{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:7OGH43QS2EPIY7NAE3IZ5B6CX5","short_pith_number":"pith:7OGH43QS","schema_version":"1.0","canonical_sha256":"fb8c7e6e12d11e8c7da026d19e87c2bf60b83c246a3a99d54db2949324551521","source":{"kind":"arxiv","id":"2110.07439","version":2},"attestation_state":"computed","paper":{"title":"Inverse Problems Leveraging Pre-trained Contrastive Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Alexandros G. Dimakis, Georgios Smyrnis, Matt Jordan, Sriram Ravula","submitted_at":"2021-10-14T15:06:30Z","abstract_excerpt":"We study a new family of inverse problems for recovering representations of corrupted data. We assume access to a pre-trained representation learning network R(x) that operates on clean images, like CLIP. The problem is to recover the representation of an image R(x), if we are only given a corrupted version A(x), for some known forward operator A. We propose a supervised inversion method that uses a contrastive objective to obtain excellent representations for highly corrupted images. Using a linear probe on our robust representations, we achieve a higher accuracy than end-to-end supervised ba"},"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":"2110.07439","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-14T15:06:30Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"ee227687b063d161f6921b9d8b2a6a1cdd441b8f85d9d4b51351be2ee79bbeb2","abstract_canon_sha256":"76ee6ed242e3060e78c597e9eb081c03ad59eecb3818442a3932b7acb65ad808"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:26:21.070964Z","signature_b64":"XBgnphnQZJHelApaBmkWfmoLKvm0674YUoGMVfaUWbNFl/683/4qP5UYKwdJFX6z39PKTh5fmG1+/yjpl2R/Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fb8c7e6e12d11e8c7da026d19e87c2bf60b83c246a3a99d54db2949324551521","last_reissued_at":"2026-07-05T03:26:21.070532Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:26:21.070532Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Inverse Problems Leveraging Pre-trained Contrastive Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Alexandros G. Dimakis, Georgios Smyrnis, Matt Jordan, Sriram Ravula","submitted_at":"2021-10-14T15:06:30Z","abstract_excerpt":"We study a new family of inverse problems for recovering representations of corrupted data. We assume access to a pre-trained representation learning network R(x) that operates on clean images, like CLIP. The problem is to recover the representation of an image R(x), if we are only given a corrupted version A(x), for some known forward operator A. We propose a supervised inversion method that uses a contrastive objective to obtain excellent representations for highly corrupted images. Using a linear probe on our robust representations, we achieve a higher accuracy than end-to-end supervised ba"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.07439","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/2110.07439/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":"2110.07439","created_at":"2026-07-05T03:26:21.070590+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.07439v2","created_at":"2026-07-05T03:26:21.070590+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.07439","created_at":"2026-07-05T03:26:21.070590+00:00"},{"alias_kind":"pith_short_12","alias_value":"7OGH43QS2EPI","created_at":"2026-07-05T03:26:21.070590+00:00"},{"alias_kind":"pith_short_16","alias_value":"7OGH43QS2EPIY7NA","created_at":"2026-07-05T03:26:21.070590+00:00"},{"alias_kind":"pith_short_8","alias_value":"7OGH43QS","created_at":"2026-07-05T03:26:21.070590+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5","json":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5.json","graph_json":"https://pith.science/api/pith-number/7OGH43QS2EPIY7NAE3IZ5B6CX5/graph.json","events_json":"https://pith.science/api/pith-number/7OGH43QS2EPIY7NAE3IZ5B6CX5/events.json","paper":"https://pith.science/paper/7OGH43QS"},"agent_actions":{"view_html":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5","download_json":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5.json","view_paper":"https://pith.science/paper/7OGH43QS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.07439&json=true","fetch_graph":"https://pith.science/api/pith-number/7OGH43QS2EPIY7NAE3IZ5B6CX5/graph.json","fetch_events":"https://pith.science/api/pith-number/7OGH43QS2EPIY7NAE3IZ5B6CX5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5/action/storage_attestation","attest_author":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5/action/author_attestation","sign_citation":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5/action/citation_signature","submit_replication":"https://pith.science/pith/7OGH43QS2EPIY7NAE3IZ5B6CX5/action/replication_record"}},"created_at":"2026-07-05T03:26:21.070590+00:00","updated_at":"2026-07-05T03:26:21.070590+00:00"}