{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CHFFWJAOYVGHZ42QYFDD3TEKE2","short_pith_number":"pith:CHFFWJAO","schema_version":"1.0","canonical_sha256":"11ca5b240ec54c7cf350c1463dcc8a26a01e8667301ce342cb9da42a9da77ec2","source":{"kind":"arxiv","id":"2504.00147","version":1},"attestation_state":"computed","paper":{"title":"Universal Zero-shot Embedding Inversion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.CL","authors_text":"Collin Zhang, John X. Morris, Vitaly Shmatikov","submitted_at":"2025-03-31T18:55:01Z","abstract_excerpt":"Embedding inversion, i.e., reconstructing text given its embedding and black-box access to the embedding encoder, is a fundamental problem in both NLP and security. From the NLP perspective, it helps determine how much semantic information about the input is retained in the embedding. From the security perspective, it measures how much information is leaked by vector databases and embedding-based retrieval systems. State-of-the-art methods for embedding inversion, such as vec2text, have high accuracy but require (a) training a separate model for each embedding, and (b) a large number of querie"},"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":"2504.00147","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-31T18:55:01Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"632f24c540df5a665bfd281ddcb75f797cb1f749b9fe25fbe3a9d72bf758b65f","abstract_canon_sha256":"056472168189396873dab7c3f909ce9e31f9162397eb575c3c6aeeaf70cd5bbc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:42:20.799145Z","signature_b64":"Rsny4U8d+x2tdiFl5ioW1qAIvY8h1uiwtMYmv1ZhbPU/Yc50Da+CQZn2KJsA6M788bMZnrBG6lbJrCbO6Jw7Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"11ca5b240ec54c7cf350c1463dcc8a26a01e8667301ce342cb9da42a9da77ec2","last_reissued_at":"2026-07-05T10:42:20.798661Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:42:20.798661Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Universal Zero-shot Embedding Inversion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.CL","authors_text":"Collin Zhang, John X. Morris, Vitaly Shmatikov","submitted_at":"2025-03-31T18:55:01Z","abstract_excerpt":"Embedding inversion, i.e., reconstructing text given its embedding and black-box access to the embedding encoder, is a fundamental problem in both NLP and security. From the NLP perspective, it helps determine how much semantic information about the input is retained in the embedding. From the security perspective, it measures how much information is leaked by vector databases and embedding-based retrieval systems. State-of-the-art methods for embedding inversion, such as vec2text, have high accuracy but require (a) training a separate model for each embedding, and (b) a large number of querie"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.00147","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/2504.00147/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":"2504.00147","created_at":"2026-07-05T10:42:20.798720+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.00147v1","created_at":"2026-07-05T10:42:20.798720+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.00147","created_at":"2026-07-05T10:42:20.798720+00:00"},{"alias_kind":"pith_short_12","alias_value":"CHFFWJAOYVGH","created_at":"2026-07-05T10:42:20.798720+00:00"},{"alias_kind":"pith_short_16","alias_value":"CHFFWJAOYVGHZ42Q","created_at":"2026-07-05T10:42:20.798720+00:00"},{"alias_kind":"pith_short_8","alias_value":"CHFFWJAO","created_at":"2026-07-05T10:42:20.798720+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26373","citing_title":"Hybrid privacy-aware semantic search: SVD-truncated document geometry and CKKS-encrypted query reranking under a restricted threat model","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01276","citing_title":"Embedding Inference Attack","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27976","citing_title":"SHARD: cell-keyed residual splitting for alignment-resistant private dense retrieval","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27976","citing_title":"SHARD: cell-keyed residual splitting for alignment-resistant private dense retrieval","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26373","citing_title":"Hybrid privacy-aware semantic search: SVD-truncated document geometry and CKKS-encrypted query reranking under a restricted threat model","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06831","citing_title":"Towards Privacy-Preserving Large Language Model: Text-free Inference Through Alignment and Adaptation","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05480","citing_title":"Can You Trust the Vectors in Your Vector Database? Black-Hole Attack from Embedding Space Defects","ref_index":63,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2","json":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2.json","graph_json":"https://pith.science/api/pith-number/CHFFWJAOYVGHZ42QYFDD3TEKE2/graph.json","events_json":"https://pith.science/api/pith-number/CHFFWJAOYVGHZ42QYFDD3TEKE2/events.json","paper":"https://pith.science/paper/CHFFWJAO"},"agent_actions":{"view_html":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2","download_json":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2.json","view_paper":"https://pith.science/paper/CHFFWJAO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.00147&json=true","fetch_graph":"https://pith.science/api/pith-number/CHFFWJAOYVGHZ42QYFDD3TEKE2/graph.json","fetch_events":"https://pith.science/api/pith-number/CHFFWJAOYVGHZ42QYFDD3TEKE2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2/action/storage_attestation","attest_author":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2/action/author_attestation","sign_citation":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2/action/citation_signature","submit_replication":"https://pith.science/pith/CHFFWJAOYVGHZ42QYFDD3TEKE2/action/replication_record"}},"created_at":"2026-07-05T10:42:20.798720+00:00","updated_at":"2026-07-05T10:42:20.798720+00:00"}