{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:ZZN5QLSHGFD2VQZV6SKPQMZKLK","short_pith_number":"pith:ZZN5QLSH","schema_version":"1.0","canonical_sha256":"ce5bd82e473147aac335f494f8332a5a874db2749ede3422cb6048c29b18d693","source":{"kind":"arxiv","id":"2311.17921","version":3},"attestation_state":"computed","paper":{"title":"Do text-free diffusion models learn discriminative visual representations?","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Abhinav Shrivastava, Archana Swaminathan, Jun Ohya, Matthew Gwilliam, Namitha Padmanabhan, Soumik Mukhopadhyay, Tianyi Zhou, Vatsal Agarwal, Yosuke Yamaguchi","submitted_at":"2023-11-29T18:59:59Z","abstract_excerpt":"While many unsupervised learning models focus on one family of tasks, either generative or discriminative, we explore the possibility of a unified representation learner: a model which addresses both families of tasks simultaneously. We identify diffusion models, a state-of-the-art method for generative tasks, as a prime candidate. Such models involve training a U-Net to iteratively predict and remove noise, and the resulting model can synthesize high-fidelity, diverse, novel images. We find that the intermediate feature maps of the U-Net are diverse, discriminative feature representations. We"},"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":"2311.17921","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2023-11-29T18:59:59Z","cross_cats_sorted":[],"title_canon_sha256":"bf132c095abc0d9b41b8a238fa68fd9db3df95d58ffb0013db72e8c6c1a42ea2","abstract_canon_sha256":"01fcfdb9ea5d48d299eace76e1e4e4c861a5267caf94cbf89007d441b9e29828"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:11:12.404946Z","signature_b64":"WVQ9SHrRvAu8lsMwxSWn6eM0mO8Z9t+TtrBNQRLRTk67nviOPXaH+UnORJu+VrClCn6xGAAuecc00KDByyE1CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ce5bd82e473147aac335f494f8332a5a874db2749ede3422cb6048c29b18d693","last_reissued_at":"2026-07-05T09:11:12.404383Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:11:12.404383Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do text-free diffusion models learn discriminative visual representations?","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Abhinav Shrivastava, Archana Swaminathan, Jun Ohya, Matthew Gwilliam, Namitha Padmanabhan, Soumik Mukhopadhyay, Tianyi Zhou, Vatsal Agarwal, Yosuke Yamaguchi","submitted_at":"2023-11-29T18:59:59Z","abstract_excerpt":"While many unsupervised learning models focus on one family of tasks, either generative or discriminative, we explore the possibility of a unified representation learner: a model which addresses both families of tasks simultaneously. We identify diffusion models, a state-of-the-art method for generative tasks, as a prime candidate. Such models involve training a U-Net to iteratively predict and remove noise, and the resulting model can synthesize high-fidelity, diverse, novel images. We find that the intermediate feature maps of the U-Net are diverse, discriminative feature representations. We"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.17921","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/2311.17921/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":"2311.17921","created_at":"2026-07-05T09:11:12.404449+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.17921v3","created_at":"2026-07-05T09:11:12.404449+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.17921","created_at":"2026-07-05T09:11:12.404449+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZZN5QLSHGFD2","created_at":"2026-07-05T09:11:12.404449+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZZN5QLSHGFD2VQZV","created_at":"2026-07-05T09:11:12.404449+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZZN5QLSH","created_at":"2026-07-05T09:11:12.404449+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07693","citing_title":"Selective Timestep Weighting and Advantage-Based Replay for Sample-Efficient Diffusion RLHF","ref_index":25,"is_internal_anchor":true},{"citing_arxiv_id":"2605.25012","citing_title":"Learning from Semantic Dictionaries: Discriminative Codebook Contrastive Learning for Unified Visual Representation and Generation","ref_index":38,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK","json":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK.json","graph_json":"https://pith.science/api/pith-number/ZZN5QLSHGFD2VQZV6SKPQMZKLK/graph.json","events_json":"https://pith.science/api/pith-number/ZZN5QLSHGFD2VQZV6SKPQMZKLK/events.json","paper":"https://pith.science/paper/ZZN5QLSH"},"agent_actions":{"view_html":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK","download_json":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK.json","view_paper":"https://pith.science/paper/ZZN5QLSH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.17921&json=true","fetch_graph":"https://pith.science/api/pith-number/ZZN5QLSHGFD2VQZV6SKPQMZKLK/graph.json","fetch_events":"https://pith.science/api/pith-number/ZZN5QLSHGFD2VQZV6SKPQMZKLK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK/action/storage_attestation","attest_author":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK/action/author_attestation","sign_citation":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK/action/citation_signature","submit_replication":"https://pith.science/pith/ZZN5QLSHGFD2VQZV6SKPQMZKLK/action/replication_record"}},"created_at":"2026-07-05T09:11:12.404449+00:00","updated_at":"2026-07-05T09:11:12.404449+00:00"}