{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:PRLEEQEOAGU62NV3PYIMI2CN26","short_pith_number":"pith:PRLEEQEO","schema_version":"1.0","canonical_sha256":"7c5642408e01a9ed36bb7e10c4684dd7a29a61a4e8d5b8d04682b177148c5027","source":{"kind":"arxiv","id":"2306.05268","version":2},"attestation_state":"computed","paper":{"title":"Factorized Contrastive Learning: Going Beyond Multi-view Redundancy","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CV","cs.MM"],"primary_cat":"cs.LG","authors_text":"James Zou, Louis-Philippe Morency, Martin Ma, Paul Pu Liang, Ruslan Salakhutdinov, Zihao Deng","submitted_at":"2023-06-08T15:17:04Z","abstract_excerpt":"In a wide range of multimodal tasks, contrastive learning has become a particularly appealing approach since it can successfully learn representations from abundant unlabeled data with only pairing information (e.g., image-caption or video-audio pairs). Underpinning these approaches is the assumption of multi-view redundancy - that shared information between modalities is necessary and sufficient for downstream tasks. However, in many real-world settings, task-relevant information is also contained in modality-unique regions: information that is only present in one modality but still relevant "},"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":"2306.05268","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-06-08T15:17:04Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CV","cs.MM"],"title_canon_sha256":"d38f18e766159bc2771d1e9c8071f738a07a21f239ec8ef56a9ee2550f2033a4","abstract_canon_sha256":"f75e11d7a73ea41660141edab64363fe962dc0be61342dac29c8be66f7927a0c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:06:29.682999Z","signature_b64":"HKp9n6ifDPAU7mGHdb1POEuoZxjJ4C9JsuniqESHjd2u+XFgCfG3uVKy+rlj4p7Psx2ibhnDNWB+QUeaeSMFAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7c5642408e01a9ed36bb7e10c4684dd7a29a61a4e8d5b8d04682b177148c5027","last_reissued_at":"2026-07-05T07:06:29.682510Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:06:29.682510Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Factorized Contrastive Learning: Going Beyond Multi-view Redundancy","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CV","cs.MM"],"primary_cat":"cs.LG","authors_text":"James Zou, Louis-Philippe Morency, Martin Ma, Paul Pu Liang, Ruslan Salakhutdinov, Zihao Deng","submitted_at":"2023-06-08T15:17:04Z","abstract_excerpt":"In a wide range of multimodal tasks, contrastive learning has become a particularly appealing approach since it can successfully learn representations from abundant unlabeled data with only pairing information (e.g., image-caption or video-audio pairs). Underpinning these approaches is the assumption of multi-view redundancy - that shared information between modalities is necessary and sufficient for downstream tasks. However, in many real-world settings, task-relevant information is also contained in modality-unique regions: information that is only present in one modality but still relevant "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.05268","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/2306.05268/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":"2306.05268","created_at":"2026-07-05T07:06:29.682572+00:00"},{"alias_kind":"arxiv_version","alias_value":"2306.05268v2","created_at":"2026-07-05T07:06:29.682572+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.05268","created_at":"2026-07-05T07:06:29.682572+00:00"},{"alias_kind":"pith_short_12","alias_value":"PRLEEQEOAGU6","created_at":"2026-07-05T07:06:29.682572+00:00"},{"alias_kind":"pith_short_16","alias_value":"PRLEEQEOAGU62NV3","created_at":"2026-07-05T07:06:29.682572+00:00"},{"alias_kind":"pith_short_8","alias_value":"PRLEEQEO","created_at":"2026-07-05T07:06:29.682572+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.31504","citing_title":"When Are Multimodal Predictions Biologically Supported? A Diagnostic Evaluation Framework","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2401.01335","citing_title":"Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26","json":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26.json","graph_json":"https://pith.science/api/pith-number/PRLEEQEOAGU62NV3PYIMI2CN26/graph.json","events_json":"https://pith.science/api/pith-number/PRLEEQEOAGU62NV3PYIMI2CN26/events.json","paper":"https://pith.science/paper/PRLEEQEO"},"agent_actions":{"view_html":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26","download_json":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26.json","view_paper":"https://pith.science/paper/PRLEEQEO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2306.05268&json=true","fetch_graph":"https://pith.science/api/pith-number/PRLEEQEOAGU62NV3PYIMI2CN26/graph.json","fetch_events":"https://pith.science/api/pith-number/PRLEEQEOAGU62NV3PYIMI2CN26/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26/action/storage_attestation","attest_author":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26/action/author_attestation","sign_citation":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26/action/citation_signature","submit_replication":"https://pith.science/pith/PRLEEQEOAGU62NV3PYIMI2CN26/action/replication_record"}},"created_at":"2026-07-05T07:06:29.682572+00:00","updated_at":"2026-07-05T07:06:29.682572+00:00"}