{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:PVV6OFE6L7ONYM7FYEOYFPCF4Y","short_pith_number":"pith:PVV6OFE6","schema_version":"1.0","canonical_sha256":"7d6be7149e5fdcdc33e5c11d82bc45e60b604cd52efcedbd08c5f9a0496e2e65","source":{"kind":"arxiv","id":"2206.08954","version":2},"attestation_state":"computed","paper":{"title":"Bag of Image Patch Embedding Behind the Success of Self-Supervised Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Adrien Bardes, Yann LeCun, Yubei Chen, Zengyi Li","submitted_at":"2022-06-17T18:11:23Z","abstract_excerpt":"Self-supervised learning (SSL) has recently achieved tremendous empirical advancements in learning image representation. However, our understanding of the principle behind learning such a representation is still limited. This work shows that joint-embedding SSL approaches primarily learn a representation of image patches, which reflects their co-occurrence. Such a connection to co-occurrence modeling can be established formally, and it supplements the prevailing invariance perspective. We empirically show that learning a representation for fixed-scale patches and aggregating local patch repres"},"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":"2206.08954","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-06-17T18:11:23Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8e42720809e5b195bd1cd853aea96f78e122ce6e73334ffaf64c003a237b4f39","abstract_canon_sha256":"771d16ff55574e55f38de4087dbce09024d4e7a37572287f105ae519366730e5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:19:52.294035Z","signature_b64":"MKlTuYbTkgWMGuH74Ts949pJQ5O6LU6rYPyuO5oHWc0CWQodI0ybuOeTAht6g7fKzNFzMbYKP1bMyy7sf0U0CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7d6be7149e5fdcdc33e5c11d82bc45e60b604cd52efcedbd08c5f9a0496e2e65","last_reissued_at":"2026-07-05T06:19:52.293616Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:19:52.293616Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Bag of Image Patch Embedding Behind the Success of Self-Supervised Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Adrien Bardes, Yann LeCun, Yubei Chen, Zengyi Li","submitted_at":"2022-06-17T18:11:23Z","abstract_excerpt":"Self-supervised learning (SSL) has recently achieved tremendous empirical advancements in learning image representation. However, our understanding of the principle behind learning such a representation is still limited. This work shows that joint-embedding SSL approaches primarily learn a representation of image patches, which reflects their co-occurrence. Such a connection to co-occurrence modeling can be established formally, and it supplements the prevailing invariance perspective. We empirically show that learning a representation for fixed-scale patches and aggregating local patch repres"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2206.08954","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/2206.08954/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":"2206.08954","created_at":"2026-07-05T06:19:52.293675+00:00"},{"alias_kind":"arxiv_version","alias_value":"2206.08954v2","created_at":"2026-07-05T06:19:52.293675+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2206.08954","created_at":"2026-07-05T06:19:52.293675+00:00"},{"alias_kind":"pith_short_12","alias_value":"PVV6OFE6L7ON","created_at":"2026-07-05T06:19:52.293675+00:00"},{"alias_kind":"pith_short_16","alias_value":"PVV6OFE6L7ONYM7F","created_at":"2026-07-05T06:19:52.293675+00:00"},{"alias_kind":"pith_short_8","alias_value":"PVV6OFE6","created_at":"2026-07-05T06:19:52.293675+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2404.08471","citing_title":"Revisiting Feature Prediction for Learning Visual Representations from Video","ref_index":137,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y","json":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y.json","graph_json":"https://pith.science/api/pith-number/PVV6OFE6L7ONYM7FYEOYFPCF4Y/graph.json","events_json":"https://pith.science/api/pith-number/PVV6OFE6L7ONYM7FYEOYFPCF4Y/events.json","paper":"https://pith.science/paper/PVV6OFE6"},"agent_actions":{"view_html":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y","download_json":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y.json","view_paper":"https://pith.science/paper/PVV6OFE6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2206.08954&json=true","fetch_graph":"https://pith.science/api/pith-number/PVV6OFE6L7ONYM7FYEOYFPCF4Y/graph.json","fetch_events":"https://pith.science/api/pith-number/PVV6OFE6L7ONYM7FYEOYFPCF4Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y/action/storage_attestation","attest_author":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y/action/author_attestation","sign_citation":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y/action/citation_signature","submit_replication":"https://pith.science/pith/PVV6OFE6L7ONYM7FYEOYFPCF4Y/action/replication_record"}},"created_at":"2026-07-05T06:19:52.293675+00:00","updated_at":"2026-07-05T06:19:52.293675+00:00"}