{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:WOGWWQYPBRLU5Z4QZL42WDOZAB","short_pith_number":"pith:WOGWWQYP","schema_version":"1.0","canonical_sha256":"b38d6b430f0c574ee790caf9ab0dd9005a03fd79578f7d757d9752d4c7125b4e","source":{"kind":"arxiv","id":"2303.17376","version":1},"attestation_state":"computed","paper":{"title":"A Study of Autoregressive Decoders for Multi-Tasking in Computer Vision","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexander Kolesnikov, Andreas Steiner, Andr\\'e Susano Pinto, Bo Wan, Emanuele Bugliarello, Filip Pavetic, Gagan Madan, Liang-Chieh Chen, Lucas Beyer, Qihang Yu, Xiaohua Zhai, Xiao Wang","submitted_at":"2023-03-30T13:42:58Z","abstract_excerpt":"There has been a recent explosion of computer vision models which perform many tasks and are composed of an image encoder (usually a ViT) and an autoregressive decoder (usually a Transformer). However, most of this work simply presents one system and its results, leaving many questions regarding design decisions and trade-offs of such systems unanswered. In this work, we aim to provide such answers. We take a close look at autoregressive decoders for multi-task learning in multimodal computer vision, including classification, captioning, visual question answering, and optical character recogni"},"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":"2303.17376","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-03-30T13:42:58Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"4c5030ca691bf5a45059542b978c3b5d7af6871b8e23158232c31a574679c166","abstract_canon_sha256":"da846752deb7e5579bc3678e47c9287c3e4e38aa2c3a1e2249b1fdd70cfc8580"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:56:33.367421Z","signature_b64":"HBV0axlKzN7OdxzcGVDY0zL69yQClO8k72PmXGcYUq0AdV4oXwrAR09VVTw94OkW8+wSEkYjH4h62az8hq0WAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b38d6b430f0c574ee790caf9ab0dd9005a03fd79578f7d757d9752d4c7125b4e","last_reissued_at":"2026-07-05T05:56:33.366912Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:56:33.366912Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Study of Autoregressive Decoders for Multi-Tasking in Computer Vision","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexander Kolesnikov, Andreas Steiner, Andr\\'e Susano Pinto, Bo Wan, Emanuele Bugliarello, Filip Pavetic, Gagan Madan, Liang-Chieh Chen, Lucas Beyer, Qihang Yu, Xiaohua Zhai, Xiao Wang","submitted_at":"2023-03-30T13:42:58Z","abstract_excerpt":"There has been a recent explosion of computer vision models which perform many tasks and are composed of an image encoder (usually a ViT) and an autoregressive decoder (usually a Transformer). However, most of this work simply presents one system and its results, leaving many questions regarding design decisions and trade-offs of such systems unanswered. In this work, we aim to provide such answers. We take a close look at autoregressive decoders for multi-task learning in multimodal computer vision, including classification, captioning, visual question answering, and optical character recogni"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.17376","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/2303.17376/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":"2303.17376","created_at":"2026-07-05T05:56:33.366968+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.17376v1","created_at":"2026-07-05T05:56:33.366968+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.17376","created_at":"2026-07-05T05:56:33.366968+00:00"},{"alias_kind":"pith_short_12","alias_value":"WOGWWQYPBRLU","created_at":"2026-07-05T05:56:33.366968+00:00"},{"alias_kind":"pith_short_16","alias_value":"WOGWWQYPBRLU5Z4Q","created_at":"2026-07-05T05:56:33.366968+00:00"},{"alias_kind":"pith_short_8","alias_value":"WOGWWQYP","created_at":"2026-07-05T05:56:33.366968+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.03245","citing_title":"Text-Conditional JEPA for Learning Semantically Rich Visual Representations","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2407.07726","citing_title":"PaliGemma: A versatile 3B VLM for transfer","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB","json":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB.json","graph_json":"https://pith.science/api/pith-number/WOGWWQYPBRLU5Z4QZL42WDOZAB/graph.json","events_json":"https://pith.science/api/pith-number/WOGWWQYPBRLU5Z4QZL42WDOZAB/events.json","paper":"https://pith.science/paper/WOGWWQYP"},"agent_actions":{"view_html":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB","download_json":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB.json","view_paper":"https://pith.science/paper/WOGWWQYP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.17376&json=true","fetch_graph":"https://pith.science/api/pith-number/WOGWWQYPBRLU5Z4QZL42WDOZAB/graph.json","fetch_events":"https://pith.science/api/pith-number/WOGWWQYPBRLU5Z4QZL42WDOZAB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB/action/storage_attestation","attest_author":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB/action/author_attestation","sign_citation":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB/action/citation_signature","submit_replication":"https://pith.science/pith/WOGWWQYPBRLU5Z4QZL42WDOZAB/action/replication_record"}},"created_at":"2026-07-05T05:56:33.366968+00:00","updated_at":"2026-07-05T05:56:33.366968+00:00"}