{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:5IEVLLHVL6CERV2W3RSDSPWRYX","short_pith_number":"pith:5IEVLLHV","canonical_record":{"source":{"id":"2105.08054","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-05-17T17:59:03Z","cross_cats_sorted":[],"title_canon_sha256":"10af960e526959812a9cd44782af28643fcdcdb2488203c1b6b279bd2edc0f29","abstract_canon_sha256":"e3feec5f144010ed79cbc368fd3603e3ae35ed8bb09a16ea9b246e77c23e1fff"},"schema_version":"1.0"},"canonical_sha256":"ea0955acf55f8448d756dc64393ed1c5d9f09ceba8b6c7262f0274783e419d33","source":{"kind":"arxiv","id":"2105.08054","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2105.08054","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"arxiv_version","alias_value":"2105.08054v1","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.08054","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"pith_short_12","alias_value":"5IEVLLHVL6CE","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"pith_short_16","alias_value":"5IEVLLHVL6CERV2W","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"pith_short_8","alias_value":"5IEVLLHV","created_at":"2026-07-05T02:40:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:5IEVLLHVL6CERV2W3RSDSPWRYX","target":"record","payload":{"canonical_record":{"source":{"id":"2105.08054","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-05-17T17:59:03Z","cross_cats_sorted":[],"title_canon_sha256":"10af960e526959812a9cd44782af28643fcdcdb2488203c1b6b279bd2edc0f29","abstract_canon_sha256":"e3feec5f144010ed79cbc368fd3603e3ae35ed8bb09a16ea9b246e77c23e1fff"},"schema_version":"1.0"},"canonical_sha256":"ea0955acf55f8448d756dc64393ed1c5d9f09ceba8b6c7262f0274783e419d33","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:40:47.335668Z","signature_b64":"L3Sy6PEx9fCiPhRrP8Qgr/6IFFL9J7T8HkMvfWZB9xYuarZVEm6xVwIMGOnqw8WJsfm7f6jPU6IjIgJaoCR7AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ea0955acf55f8448d756dc64393ed1c5d9f09ceba8b6c7262f0274783e419d33","last_reissued_at":"2026-07-05T02:40:47.335126Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:40:47.335126Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2105.08054","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T02:40:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SPVC4ZxanE/kWa3d9w/wOGGeikpEONISu2RsY4xj2bfvOXVn0nQjAYR+JI3OhJovsWf+vx5fg8mbWv/3f8IyBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T00:59:36.231922Z"},"content_sha256":"6c4b2ede21b1ac629136ca2cfecca8667d13430a23f33f617abb6d3ff4fe7909","schema_version":"1.0","event_id":"sha256:6c4b2ede21b1ac629136ca2cfecca8667d13430a23f33f617abb6d3ff4fe7909"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:5IEVLLHVL6CERV2W3RSDSPWRYX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Divide and Contrast: Self-supervised Learning from Uncurated Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Aaron van den Oord, Olivier J. Henaff, Yonglong Tian","submitted_at":"2021-05-17T17:59:03Z","abstract_excerpt":"Self-supervised learning holds promise in leveraging large amounts of unlabeled data, however much of its progress has thus far been limited to highly curated pre-training data such as ImageNet. We explore the effects of contrastive learning from larger, less-curated image datasets such as YFCC, and find there is indeed a large difference in the resulting representation quality. We hypothesize that this curation gap is due to a shift in the distribution of image classes -- which is more diverse and heavy-tailed -- resulting in less relevant negative samples to learn from. We test this hypothes"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.08054","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/2105.08054/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T02:40:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3G6sexCCnDrhXpahTB8rKyxmspRadPwymlgaacZMgTjXfLYpleV0Dw8tSTr9jgt1T7bIU3iSCGjxdIt7HbP1Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T00:59:36.232438Z"},"content_sha256":"9cde31c0540c96e057d6bedee1e53d98ff49ff5f020902d4011ee94a51e051b1","schema_version":"1.0","event_id":"sha256:9cde31c0540c96e057d6bedee1e53d98ff49ff5f020902d4011ee94a51e051b1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5IEVLLHVL6CERV2W3RSDSPWRYX/bundle.json","state_url":"https://pith.science/pith/5IEVLLHVL6CERV2W3RSDSPWRYX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5IEVLLHVL6CERV2W3RSDSPWRYX/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-04T00:59:36Z","links":{"resolver":"https://pith.science/pith/5IEVLLHVL6CERV2W3RSDSPWRYX","bundle":"https://pith.science/pith/5IEVLLHVL6CERV2W3RSDSPWRYX/bundle.json","state":"https://pith.science/pith/5IEVLLHVL6CERV2W3RSDSPWRYX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5IEVLLHVL6CERV2W3RSDSPWRYX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:5IEVLLHVL6CERV2W3RSDSPWRYX","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"e3feec5f144010ed79cbc368fd3603e3ae35ed8bb09a16ea9b246e77c23e1fff","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-05-17T17:59:03Z","title_canon_sha256":"10af960e526959812a9cd44782af28643fcdcdb2488203c1b6b279bd2edc0f29"},"schema_version":"1.0","source":{"id":"2105.08054","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2105.08054","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"arxiv_version","alias_value":"2105.08054v1","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.08054","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"pith_short_12","alias_value":"5IEVLLHVL6CE","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"pith_short_16","alias_value":"5IEVLLHVL6CERV2W","created_at":"2026-07-05T02:40:47Z"},{"alias_kind":"pith_short_8","alias_value":"5IEVLLHV","created_at":"2026-07-05T02:40:47Z"}],"graph_snapshots":[{"event_id":"sha256:9cde31c0540c96e057d6bedee1e53d98ff49ff5f020902d4011ee94a51e051b1","target":"graph","created_at":"2026-07-05T02:40:47Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2105.08054/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Self-supervised learning holds promise in leveraging large amounts of unlabeled data, however much of its progress has thus far been limited to highly curated pre-training data such as ImageNet. We explore the effects of contrastive learning from larger, less-curated image datasets such as YFCC, and find there is indeed a large difference in the resulting representation quality. We hypothesize that this curation gap is due to a shift in the distribution of image classes -- which is more diverse and heavy-tailed -- resulting in less relevant negative samples to learn from. We test this hypothes","authors_text":"Aaron van den Oord, Olivier J. Henaff, Yonglong Tian","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-05-17T17:59:03Z","title":"Divide and Contrast: Self-supervised Learning from Uncurated Data"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.08054","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6c4b2ede21b1ac629136ca2cfecca8667d13430a23f33f617abb6d3ff4fe7909","target":"record","created_at":"2026-07-05T02:40:47Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"e3feec5f144010ed79cbc368fd3603e3ae35ed8bb09a16ea9b246e77c23e1fff","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-05-17T17:59:03Z","title_canon_sha256":"10af960e526959812a9cd44782af28643fcdcdb2488203c1b6b279bd2edc0f29"},"schema_version":"1.0","source":{"id":"2105.08054","kind":"arxiv","version":1}},"canonical_sha256":"ea0955acf55f8448d756dc64393ed1c5d9f09ceba8b6c7262f0274783e419d33","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ea0955acf55f8448d756dc64393ed1c5d9f09ceba8b6c7262f0274783e419d33","first_computed_at":"2026-07-05T02:40:47.335126Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T02:40:47.335126Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"L3Sy6PEx9fCiPhRrP8Qgr/6IFFL9J7T8HkMvfWZB9xYuarZVEm6xVwIMGOnqw8WJsfm7f6jPU6IjIgJaoCR7AQ==","signature_status":"signed_v1","signed_at":"2026-07-05T02:40:47.335668Z","signed_message":"canonical_sha256_bytes"},"source_id":"2105.08054","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6c4b2ede21b1ac629136ca2cfecca8667d13430a23f33f617abb6d3ff4fe7909","sha256:9cde31c0540c96e057d6bedee1e53d98ff49ff5f020902d4011ee94a51e051b1"],"state_sha256":"99ce831e328c4aa8b23f0851696245c8969ef4f134a11862c8282f32dcdb8ee1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"27bvg32puVIj7RLIoIMUZCKnn2RC4xHdCYgaILgBCnlV5GINlYcwg4ksQeLD0/JkZ2tFr8JajyToW4cK6xnaBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T00:59:36.237157Z","bundle_sha256":"2ab834c519170c52a50474f81c6d3282987fd513204452e05d758cb175feab30"}}