{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:DIVJE7DAXHCMOLEW6KF7DJ7XFJ","short_pith_number":"pith:DIVJE7DA","schema_version":"1.0","canonical_sha256":"1a2a927c60b9c4c72c96f28bf1a7f72a75ecb46c5eaa22cddc37ceda29c24096","source":{"kind":"arxiv","id":"2212.08378","version":1},"attestation_state":"computed","paper":{"title":"Feature Dropout: Revisiting the Role of Augmentations in Contrastive Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.SD","eess.AS"],"primary_cat":"cs.LG","authors_text":"Alex Tamkin, Margalit Glasgow, Noah Goodman, Xiluo He","submitted_at":"2022-12-16T10:08:38Z","abstract_excerpt":"What role do augmentations play in contrastive learning? Recent work suggests that good augmentations are label-preserving with respect to a specific downstream task. We complicate this picture by showing that label-destroying augmentations can be useful in the foundation model setting, where the goal is to learn diverse, general-purpose representations for multiple downstream tasks. We perform contrastive learning experiments on a range of image and audio datasets with multiple downstream tasks (e.g. for digits superimposed on photographs, predicting the class of one vs. the other). We find t"},"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":"2212.08378","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-12-16T10:08:38Z","cross_cats_sorted":["cs.CV","cs.SD","eess.AS"],"title_canon_sha256":"e3e0521c3c8fbd6b6164eed70dbdc40419914d96e74da31d45101631ea7ca96e","abstract_canon_sha256":"76469383700fcabbf2fbbc115238efb56a8c73a3db1b7bdf058b0e7e9e52161d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:25:56.591968Z","signature_b64":"xURz5iuntEeE7CY/wvYy4FtoO/NIOz4cPV2NoDnGjx/b1clzQ/Fr8QaWsoxETw42qRu8h5ag4NTLMeQccH4OCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1a2a927c60b9c4c72c96f28bf1a7f72a75ecb46c5eaa22cddc37ceda29c24096","last_reissued_at":"2026-07-05T05:25:56.591616Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:25:56.591616Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Feature Dropout: Revisiting the Role of Augmentations in Contrastive Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.SD","eess.AS"],"primary_cat":"cs.LG","authors_text":"Alex Tamkin, Margalit Glasgow, Noah Goodman, Xiluo He","submitted_at":"2022-12-16T10:08:38Z","abstract_excerpt":"What role do augmentations play in contrastive learning? Recent work suggests that good augmentations are label-preserving with respect to a specific downstream task. We complicate this picture by showing that label-destroying augmentations can be useful in the foundation model setting, where the goal is to learn diverse, general-purpose representations for multiple downstream tasks. We perform contrastive learning experiments on a range of image and audio datasets with multiple downstream tasks (e.g. for digits superimposed on photographs, predicting the class of one vs. the other). We find t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.08378","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/2212.08378/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":"2212.08378","created_at":"2026-07-05T05:25:56.591678+00:00"},{"alias_kind":"arxiv_version","alias_value":"2212.08378v1","created_at":"2026-07-05T05:25:56.591678+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.08378","created_at":"2026-07-05T05:25:56.591678+00:00"},{"alias_kind":"pith_short_12","alias_value":"DIVJE7DAXHCM","created_at":"2026-07-05T05:25:56.591678+00:00"},{"alias_kind":"pith_short_16","alias_value":"DIVJE7DAXHCMOLEW","created_at":"2026-07-05T05:25:56.591678+00:00"},{"alias_kind":"pith_short_8","alias_value":"DIVJE7DA","created_at":"2026-07-05T05:25:56.591678+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.18478","citing_title":"Beyond Self-Consistency: Loss-Balanced Perturbation-Based Regularization Improves Industrial-Scale Ads Ranking","ref_index":11,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ","json":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ.json","graph_json":"https://pith.science/api/pith-number/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/graph.json","events_json":"https://pith.science/api/pith-number/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/events.json","paper":"https://pith.science/paper/DIVJE7DA"},"agent_actions":{"view_html":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ","download_json":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ.json","view_paper":"https://pith.science/paper/DIVJE7DA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2212.08378&json=true","fetch_graph":"https://pith.science/api/pith-number/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/graph.json","fetch_events":"https://pith.science/api/pith-number/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/action/storage_attestation","attest_author":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/action/author_attestation","sign_citation":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/action/citation_signature","submit_replication":"https://pith.science/pith/DIVJE7DAXHCMOLEW6KF7DJ7XFJ/action/replication_record"}},"created_at":"2026-07-05T05:25:56.591678+00:00","updated_at":"2026-07-05T05:25:56.591678+00:00"}