{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:O7DVF5PWOHW7KHJVWQUB3QZEF7","short_pith_number":"pith:O7DVF5PW","schema_version":"1.0","canonical_sha256":"77c752f5f671edf51d35b4281dc3242fd6cd3f6dead9a48b7d3c044a99d597dc","source":{"kind":"arxiv","id":"2409.10365","version":2},"attestation_state":"computed","paper":{"title":"Robust image representations with counterfactual contrastive learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Ben Glocker, Fabio De Sousa Ribeiro, Galvin Khara, M\\'elanie Roschewitz, Tian Xia","submitted_at":"2024-09-16T15:11:00Z","abstract_excerpt":"Contrastive pretraining can substantially increase model generalisation and downstream performance. However, the quality of the learned representations is highly dependent on the data augmentation strategy applied to generate positive pairs. Positive contrastive pairs should preserve semantic meaning while discarding unwanted variations related to the data acquisition domain. Traditional contrastive pipelines attempt to simulate domain shifts through pre-defined generic image transformations. However, these do not always mimic realistic and relevant domain variations for medical imaging, such "},"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":"2409.10365","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-09-16T15:11:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3358aafdfb0c9a6d6cf8d9499f653c1ea32b70af505efd58fe65d3a3729c35c3","abstract_canon_sha256":"35262f944c63d7cd401ffbdfa506c0c9bf1a274fcd0feb9ed63a035bb6cec6a8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:21:35.810383Z","signature_b64":"3yO64Qr7yxfLQBJvktp/A5NYw7j3TskCPU65kV7VGYsOldEYeV3YmBbJy0VP9xOnOnDAxwjhoZzjjwmF1mmxCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"77c752f5f671edf51d35b4281dc3242fd6cd3f6dead9a48b7d3c044a99d597dc","last_reissued_at":"2026-07-05T11:21:35.809847Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:21:35.809847Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Robust image representations with counterfactual contrastive learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Ben Glocker, Fabio De Sousa Ribeiro, Galvin Khara, M\\'elanie Roschewitz, Tian Xia","submitted_at":"2024-09-16T15:11:00Z","abstract_excerpt":"Contrastive pretraining can substantially increase model generalisation and downstream performance. However, the quality of the learned representations is highly dependent on the data augmentation strategy applied to generate positive pairs. Positive contrastive pairs should preserve semantic meaning while discarding unwanted variations related to the data acquisition domain. Traditional contrastive pipelines attempt to simulate domain shifts through pre-defined generic image transformations. However, these do not always mimic realistic and relevant domain variations for medical imaging, such "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.10365","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/2409.10365/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":"2409.10365","created_at":"2026-07-05T11:21:35.809912+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.10365v2","created_at":"2026-07-05T11:21:35.809912+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.10365","created_at":"2026-07-05T11:21:35.809912+00:00"},{"alias_kind":"pith_short_12","alias_value":"O7DVF5PWOHW7","created_at":"2026-07-05T11:21:35.809912+00:00"},{"alias_kind":"pith_short_16","alias_value":"O7DVF5PWOHW7KHJV","created_at":"2026-07-05T11:21:35.809912+00:00"},{"alias_kind":"pith_short_8","alias_value":"O7DVF5PW","created_at":"2026-07-05T11:21:35.809912+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.12057","citing_title":"Unified 3D MRI Representations via Sequence-Invariant Contrastive Learning","ref_index":22,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7","json":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7.json","graph_json":"https://pith.science/api/pith-number/O7DVF5PWOHW7KHJVWQUB3QZEF7/graph.json","events_json":"https://pith.science/api/pith-number/O7DVF5PWOHW7KHJVWQUB3QZEF7/events.json","paper":"https://pith.science/paper/O7DVF5PW"},"agent_actions":{"view_html":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7","download_json":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7.json","view_paper":"https://pith.science/paper/O7DVF5PW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.10365&json=true","fetch_graph":"https://pith.science/api/pith-number/O7DVF5PWOHW7KHJVWQUB3QZEF7/graph.json","fetch_events":"https://pith.science/api/pith-number/O7DVF5PWOHW7KHJVWQUB3QZEF7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7/action/storage_attestation","attest_author":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7/action/author_attestation","sign_citation":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7/action/citation_signature","submit_replication":"https://pith.science/pith/O7DVF5PWOHW7KHJVWQUB3QZEF7/action/replication_record"}},"created_at":"2026-07-05T11:21:35.809912+00:00","updated_at":"2026-07-05T11:21:35.809912+00:00"}