{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:ZRG2YGSCE5KP7DX2SXVLYSOT6I","short_pith_number":"pith:ZRG2YGSC","schema_version":"1.0","canonical_sha256":"cc4dac1a422754ff8efa95eabc49d3f2197d9c0ef1e4eabd4c4d5a3f0b9a7064","source":{"kind":"arxiv","id":"2107.07651","version":2},"attestation_state":"computed","paper":{"title":"Align before Fuse: Vision and Language Representation Learning with Momentum Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Akhilesh Deepak Gotmare, Caiming Xiong, Junnan Li, Ramprasaath R. Selvaraju, Shafiq Joty, Steven Hoi","submitted_at":"2021-07-16T00:19:22Z","abstract_excerpt":"Large-scale vision and language representation learning has shown promising improvements on various vision-language tasks. Most existing methods employ a transformer-based multimodal encoder to jointly model visual tokens (region-based image features) and word tokens. Because the visual tokens and word tokens are unaligned, it is challenging for the multimodal encoder to learn image-text interactions. In this paper, we introduce a contrastive loss to ALign the image and text representations BEfore Fusing (ALBEF) them through cross-modal attention, which enables more grounded vision and languag"},"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":"2107.07651","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-07-16T00:19:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"73abf5702cfe6296da282ec7214ae3e30c5c3776033e810241f25cd6df7904d4","abstract_canon_sha256":"c4075835e291128ae9a4cb2cc6b6f37f097d3fdeb3109ec4df54f990a43f2770"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:20:36.971253Z","signature_b64":"WdSvHHKHQ1X63SVAlPfUlZtSQUPPOji+HNQQg2mBBMK8Ll3VN+ceku88j6EjEwvdRQcMIoxBvrD3gLZZBBwyAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cc4dac1a422754ff8efa95eabc49d3f2197d9c0ef1e4eabd4c4d5a3f0b9a7064","last_reissued_at":"2026-07-05T03:20:36.970700Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:20:36.970700Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Align before Fuse: Vision and Language Representation Learning with Momentum Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Akhilesh Deepak Gotmare, Caiming Xiong, Junnan Li, Ramprasaath R. Selvaraju, Shafiq Joty, Steven Hoi","submitted_at":"2021-07-16T00:19:22Z","abstract_excerpt":"Large-scale vision and language representation learning has shown promising improvements on various vision-language tasks. Most existing methods employ a transformer-based multimodal encoder to jointly model visual tokens (region-based image features) and word tokens. Because the visual tokens and word tokens are unaligned, it is challenging for the multimodal encoder to learn image-text interactions. In this paper, we introduce a contrastive loss to ALign the image and text representations BEfore Fusing (ALBEF) them through cross-modal attention, which enables more grounded vision and languag"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2107.07651","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/2107.07651/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":"2107.07651","created_at":"2026-07-05T03:20:36.970762+00:00"},{"alias_kind":"arxiv_version","alias_value":"2107.07651v2","created_at":"2026-07-05T03:20:36.970762+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2107.07651","created_at":"2026-07-05T03:20:36.970762+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZRG2YGSCE5KP","created_at":"2026-07-05T03:20:36.970762+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZRG2YGSCE5KP7DX2","created_at":"2026-07-05T03:20:36.970762+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZRG2YGSC","created_at":"2026-07-05T03:20:36.970762+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.10236","citing_title":"Navigating the Challenges of AI-Generated Image Detection in the Wild: What Truly Matters?","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2510.16416","citing_title":"SSL4RL: Revisiting Self-supervised Learning as Intrinsic Reward for Visual-Language Reasoning","ref_index":36,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I","json":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I.json","graph_json":"https://pith.science/api/pith-number/ZRG2YGSCE5KP7DX2SXVLYSOT6I/graph.json","events_json":"https://pith.science/api/pith-number/ZRG2YGSCE5KP7DX2SXVLYSOT6I/events.json","paper":"https://pith.science/paper/ZRG2YGSC"},"agent_actions":{"view_html":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I","download_json":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I.json","view_paper":"https://pith.science/paper/ZRG2YGSC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2107.07651&json=true","fetch_graph":"https://pith.science/api/pith-number/ZRG2YGSCE5KP7DX2SXVLYSOT6I/graph.json","fetch_events":"https://pith.science/api/pith-number/ZRG2YGSCE5KP7DX2SXVLYSOT6I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I/action/storage_attestation","attest_author":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I/action/author_attestation","sign_citation":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I/action/citation_signature","submit_replication":"https://pith.science/pith/ZRG2YGSCE5KP7DX2SXVLYSOT6I/action/replication_record"}},"created_at":"2026-07-05T03:20:36.970762+00:00","updated_at":"2026-07-05T03:20:36.970762+00:00"}