{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:DURSVGYQFNESH3UGIEJ3NBYVEX","short_pith_number":"pith:DURSVGYQ","canonical_record":{"source":{"id":"2605.17003","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T14:01:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c2e93a4e65f78ce28ae390ea3dd8dff3c9f8d6403b9e3cd714e3addd54e1401b","abstract_canon_sha256":"1911ce215d4d6d854cfa5b80d1b0ff7a29068dbc8f9081e7df3daf2d84478c7f"},"schema_version":"1.0"},"canonical_sha256":"1d232a9b102b4923ee864113b6871525e22debc7d25b111a44127be0bfd54e28","source":{"kind":"arxiv","id":"2605.17003","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17003","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17003v1","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17003","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_12","alias_value":"DURSVGYQFNES","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_16","alias_value":"DURSVGYQFNESH3UG","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_8","alias_value":"DURSVGYQ","created_at":"2026-05-20T00:03:35Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:DURSVGYQFNESH3UGIEJ3NBYVEX","target":"record","payload":{"canonical_record":{"source":{"id":"2605.17003","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T14:01:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c2e93a4e65f78ce28ae390ea3dd8dff3c9f8d6403b9e3cd714e3addd54e1401b","abstract_canon_sha256":"1911ce215d4d6d854cfa5b80d1b0ff7a29068dbc8f9081e7df3daf2d84478c7f"},"schema_version":"1.0"},"canonical_sha256":"1d232a9b102b4923ee864113b6871525e22debc7d25b111a44127be0bfd54e28","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:35.505578Z","signature_b64":"x0MRt6telqCmWqK2Ybe+d2DDaKxL8Y3DAGCpAiXDL8oNhxwyXXAkZTlIAHq2OQltx55noi9CJv42h/nVu/5mCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1d232a9b102b4923ee864113b6871525e22debc7d25b111a44127be0bfd54e28","last_reissued_at":"2026-05-20T00:03:35.504868Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:35.504868Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.17003","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":"2026-05-20T00:03:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"msuAqDGiCVY0mH+M52e6Qp62GMcv9FhW8vWhE63J8zmrVkgsWmW/63Cpk/f6xWHg4SZq1SAU4BXxv7SXPK8VDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T20:39:28.475145Z"},"content_sha256":"e7a564b7593c9ba501c0d73eec049346c0c70fc81c2fd201b595e5b9af0890c5","schema_version":"1.0","event_id":"sha256:e7a564b7593c9ba501c0d73eec049346c0c70fc81c2fd201b595e5b9af0890c5"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:DURSVGYQFNESH3UGIEJ3NBYVEX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning-Zone Energy: Online Data Selection for Efficient RL Post-Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A closed-form Learning-Zone Energy score selects prompts aligned with large policy gradient updates for efficient LLM RL post-training.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Boyao Yang, Jun Zhu, Peng Cui","submitted_at":"2026-05-16T14:01:12Z","abstract_excerpt":"Reinforcement Learning (RL) post-training has emerged as the dominant paradigm for eliciting mathematical reasoning in Large Language Models (LLMs), yet prevailing techniques such as GRPO and DAPO distribute rollout and gradient budgets nearly uniformly across prompts, squandering compute on samples that are already mastered or remain far beyond the model's current capability. To address this fundamental inefficiency, we propose Learning-Zone Energy (LZE), a theoretically grounded, fully online data selection framework that concentrates computation on the model's active learning frontier. At i"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"A closed-form Learning-Zone Energy Score that fuses an initial-difficulty anchor, a normalized outcome-uncertainty term, and a pass-rate momentum into a single scalar that is provably aligned with the expected magnitude of group-relative policy gradient updates.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The forward pruner with replay maintains training stability and does not introduce harmful distributional shift or undetected forgetting, even while permanently skipping rollout generation for persistently solved prompts (as described in the framework section of the abstract).","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LZE is an online data selection method for RL post-training that fuses difficulty, uncertainty, and momentum signals into a closed-form score aligned with policy gradient magnitude, retaining 40% of data while matching or exceeding full-data baselines on math tasks with 36% fewer FLOPs.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A closed-form Learning-Zone Energy score selects prompts aligned with large policy gradient updates for efficient LLM RL post-training.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"c0731f8ec63412e930da64ef104ee09339858b880818936cb2a523b5d47cfec5"},"source":{"id":"2605.17003","kind":"arxiv","version":1},"verdict":{"id":"64523a30-6119-4388-a08e-5d8bfa4922fa","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T20:38:57.173146Z","strongest_claim":"A closed-form Learning-Zone Energy Score that fuses an initial-difficulty anchor, a normalized outcome-uncertainty term, and a pass-rate momentum into a single scalar that is provably aligned with the expected magnitude of group-relative policy gradient updates.","one_line_summary":"LZE is an online data selection method for RL post-training that fuses difficulty, uncertainty, and momentum signals into a closed-form score aligned with policy gradient magnitude, retaining 40% of data while matching or exceeding full-data baselines on math tasks with 36% fewer FLOPs.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The forward pruner with replay maintains training stability and does not introduce harmful distributional shift or undetected forgetting, even while permanently skipping rollout generation for persistently solved prompts (as described in the framework section of the abstract).","pith_extraction_headline":"A closed-form Learning-Zone Energy score selects prompts aligned with large policy gradient updates for efficient LLM RL post-training."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.17003/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T21:01:19.039573Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T20:50:49.261232Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"citation_quote_validity","ran_at":"2026-05-19T19:49:55.851379Z","status":"skipped","version":"0.1.0","findings_count":0},{"name":"cited_work_retraction","ran_at":"2026-05-19T19:23:34.822550Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.197245Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.287187Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"78dcfbb8d1965c2042f079a133f96369b2fe006a3f8c804872e62ee59c25a146"},"references":{"count":50,"sample":[{"doi":"","year":2023,"title":"SemDeDup: Data-efficient learning at web-scale through semantic deduplication","work_id":"492d4320-a8d4-4094-b226-ea8d784560d9","ref_index":1,"cited_arxiv_id":"2303.09540","is_internal_anchor":true},{"doi":"10.1038/s41586-025-094","year":2025,"title":"Deepseek-r1 incentivizes reasoning in llms through reinforcement learning","work_id":"f9b882e7-7f02-44e0-b1a6-5b0e5e2e1b18","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.18653/v1/2024","year":2024,"title":"NAACL-LONG.102","work_id":"557a574b-31bd-4f07-9bd0-47bc7c74bcd8","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1145/1553374.1553380","year":2009,"title":"In Proceedings of the 26th Annual International Conference on Machine Learning (Montreal, Quebec, Canada) (ICML ’09)","work_id":"5d30e737-af53-4f1d-8f90-c8e0159a22ba","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1901,"title":"Language models are few-shot learners","work_id":"230b7b6b-51e6-4490-b1b0-890f3b0bc1f3","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":50,"snapshot_sha256":"97da7ddbf12a80f45e1d245cb0451026cea3ffe68f49466402c6a7f11c829791","internal_anchors":1},"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"},"verdict_id":"64523a30-6119-4388-a08e-5d8bfa4922fa"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:03:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uxPJr+JhjEG+lYSaju6VHNqwYr3hDobp1AbMb4yzbu1XrRxMGgz+lN11F/F64pesR//T0wvioIGCo8OHYpkpDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T20:39:28.475982Z"},"content_sha256":"7f787c4d136312cc9212e10b9abd5226a8dd078bd67359c24621a8cfee8bff44","schema_version":"1.0","event_id":"sha256:7f787c4d136312cc9212e10b9abd5226a8dd078bd67359c24621a8cfee8bff44"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DURSVGYQFNESH3UGIEJ3NBYVEX/bundle.json","state_url":"https://pith.science/pith/DURSVGYQFNESH3UGIEJ3NBYVEX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DURSVGYQFNESH3UGIEJ3NBYVEX/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":"2026-05-26T20:39:28Z","links":{"resolver":"https://pith.science/pith/DURSVGYQFNESH3UGIEJ3NBYVEX","bundle":"https://pith.science/pith/DURSVGYQFNESH3UGIEJ3NBYVEX/bundle.json","state":"https://pith.science/pith/DURSVGYQFNESH3UGIEJ3NBYVEX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DURSVGYQFNESH3UGIEJ3NBYVEX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DURSVGYQFNESH3UGIEJ3NBYVEX","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":"1911ce215d4d6d854cfa5b80d1b0ff7a29068dbc8f9081e7df3daf2d84478c7f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T14:01:12Z","title_canon_sha256":"c2e93a4e65f78ce28ae390ea3dd8dff3c9f8d6403b9e3cd714e3addd54e1401b"},"schema_version":"1.0","source":{"id":"2605.17003","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17003","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17003v1","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17003","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_12","alias_value":"DURSVGYQFNES","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_16","alias_value":"DURSVGYQFNESH3UG","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_8","alias_value":"DURSVGYQ","created_at":"2026-05-20T00:03:35Z"}],"graph_snapshots":[{"event_id":"sha256:7f787c4d136312cc9212e10b9abd5226a8dd078bd67359c24621a8cfee8bff44","target":"graph","created_at":"2026-05-20T00:03:35Z","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":"A closed-form Learning-Zone Energy Score that fuses an initial-difficulty anchor, a normalized outcome-uncertainty term, and a pass-rate momentum into a single scalar that is provably aligned with the expected magnitude of group-relative policy gradient updates."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The forward pruner with replay maintains training stability and does not introduce harmful distributional shift or undetected forgetting, even while permanently skipping rollout generation for persistently solved prompts (as described in the framework section of the abstract)."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"LZE is an online data selection method for RL post-training that fuses difficulty, uncertainty, and momentum signals into a closed-form score aligned with policy gradient magnitude, retaining 40% of data while matching or exceeding full-data baselines on math tasks with 36% fewer FLOPs."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A closed-form Learning-Zone Energy score selects prompts aligned with large policy gradient updates for efficient LLM RL post-training."}],"snapshot_sha256":"c0731f8ec63412e930da64ef104ee09339858b880818936cb2a523b5d47cfec5"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T21:01:19.039573Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T20:50:49.261232Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"citation_quote_validity","ran_at":"2026-05-19T19:49:55.851379Z","status":"skipped","version":"0.1.0"},{"findings_count":0,"name":"cited_work_retraction","ran_at":"2026-05-19T19:23:34.822550Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.197245Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.287187Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.17003/integrity.json","findings":[],"snapshot_sha256":"78dcfbb8d1965c2042f079a133f96369b2fe006a3f8c804872e62ee59c25a146","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement Learning (RL) post-training has emerged as the dominant paradigm for eliciting mathematical reasoning in Large Language Models (LLMs), yet prevailing techniques such as GRPO and DAPO distribute rollout and gradient budgets nearly uniformly across prompts, squandering compute on samples that are already mastered or remain far beyond the model's current capability. To address this fundamental inefficiency, we propose Learning-Zone Energy (LZE), a theoretically grounded, fully online data selection framework that concentrates computation on the model's active learning frontier. At i","authors_text":"Boyao Yang, Jun Zhu, Peng Cui","cross_cats":["cs.AI"],"headline":"A closed-form Learning-Zone Energy score selects prompts aligned with large policy gradient updates for efficient LLM RL post-training.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T14:01:12Z","title":"Learning-Zone Energy: Online Data Selection for Efficient RL Post-Training"},"references":{"count":50,"internal_anchors":1,"resolved_work":50,"sample":[{"cited_arxiv_id":"2303.09540","doi":"","is_internal_anchor":true,"ref_index":1,"title":"SemDeDup: Data-efficient learning at web-scale through semantic deduplication","work_id":"492d4320-a8d4-4094-b226-ea8d784560d9","year":2023},{"cited_arxiv_id":"","doi":"10.1038/s41586-025-094","is_internal_anchor":false,"ref_index":2,"title":"Deepseek-r1 incentivizes reasoning in llms through reinforcement learning","work_id":"f9b882e7-7f02-44e0-b1a6-5b0e5e2e1b18","year":2025},{"cited_arxiv_id":"","doi":"10.18653/v1/2024","is_internal_anchor":false,"ref_index":3,"title":"NAACL-LONG.102","work_id":"557a574b-31bd-4f07-9bd0-47bc7c74bcd8","year":2024},{"cited_arxiv_id":"","doi":"10.1145/1553374.1553380","is_internal_anchor":false,"ref_index":4,"title":"In Proceedings of the 26th Annual International Conference on Machine Learning (Montreal, Quebec, Canada) (ICML ’09)","work_id":"5d30e737-af53-4f1d-8f90-c8e0159a22ba","year":2009},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Language models are few-shot learners","work_id":"230b7b6b-51e6-4490-b1b0-890f3b0bc1f3","year":1901}],"snapshot_sha256":"97da7ddbf12a80f45e1d245cb0451026cea3ffe68f49466402c6a7f11c829791"},"source":{"id":"2605.17003","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T20:38:57.173146Z","id":"64523a30-6119-4388-a08e-5d8bfa4922fa","model_set":{"reader":"grok-4.3"},"one_line_summary":"LZE is an online data selection method for RL post-training that fuses difficulty, uncertainty, and momentum signals into a closed-form score aligned with policy gradient magnitude, retaining 40% of data while matching or exceeding full-data baselines on math tasks with 36% fewer FLOPs.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A closed-form Learning-Zone Energy score selects prompts aligned with large policy gradient updates for efficient LLM RL post-training.","strongest_claim":"A closed-form Learning-Zone Energy Score that fuses an initial-difficulty anchor, a normalized outcome-uncertainty term, and a pass-rate momentum into a single scalar that is provably aligned with the expected magnitude of group-relative policy gradient updates.","weakest_assumption":"The forward pruner with replay maintains training stability and does not introduce harmful distributional shift or undetected forgetting, even while permanently skipping rollout generation for persistently solved prompts (as described in the framework section of the abstract)."}},"verdict_id":"64523a30-6119-4388-a08e-5d8bfa4922fa"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:e7a564b7593c9ba501c0d73eec049346c0c70fc81c2fd201b595e5b9af0890c5","target":"record","created_at":"2026-05-20T00:03:35Z","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":"1911ce215d4d6d854cfa5b80d1b0ff7a29068dbc8f9081e7df3daf2d84478c7f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T14:01:12Z","title_canon_sha256":"c2e93a4e65f78ce28ae390ea3dd8dff3c9f8d6403b9e3cd714e3addd54e1401b"},"schema_version":"1.0","source":{"id":"2605.17003","kind":"arxiv","version":1}},"canonical_sha256":"1d232a9b102b4923ee864113b6871525e22debc7d25b111a44127be0bfd54e28","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1d232a9b102b4923ee864113b6871525e22debc7d25b111a44127be0bfd54e28","first_computed_at":"2026-05-20T00:03:35.504868Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:35.504868Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"x0MRt6telqCmWqK2Ybe+d2DDaKxL8Y3DAGCpAiXDL8oNhxwyXXAkZTlIAHq2OQltx55noi9CJv42h/nVu/5mCw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:35.505578Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.17003","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e7a564b7593c9ba501c0d73eec049346c0c70fc81c2fd201b595e5b9af0890c5","sha256:7f787c4d136312cc9212e10b9abd5226a8dd078bd67359c24621a8cfee8bff44"],"state_sha256":"188c2570dff69e287acd1315678edef040193f3ba0a741ee69ed3146d2941773"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uuONrv/EP8gdDowEknNXIRH2vjcZ8GP51ocYE39xtMFRKpmlWZCxjG7/qlI4VdJI6vPCMG2lmFLXG8/h/CIGAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-26T20:39:28.479400Z","bundle_sha256":"d8ef4c22cabad3c9090d8c2674b5ae976d97f2cf4a9fcfd50fd86b4d20218542"}}