{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:G2LZ6VBD5AB7253NR5PD57LD2Z","short_pith_number":"pith:G2LZ6VBD","schema_version":"1.0","canonical_sha256":"36979f5423e803fd776d8f5e3efd63d67961845a7b7a0ced67fa568b3e581770","source":{"kind":"arxiv","id":"2506.22881","version":3},"attestation_state":"computed","paper":{"title":"CLIP-like Model as a Foundational Density Ratio Estimator","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Fumiya Uchiyama, Hirokatsu Kataoka, Masahiro Suzuki, Rintaro Yanagi, Shohei Taniguchi, Shota Takashiro, Yusuke Iwasawa, Yutaka Matsuo","submitted_at":"2025-06-28T13:36:44Z","abstract_excerpt":"Density ratio estimation is a core concept in statistical machine learning because it provides a unified mechanism for tasks such as importance weighting, divergence estimation, and likelihood-free inference, but its potential in vision and language models has not been fully explored. Modern vision-language encoders such as CLIP and SigLIP are trained with contrastive objectives that implicitly optimize log density ratios between joint and marginal image-text distributions, which implicitly learn similarity scores proportional to log density ratios. However, prior work has largely focused on 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":"2506.22881","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-06-28T13:36:44Z","cross_cats_sorted":[],"title_canon_sha256":"40a8f7860e70a8e241e20ba8e6f08a778df1c4e9498dccadc4b3b825077c838e","abstract_canon_sha256":"c5df4ed14fd84f8b9434508767e9f1329d50b1cbeac4fb3a239c4be8485eef16"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:06.865272Z","signature_b64":"ld6Qxq0wm3ije8f3T6Vo5H4Xt2iyYXqK+qXy7T5LucPgi9Z9oKMGMIlsZkqBpl9cpDSt9VSalGB76ZlI3FotBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"36979f5423e803fd776d8f5e3efd63d67961845a7b7a0ced67fa568b3e581770","last_reissued_at":"2026-06-02T02:04:06.864760Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:06.864760Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CLIP-like Model as a Foundational Density Ratio Estimator","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Fumiya Uchiyama, Hirokatsu Kataoka, Masahiro Suzuki, Rintaro Yanagi, Shohei Taniguchi, Shota Takashiro, Yusuke Iwasawa, Yutaka Matsuo","submitted_at":"2025-06-28T13:36:44Z","abstract_excerpt":"Density ratio estimation is a core concept in statistical machine learning because it provides a unified mechanism for tasks such as importance weighting, divergence estimation, and likelihood-free inference, but its potential in vision and language models has not been fully explored. Modern vision-language encoders such as CLIP and SigLIP are trained with contrastive objectives that implicitly optimize log density ratios between joint and marginal image-text distributions, which implicitly learn similarity scores proportional to log density ratios. However, prior work has largely focused on t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.22881","kind":"arxiv","version":3},"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/2506.22881/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":"2506.22881","created_at":"2026-06-02T02:04:06.864836+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.22881v3","created_at":"2026-06-02T02:04:06.864836+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.22881","created_at":"2026-06-02T02:04:06.864836+00:00"},{"alias_kind":"pith_short_12","alias_value":"G2LZ6VBD5AB7","created_at":"2026-06-02T02:04:06.864836+00:00"},{"alias_kind":"pith_short_16","alias_value":"G2LZ6VBD5AB7253N","created_at":"2026-06-02T02:04:06.864836+00:00"},{"alias_kind":"pith_short_8","alias_value":"G2LZ6VBD","created_at":"2026-06-02T02:04:06.864836+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/G2LZ6VBD5AB7253NR5PD57LD2Z","json":"https://pith.science/pith/G2LZ6VBD5AB7253NR5PD57LD2Z.json","graph_json":"https://pith.science/api/pith-number/G2LZ6VBD5AB7253NR5PD57LD2Z/graph.json","events_json":"https://pith.science/api/pith-number/G2LZ6VBD5AB7253NR5PD57LD2Z/events.json","paper":"https://pith.science/paper/G2LZ6VBD"},"agent_actions":{"view_html":"https://pith.science/pith/G2LZ6VBD5AB7253NR5PD57LD2Z","download_json":"https://pith.science/pith/G2LZ6VBD5AB7253NR5PD57LD2Z.json","view_paper":"https://pith.science/paper/G2LZ6VBD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.22881&json=true","fetch_graph":"https://pith.science/api/pith-number/G2LZ6VBD5AB7253NR5PD57LD2Z/graph.json","fetch_events":"https://pith.science/api/pith-number/G2LZ6VBD5AB7253NR5PD57LD2Z/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G2LZ6VBD5AB7253NR5PD57LD2Z/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G2LZ6VBD5AB7253NR5PD57LD2Z/action/storage_attestation","attest_author":"https://pith.science/pith/G2LZ6VBD5AB7253NR5PD57LD2Z/action/author_attestation","sign_citation":"https://pith.science/pith/G2LZ6VBD5AB7253NR5PD57LD2Z/action/citation_signature","submit_replication":"https://pith.science/pith/G2LZ6VBD5AB7253NR5PD57LD2Z/action/replication_record"}},"created_at":"2026-06-02T02:04:06.864836+00:00","updated_at":"2026-06-02T02:04:06.864836+00:00"}