{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CI67QMUI33BKQ6IZIOZ2Y4VII3","short_pith_number":"pith:CI67QMUI","schema_version":"1.0","canonical_sha256":"123df83288dec2a8791943b3ac72a846c14ae4702b82015e750e10d01de093b5","source":{"kind":"arxiv","id":"2502.09963","version":1},"attestation_state":"computed","paper":{"title":"Generating on Generated: An Approach Towards Self-Evolving Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jiaxin Wu, Jinlin Wu, Qing Li, Xiaoyong Wei, Xulu Zhang, Zhaoxiang Zhang, Zhen Lei","submitted_at":"2025-02-14T07:41:47Z","abstract_excerpt":"Recursive Self-Improvement (RSI) enables intelligence systems to autonomously refine their capabilities. This paper explores the application of RSI in text-to-image diffusion models, addressing the challenge of training collapse caused by synthetic data. We identify two key factors contributing to this collapse: the lack of perceptual alignment and the accumulation of generative hallucinations. To mitigate these issues, we propose three strategies: (1) a prompt construction and filtering pipeline designed to facilitate the generation of perceptual aligned data, (2) a preference sampling method"},"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":"2502.09963","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-02-14T07:41:47Z","cross_cats_sorted":[],"title_canon_sha256":"20b7292776a31cfad0b7fc1a2f5c05952ed5fa6db6a07b3219a44d3b3f3387ad","abstract_canon_sha256":"8a798459ff503565278b6e81b774c19eb50d76614a8c1da237a72abd6f3bf8bb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:14:15.176009Z","signature_b64":"duNB/sqEGKVsd8Vho8UrwiCOtMl+hfcAX7vGgA4h6sHAEV3Ml0QowCLtqejJ+0CiDpzEkd2xs5eYn1MTcCYcAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"123df83288dec2a8791943b3ac72a846c14ae4702b82015e750e10d01de093b5","last_reissued_at":"2026-07-05T10:14:15.175550Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:14:15.175550Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generating on Generated: An Approach Towards Self-Evolving Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jiaxin Wu, Jinlin Wu, Qing Li, Xiaoyong Wei, Xulu Zhang, Zhaoxiang Zhang, Zhen Lei","submitted_at":"2025-02-14T07:41:47Z","abstract_excerpt":"Recursive Self-Improvement (RSI) enables intelligence systems to autonomously refine their capabilities. This paper explores the application of RSI in text-to-image diffusion models, addressing the challenge of training collapse caused by synthetic data. We identify two key factors contributing to this collapse: the lack of perceptual alignment and the accumulation of generative hallucinations. To mitigate these issues, we propose three strategies: (1) a prompt construction and filtering pipeline designed to facilitate the generation of perceptual aligned data, (2) a preference sampling method"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.09963","kind":"arxiv","version":1},"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/2502.09963/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":"2502.09963","created_at":"2026-07-05T10:14:15.175608+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.09963v1","created_at":"2026-07-05T10:14:15.175608+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.09963","created_at":"2026-07-05T10:14:15.175608+00:00"},{"alias_kind":"pith_short_12","alias_value":"CI67QMUI33BK","created_at":"2026-07-05T10:14:15.175608+00:00"},{"alias_kind":"pith_short_16","alias_value":"CI67QMUI33BKQ6IZ","created_at":"2026-07-05T10:14:15.175608+00:00"},{"alias_kind":"pith_short_8","alias_value":"CI67QMUI","created_at":"2026-07-05T10:14:15.175608+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07663","citing_title":"Recursive Self-Improvement in AI: From Bounded Self-Refinement to Autonomous Research Loops","ref_index":156,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3","json":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3.json","graph_json":"https://pith.science/api/pith-number/CI67QMUI33BKQ6IZIOZ2Y4VII3/graph.json","events_json":"https://pith.science/api/pith-number/CI67QMUI33BKQ6IZIOZ2Y4VII3/events.json","paper":"https://pith.science/paper/CI67QMUI"},"agent_actions":{"view_html":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3","download_json":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3.json","view_paper":"https://pith.science/paper/CI67QMUI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.09963&json=true","fetch_graph":"https://pith.science/api/pith-number/CI67QMUI33BKQ6IZIOZ2Y4VII3/graph.json","fetch_events":"https://pith.science/api/pith-number/CI67QMUI33BKQ6IZIOZ2Y4VII3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3/action/storage_attestation","attest_author":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3/action/author_attestation","sign_citation":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3/action/citation_signature","submit_replication":"https://pith.science/pith/CI67QMUI33BKQ6IZIOZ2Y4VII3/action/replication_record"}},"created_at":"2026-07-05T10:14:15.175608+00:00","updated_at":"2026-07-05T10:14:15.175608+00:00"}