{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:U5TMGCJRBMDJSPL34S2XD52LVS","short_pith_number":"pith:U5TMGCJR","schema_version":"1.0","canonical_sha256":"a766c309310b06993d7be4b571f74bac989f04f254bc04dfe8b1ab6ddc86306e","source":{"kind":"arxiv","id":"2606.22766","version":1},"attestation_state":"computed","paper":{"title":"READ More than What You See: Reinforcement Learning for Accurate and Coherent Audio Description Generations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Antoni B. Chan, Bo Fang, Hang Zhou, Hui Zhang, Xinyao Zhang, Yuxin Song","submitted_at":"2026-06-22T02:05:38Z","abstract_excerpt":"Audio Description aims to generate concise narrations of essential visual content in audio-visual media for blind and low-vision audiences. Existing methods either rely on prompting off-the-shelf multimodal models, which often mismatch AD style, or partially optimize training-based systems with next-token prediction, which under-explores model capacity and biases generation toward generic expressions. We present READ, the first reinforcement-learning (RL) framework for training-based AD generation. READ formulates AD as sequence-level optimization with reference-matching, length, and format re"},"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":"2606.22766","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-22T02:05:38Z","cross_cats_sorted":[],"title_canon_sha256":"7c51a19d99cf391d066a60d8319cc394ae258c773590db81ba0b77858996b907","abstract_canon_sha256":"8baacf5223dbecc01fc5e8c12f08e6d00323d9aacd49f3aabeae62bfbe996fe8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:53.937101Z","signature_b64":"l7wOGm21vUfxNJbmgu6rpDfq7XQSNCRn2YwSxi8ZNMA1iPZfwEOv5b5C1E+utvqfQVrVjDvNuSNgbbUI9Dt3Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a766c309310b06993d7be4b571f74bac989f04f254bc04dfe8b1ab6ddc86306e","last_reissued_at":"2026-06-23T02:13:53.936684Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:53.936684Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"READ More than What You See: Reinforcement Learning for Accurate and Coherent Audio Description Generations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Antoni B. Chan, Bo Fang, Hang Zhou, Hui Zhang, Xinyao Zhang, Yuxin Song","submitted_at":"2026-06-22T02:05:38Z","abstract_excerpt":"Audio Description aims to generate concise narrations of essential visual content in audio-visual media for blind and low-vision audiences. Existing methods either rely on prompting off-the-shelf multimodal models, which often mismatch AD style, or partially optimize training-based systems with next-token prediction, which under-explores model capacity and biases generation toward generic expressions. We present READ, the first reinforcement-learning (RL) framework for training-based AD generation. READ formulates AD as sequence-level optimization with reference-matching, length, and format re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22766","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/2606.22766/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":"2606.22766","created_at":"2026-06-23T02:13:53.936751+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.22766v1","created_at":"2026-06-23T02:13:53.936751+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22766","created_at":"2026-06-23T02:13:53.936751+00:00"},{"alias_kind":"pith_short_12","alias_value":"U5TMGCJRBMDJ","created_at":"2026-06-23T02:13:53.936751+00:00"},{"alias_kind":"pith_short_16","alias_value":"U5TMGCJRBMDJSPL3","created_at":"2026-06-23T02:13:53.936751+00:00"},{"alias_kind":"pith_short_8","alias_value":"U5TMGCJR","created_at":"2026-06-23T02:13:53.936751+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/U5TMGCJRBMDJSPL34S2XD52LVS","json":"https://pith.science/pith/U5TMGCJRBMDJSPL34S2XD52LVS.json","graph_json":"https://pith.science/api/pith-number/U5TMGCJRBMDJSPL34S2XD52LVS/graph.json","events_json":"https://pith.science/api/pith-number/U5TMGCJRBMDJSPL34S2XD52LVS/events.json","paper":"https://pith.science/paper/U5TMGCJR"},"agent_actions":{"view_html":"https://pith.science/pith/U5TMGCJRBMDJSPL34S2XD52LVS","download_json":"https://pith.science/pith/U5TMGCJRBMDJSPL34S2XD52LVS.json","view_paper":"https://pith.science/paper/U5TMGCJR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.22766&json=true","fetch_graph":"https://pith.science/api/pith-number/U5TMGCJRBMDJSPL34S2XD52LVS/graph.json","fetch_events":"https://pith.science/api/pith-number/U5TMGCJRBMDJSPL34S2XD52LVS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/U5TMGCJRBMDJSPL34S2XD52LVS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/U5TMGCJRBMDJSPL34S2XD52LVS/action/storage_attestation","attest_author":"https://pith.science/pith/U5TMGCJRBMDJSPL34S2XD52LVS/action/author_attestation","sign_citation":"https://pith.science/pith/U5TMGCJRBMDJSPL34S2XD52LVS/action/citation_signature","submit_replication":"https://pith.science/pith/U5TMGCJRBMDJSPL34S2XD52LVS/action/replication_record"}},"created_at":"2026-06-23T02:13:53.936751+00:00","updated_at":"2026-06-23T02:13:53.936751+00:00"}