https://pith.science/pith/UO7YVABAGTPH7QMDREGJAU4C4G/bundle.json","state":"https://pith.science/pith/UO7YVABAGTPH7QMDREGJAU4C4G/state.json","well_known_bundle":"https://pith.science/.well-known/pith/UO7YVABAGTPH7QMDREGJAU4C4G/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:UO7YVABAGTPH7QMDREGJAU4C4G","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"75abb2fa56e8777109f9d014175ae7b44cdb8dfa017c39ee58eff302a3b4e0e5","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-05-15T12:42:54Z","title_canon_sha256":"096ce218679d892b9946064eadc53fa1b36c70b4d128ea7ce0491a96a2d2219f"},"schema_version":"1.0","source":{"id":"2605.15905","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15905","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15905v1","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15905","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"pith_short_12","alias_value":"UO7YVABAGTPH","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"pith_short_16","alias_value":"UO7YVABAGTPH7QMD","created_at":"2026-05-20T00:01:24Z"},{"alias_kind":"pith_short_8","alias_value":"UO7YVABA","created_at":"2026-05-20T00:01:24Z"}],"graph_snapshots":[{"event_id":"sha256:15da0052180e8afc5d36f643bfee9b9f93500014f7f7b7ee93c3bd80069c3046","target":"graph","created_at":"2026-05-20T00:01:24Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Based on the generation process, GenLI improves the diversity of user interests and avoids complex matching-based behavioral retrieval, achieving a better balance between accuracy and efficiency for CTR prediction."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The interest generation module produces distributions that faithfully represent multiple latent aspects of real user interests even when generated without any information about the target item, and the simple lookup operation in the behavior retrieval module selects behaviors that remain sufficiently relevant for downstream interest fusion."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"GenLI generates diverse target-independent interest distributions via an IGM, retrieves behaviors with O(1) lookup in BRM, and fuses via IFM gating to balance accuracy and efficiency in CTR prediction."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A generative module produces multiple target-independent interest distributions to capture diverse long-term user behaviors for click-through rate prediction."}],"snapshot_sha256":"12cd562837d8d82a6da849fc10d4900c822dae4fd0651b874dbff06b00a7c17e"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"9eab1a71ebb3fc19e0602d580a3e7c7386b75b7180d70c768f1fdb014bd7e3b6"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T22:31:19.642418Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T22:21:50.250465Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:46.561950Z","status":"skipped","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T17:01:55.768308Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.15905/integrity.json","findings":[],"snapshot_sha256":"8ee1c5ca738e42f12545d15a2504379104f07889f1b970ce65b23fa49745bbba","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Modeling long-term user interests with massive historical user behaviors enhances click-through rate (CTR) prediction performance in advertising and recommendation systems. Typically, a two-stage framework is widely adopted, where a general search unit (GSU) first retrieves top-$k$ relevant behaviors towards the target item, and an exact search unit (ESU) generates interest features via tailored attention. However, current target-centered GSU would ignore other latent user interests, leading to incomplete and biased interest features. Additionally, the matching-based retrieval process in GSUs ","authors_text":"Bo Zhang, Hao Fang, Huimu Ye, Jiangli Shao, Kaifu Zheng, Shu Han, Xingxing Wang, Zhiwei Liu","cross_cats":["cs.AI"],"headline":"A generative module produces multiple target-independent interest distributions to capture diverse long-term user behaviors for click-through rate prediction.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-05-15T12:42:54Z","title":"Generative Long-term User Interest Modeling for Click-Through Rate Prediction"},"references":{"count":43,"internal_anchors":3,"resolved_work":43,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Yue Cao, Xiaojiang Zhou, Jiaqi Feng, Peihao Huang, Yao Xiao, Dayao Chen, and Sheng Chen. 2022. Sampling Is All You Need on Modeling Long-Term User Behaviors for CTR Prediction. InProceedings of the 31","work_id":"d6e08fd4-02fb-4ece-a4e6-807dbeb82234","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Jianxin Chang, Chenbin Zhang, Zhiyi Fu, Xiaoxue Zang, Lin Guan, Jing Lu, Yiqun Hui, Dewei Leng, Yanan Niu, Yang Song, and Kun Gai. 2023. TWIN: TWo-stage Interest Network for Lifelong User Behavior Mod","work_id":"0d204978-f60e-4b73-be87-e66a462f23bf","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Junxuan Chen, Baigui Sun, Hao Li, Hongtao Lu, and Xian-Sheng Hua. 2016. Deep CTR Prediction in Display Advertising. InProceedings of the 24th ACM International Conference on Multimedia(Amsterdam, The ","work_id":"f70ed98a-f905-4015-a004-548ddb12bf32","year":2016},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Qiwei Chen, Changhua Pei, Shanshan Lv, Chao Li, Junfeng Ge, and Wenwu Ou","work_id":"ab15ab62-5b73-44e7-94a8-efec83833a8a","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"End-to-end user behavior retrieval in click-through rateprediction model","work_id":"01c79691-2a73-4ce2-9d21-c68702877449","year":2021}],"snapshot_sha256":"903eb963ff2ba7e6c01d4a224ee7701daafb2c37e33bfe24f9979d87f6bc47d3"},"source":{"id":"2605.15905","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T22:16:37.944334Z","id":"74e2230b-405c-4b74-8ec5-e6227644a1df","model_set":{"reader":"grok-4.3"},"one_line_summary":"GenLI generates diverse target-independent interest distributions via an IGM, retrieves behaviors with O(1) lookup in BRM, and fuses via IFM gating to balance accuracy and efficiency in CTR prediction.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A generative module produces multiple target-independent interest distributions to capture diverse long-term user behaviors for click-through rate prediction.","strongest_claim":"Based on the generation process, GenLI improves the diversity of user interests and avoids complex matching-based behavioral retrieval, achieving a better balance between accuracy and efficiency for CTR prediction.","weakest_assumption":"The interest generation module produces distributions that faithfully represent multiple latent aspects of real user interests even when generated without any information about the target item, and the simple lookup operation in the behavior retrieval module selects behaviors that remain sufficiently relevant for downstream interest fusion."}},"verdict_id":"74e2230b-405c-4b74-8ec5-e6227644a1df"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:a2378093edafddcdd0ea5a406fb8dc9c1d22a03901b1fd1a2c14ec300664a1d8","target":"record","created_at":"2026-05-20T00:01:24Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"75abb2fa56e8777109f9d014175ae7b44cdb8dfa017c39ee58eff302a3b4e0e5","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-05-15T12:42:54Z","title_canon_sha256":"096ce218679d892b9946064eadc53fa1b36c70b4d128ea7ce0491a96a2d2219f"},"schema_version":"1.0","source":{"id":"2605.15905","kind":"arxiv","version":1}},"canonical_sha256":"a3bf8a802034de7fc183890c905382e181484e42937c725251860ec30397b01a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a3bf8a802034de7fc183890c905382e181484e42937c725251860ec30397b01a","first_computed_at":"2026-05-20T00:01:24.791252Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:01:24.791252Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"bolm7nAEXnDR7D/T4fR86LHIMoX/NCUipP8hxvTY+wna92CChrG+I7TAqvvnkZJN83MK6Wv4flEmFd0X2m2VBw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:01:24.792019Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.15905","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a2378093edafddcdd0ea5a406fb8dc9c1d22a03901b1fd1a2c14ec300664a1d8","sha256:15da0052180e8afc5d36f643bfee9b9f93500014f7f7b7ee93c3bd80069c3046"],"state_sha256":"218c695a2d5f0fef3c11605f85aca4cf5a6a5fbcf5a44a9f113f0f9bdc955a3f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3K3/5y4fEwi4XvS25zMAqwPZrrFuDAzNtR6o8x9fVsFtWwwTxXWDYinI/3a9GmhnZdKDjpG6gj95RMYUuDArDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-24T17:33:02.632605Z","bundle_sha256":"6592ed67d346205116634246e532dbf10efae1ab44992019c9420496d5c893b5"}}