{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:NYGVGRJEA7DJRJ5TO3DBSROCZL","short_pith_number":"pith:NYGVGRJE","schema_version":"1.0","canonical_sha256":"6e0d53452407c698a7b376c61945c2cac00932edc35367cb9c4cd529d07cb6ac","source":{"kind":"arxiv","id":"2305.15316","version":1},"attestation_state":"computed","paper":{"title":"Training on Thin Air: Improve Image Classification with Generated Data","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Hshmat Sahak, Jimmy Ba, Yongchao Zhou","submitted_at":"2023-05-24T16:33:02Z","abstract_excerpt":"Acquiring high-quality data for training discriminative models is a crucial yet challenging aspect of building effective predictive systems. In this paper, we present Diffusion Inversion, a simple yet effective method that leverages the pre-trained generative model, Stable Diffusion, to generate diverse, high-quality training data for image classification. Our approach captures the original data distribution and ensures data coverage by inverting images to the latent space of Stable Diffusion, and generates diverse novel training images by conditioning the generative model on noisy versions of"},"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":"2305.15316","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-05-24T16:33:02Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"36e8d7e9571f9a3e11fe0efd14c26b063a2620763c170c5d1ca95ae927865e55","abstract_canon_sha256":"8fa236076c17f0905d1478037c95b6d8fb41d669fc75d2557dd7e34157f33b12"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:13:33.755289Z","signature_b64":"CdYDlWSEK5yxteGiVNQDZ7M9s0JsXt/M2HA3L4Vgpn/qBv/WUsEW+z2/13AEM8X1QSmf0ijxGaTbvejrtjpKDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6e0d53452407c698a7b376c61945c2cac00932edc35367cb9c4cd529d07cb6ac","last_reissued_at":"2026-07-05T06:13:33.754873Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:13:33.754873Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Training on Thin Air: Improve Image Classification with Generated Data","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Hshmat Sahak, Jimmy Ba, Yongchao Zhou","submitted_at":"2023-05-24T16:33:02Z","abstract_excerpt":"Acquiring high-quality data for training discriminative models is a crucial yet challenging aspect of building effective predictive systems. In this paper, we present Diffusion Inversion, a simple yet effective method that leverages the pre-trained generative model, Stable Diffusion, to generate diverse, high-quality training data for image classification. Our approach captures the original data distribution and ensures data coverage by inverting images to the latent space of Stable Diffusion, and generates diverse novel training images by conditioning the generative model on noisy versions of"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.15316","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/2305.15316/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":"2305.15316","created_at":"2026-07-05T06:13:33.754923+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.15316v1","created_at":"2026-07-05T06:13:33.754923+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.15316","created_at":"2026-07-05T06:13:33.754923+00:00"},{"alias_kind":"pith_short_12","alias_value":"NYGVGRJEA7DJ","created_at":"2026-07-05T06:13:33.754923+00:00"},{"alias_kind":"pith_short_16","alias_value":"NYGVGRJEA7DJRJ5T","created_at":"2026-07-05T06:13:33.754923+00:00"},{"alias_kind":"pith_short_8","alias_value":"NYGVGRJE","created_at":"2026-07-05T06:13:33.754923+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.27343","citing_title":"Towards Controllable Image Generation through Representation-Conditioned Diffusion Models","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL","json":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL.json","graph_json":"https://pith.science/api/pith-number/NYGVGRJEA7DJRJ5TO3DBSROCZL/graph.json","events_json":"https://pith.science/api/pith-number/NYGVGRJEA7DJRJ5TO3DBSROCZL/events.json","paper":"https://pith.science/paper/NYGVGRJE"},"agent_actions":{"view_html":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL","download_json":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL.json","view_paper":"https://pith.science/paper/NYGVGRJE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.15316&json=true","fetch_graph":"https://pith.science/api/pith-number/NYGVGRJEA7DJRJ5TO3DBSROCZL/graph.json","fetch_events":"https://pith.science/api/pith-number/NYGVGRJEA7DJRJ5TO3DBSROCZL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL/action/storage_attestation","attest_author":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL/action/author_attestation","sign_citation":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL/action/citation_signature","submit_replication":"https://pith.science/pith/NYGVGRJEA7DJRJ5TO3DBSROCZL/action/replication_record"}},"created_at":"2026-07-05T06:13:33.754923+00:00","updated_at":"2026-07-05T06:13:33.754923+00:00"}