{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:FK2R5IPIPDWLHESLTTLDTQZDHV","short_pith_number":"pith:FK2R5IPI","schema_version":"1.0","canonical_sha256":"2ab51ea1e878ecb3924b9cd639c3233d755b50a68a150ddc13c6cac5f9444ab1","source":{"kind":"arxiv","id":"2306.08121","version":2},"attestation_state":"computed","paper":{"title":"Better Generalization with Semantic IDs: A Case Study in Ranking for Recommendations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.IR","authors_text":"Anima Singh, Devansh Tandon, Ed H. Chi, Lichan Hong, Li Wei, Lukasz Heldt, Maheswaran Sathiamoorthy, Nikhil Mehta, Raghunandan Keshavan, Trung Vu, Xinyang Yi, Yilin Zheng","submitted_at":"2023-06-13T20:34:15Z","abstract_excerpt":"Randomly-hashed item ids are used ubiquitously in recommendation models. However, the learned representations from random hashing prevents generalization across similar items, causing problems of learning unseen and long-tail items, especially when item corpus is large, power-law distributed, and evolving dynamically. In this paper, we propose using content-derived features as a replacement for random ids. We show that simply replacing ID features with content-based embeddings can cause a drop in quality due to reduced memorization capability. To strike a good balance of memorization and gener"},"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":"2306.08121","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2023-06-13T20:34:15Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f839a51203d43907212d76043988fd564fc2cf3008b8d1924018262ad237d0ae","abstract_canon_sha256":"902c7711b1d235605e9de88b69db9b7eceb4f6e6294f7150719691bd8cc426b4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:24:56.052833Z","signature_b64":"XDixFC9ibyTDwZb77w/3L2kMbrAE4Ftlc5hbUV4J1i7Jc8y4O4z3OF1vdJomK/ZirH61aPQfTi7yljGkJJr5DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2ab51ea1e878ecb3924b9cd639c3233d755b50a68a150ddc13c6cac5f9444ab1","last_reissued_at":"2026-07-05T08:24:56.052297Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:24:56.052297Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Better Generalization with Semantic IDs: A Case Study in Ranking for Recommendations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.IR","authors_text":"Anima Singh, Devansh Tandon, Ed H. Chi, Lichan Hong, Li Wei, Lukasz Heldt, Maheswaran Sathiamoorthy, Nikhil Mehta, Raghunandan Keshavan, Trung Vu, Xinyang Yi, Yilin Zheng","submitted_at":"2023-06-13T20:34:15Z","abstract_excerpt":"Randomly-hashed item ids are used ubiquitously in recommendation models. However, the learned representations from random hashing prevents generalization across similar items, causing problems of learning unseen and long-tail items, especially when item corpus is large, power-law distributed, and evolving dynamically. In this paper, we propose using content-derived features as a replacement for random ids. We show that simply replacing ID features with content-based embeddings can cause a drop in quality due to reduced memorization capability. To strike a good balance of memorization and gener"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.08121","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/2306.08121/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":"2306.08121","created_at":"2026-07-05T08:24:56.052376+00:00"},{"alias_kind":"arxiv_version","alias_value":"2306.08121v2","created_at":"2026-07-05T08:24:56.052376+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.08121","created_at":"2026-07-05T08:24:56.052376+00:00"},{"alias_kind":"pith_short_12","alias_value":"FK2R5IPIPDWL","created_at":"2026-07-05T08:24:56.052376+00:00"},{"alias_kind":"pith_short_16","alias_value":"FK2R5IPIPDWLHESL","created_at":"2026-07-05T08:24:56.052376+00:00"},{"alias_kind":"pith_short_8","alias_value":"FK2R5IPI","created_at":"2026-07-05T08:24:56.052376+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10375","citing_title":"SIDInspector: A Mapping-First Diagnostic Resource for Semantic-ID Tokenizers","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01396","citing_title":"Quantizing Intent: Cross-Domain Semantic IDs from Organic Activity for Industrial Ranking","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21832","citing_title":"FLUID: From Ephemeral IDs to Multimodal Semantic Codes for Industrial-Scale Livestreaming Recommendation","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV","json":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV.json","graph_json":"https://pith.science/api/pith-number/FK2R5IPIPDWLHESLTTLDTQZDHV/graph.json","events_json":"https://pith.science/api/pith-number/FK2R5IPIPDWLHESLTTLDTQZDHV/events.json","paper":"https://pith.science/paper/FK2R5IPI"},"agent_actions":{"view_html":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV","download_json":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV.json","view_paper":"https://pith.science/paper/FK2R5IPI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2306.08121&json=true","fetch_graph":"https://pith.science/api/pith-number/FK2R5IPIPDWLHESLTTLDTQZDHV/graph.json","fetch_events":"https://pith.science/api/pith-number/FK2R5IPIPDWLHESLTTLDTQZDHV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV/action/storage_attestation","attest_author":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV/action/author_attestation","sign_citation":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV/action/citation_signature","submit_replication":"https://pith.science/pith/FK2R5IPIPDWLHESLTTLDTQZDHV/action/replication_record"}},"created_at":"2026-07-05T08:24:56.052376+00:00","updated_at":"2026-07-05T08:24:56.052376+00:00"}