{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5DLWYQR6JAKYUTIBCA7BDPTXUX","short_pith_number":"pith:5DLWYQR6","schema_version":"1.0","canonical_sha256":"e8d76c423e48158a4d01103e11be77a5e8d7b299690affc9b0b85022fdf8a021","source":{"kind":"arxiv","id":"2409.03025","version":2},"attestation_state":"computed","paper":{"title":"No Detail Left Behind: Revisiting Self-Retrieval for Fine-Grained Image Captioning","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Darshan Singh, Makarand Tapaswi, Manu Gaur","submitted_at":"2024-09-04T18:32:39Z","abstract_excerpt":"Image captioning systems are unable to generate fine-grained captions as they are trained on data that is either noisy (alt-text) or generic (human annotations). This is further exacerbated by maximum likelihood training that encourages generation of frequently occurring phrases. Previous works have tried to address this limitation by fine-tuning captioners with a self-retrieval (SR) reward. However, we find that SR fine-tuning has a tendency to reduce caption faithfulness and even hallucinate. In this work, we circumvent this bottleneck by improving the MLE initialization of the captioning sy"},"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":"2409.03025","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2024-09-04T18:32:39Z","cross_cats_sorted":[],"title_canon_sha256":"ef7ce293636f37c9dc76515260f848d214b4a099a3b572f5a83b3f8473d3a343","abstract_canon_sha256":"01f55e08c98f10a5091429eb0ac87088ba3825e5b1faaf66cb940b78fe10454b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:46:20.222480Z","signature_b64":"yIOYNYEL3dNdTljKOCL0F63hKeWcRkmDuXWs6jBZ0GsGVpi6XF8eLhzLn1v2Y5C5Bd/5LzLqeuOGfKMxys8+AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e8d76c423e48158a4d01103e11be77a5e8d7b299690affc9b0b85022fdf8a021","last_reissued_at":"2026-07-05T10:46:20.222017Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:46:20.222017Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"No Detail Left Behind: Revisiting Self-Retrieval for Fine-Grained Image Captioning","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Darshan Singh, Makarand Tapaswi, Manu Gaur","submitted_at":"2024-09-04T18:32:39Z","abstract_excerpt":"Image captioning systems are unable to generate fine-grained captions as they are trained on data that is either noisy (alt-text) or generic (human annotations). This is further exacerbated by maximum likelihood training that encourages generation of frequently occurring phrases. Previous works have tried to address this limitation by fine-tuning captioners with a self-retrieval (SR) reward. However, we find that SR fine-tuning has a tendency to reduce caption faithfulness and even hallucinate. In this work, we circumvent this bottleneck by improving the MLE initialization of the captioning sy"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.03025","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/2409.03025/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":"2409.03025","created_at":"2026-07-05T10:46:20.222074+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.03025v2","created_at":"2026-07-05T10:46:20.222074+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.03025","created_at":"2026-07-05T10:46:20.222074+00:00"},{"alias_kind":"pith_short_12","alias_value":"5DLWYQR6JAKY","created_at":"2026-07-05T10:46:20.222074+00:00"},{"alias_kind":"pith_short_16","alias_value":"5DLWYQR6JAKYUTIB","created_at":"2026-07-05T10:46:20.222074+00:00"},{"alias_kind":"pith_short_8","alias_value":"5DLWYQR6","created_at":"2026-07-05T10:46:20.222074+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX","json":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX.json","graph_json":"https://pith.science/api/pith-number/5DLWYQR6JAKYUTIBCA7BDPTXUX/graph.json","events_json":"https://pith.science/api/pith-number/5DLWYQR6JAKYUTIBCA7BDPTXUX/events.json","paper":"https://pith.science/paper/5DLWYQR6"},"agent_actions":{"view_html":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX","download_json":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX.json","view_paper":"https://pith.science/paper/5DLWYQR6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.03025&json=true","fetch_graph":"https://pith.science/api/pith-number/5DLWYQR6JAKYUTIBCA7BDPTXUX/graph.json","fetch_events":"https://pith.science/api/pith-number/5DLWYQR6JAKYUTIBCA7BDPTXUX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX/action/storage_attestation","attest_author":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX/action/author_attestation","sign_citation":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX/action/citation_signature","submit_replication":"https://pith.science/pith/5DLWYQR6JAKYUTIBCA7BDPTXUX/action/replication_record"}},"created_at":"2026-07-05T10:46:20.222074+00:00","updated_at":"2026-07-05T10:46:20.222074+00:00"}