{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Y3MPBC5WRAP22QXBTQSXORWZLY","short_pith_number":"pith:Y3MPBC5W","schema_version":"1.0","canonical_sha256":"c6d8f08bb6881fad42e19c257746d95e08a4d9c57814303fca6b7bd3f0dac69b","source":{"kind":"arxiv","id":"2508.05700","version":2},"attestation_state":"computed","paper":{"title":"Multi-Faceted Large Embedding Tables for Pinterest Ads Ranking","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.IR","authors_text":"Degao Peng, Guangtong Bai, Han Sun, Haoyang Li, Huasen Wu, Jiacheng Li, Jiayin Jin, Jinfeng Zhuang, Kungang Li, Ling Leng, Li Tang, Prathibha Deshikachar, Runze Su, Sihan Wang, Siping Ji, Yixiong Meng, Zelun Wang, Zhifang Liu, Zhimeng Pan","submitted_at":"2025-08-07T00:31:20Z","abstract_excerpt":"Large embedding tables are indispensable in modern recommendation systems, thanks to their ability to effectively capture and memorize intricate details of interactions among diverse entities. As we explore integrating large embedding tables into Pinterest's ads ranking models, we encountered not only common challenges such as sparsity and scalability, but also several obstacles unique to our context. Notably, our initial attempts to train large embedding tables from scratch resulted in neutral metrics. To tackle this, we introduced a novel multi-faceted pretraining scheme that incorporates mu"},"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":"2508.05700","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2025-08-07T00:31:20Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"72c8086b42b2e4e64090001c73e7ce5d4403a48ce835498fb2fd8be4545b5ea0","abstract_canon_sha256":"f249554f9c4906c3a9e9637b4549b363fa73c54739fb055665626537d1202539"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:23.694481Z","signature_b64":"SQQ25HP3booVoV7D3WF7B2jIMdsJBelZeLZPezBLmXZLsjYA9mBrO2+ECDghU+SCDCc1kin5LeeTVtePLJq7Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c6d8f08bb6881fad42e19c257746d95e08a4d9c57814303fca6b7bd3f0dac69b","last_reissued_at":"2026-07-05T11:52:23.693901Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:23.693901Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multi-Faceted Large Embedding Tables for Pinterest Ads Ranking","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.IR","authors_text":"Degao Peng, Guangtong Bai, Han Sun, Haoyang Li, Huasen Wu, Jiacheng Li, Jiayin Jin, Jinfeng Zhuang, Kungang Li, Ling Leng, Li Tang, Prathibha Deshikachar, Runze Su, Sihan Wang, Siping Ji, Yixiong Meng, Zelun Wang, Zhifang Liu, Zhimeng Pan","submitted_at":"2025-08-07T00:31:20Z","abstract_excerpt":"Large embedding tables are indispensable in modern recommendation systems, thanks to their ability to effectively capture and memorize intricate details of interactions among diverse entities. As we explore integrating large embedding tables into Pinterest's ads ranking models, we encountered not only common challenges such as sparsity and scalability, but also several obstacles unique to our context. Notably, our initial attempts to train large embedding tables from scratch resulted in neutral metrics. To tackle this, we introduced a novel multi-faceted pretraining scheme that incorporates mu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.05700","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/2508.05700/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":"2508.05700","created_at":"2026-07-05T11:52:23.693964+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.05700v2","created_at":"2026-07-05T11:52:23.693964+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.05700","created_at":"2026-07-05T11:52:23.693964+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y3MPBC5WRAP2","created_at":"2026-07-05T11:52:23.693964+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y3MPBC5WRAP22QXB","created_at":"2026-07-05T11:52:23.693964+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y3MPBC5W","created_at":"2026-07-05T11:52:23.693964+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10243","citing_title":"DUET -- Dual User Embedding Transformers for Offsite Conversion Prediction","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY","json":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY.json","graph_json":"https://pith.science/api/pith-number/Y3MPBC5WRAP22QXBTQSXORWZLY/graph.json","events_json":"https://pith.science/api/pith-number/Y3MPBC5WRAP22QXBTQSXORWZLY/events.json","paper":"https://pith.science/paper/Y3MPBC5W"},"agent_actions":{"view_html":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY","download_json":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY.json","view_paper":"https://pith.science/paper/Y3MPBC5W","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.05700&json=true","fetch_graph":"https://pith.science/api/pith-number/Y3MPBC5WRAP22QXBTQSXORWZLY/graph.json","fetch_events":"https://pith.science/api/pith-number/Y3MPBC5WRAP22QXBTQSXORWZLY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY/action/storage_attestation","attest_author":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY/action/author_attestation","sign_citation":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY/action/citation_signature","submit_replication":"https://pith.science/pith/Y3MPBC5WRAP22QXBTQSXORWZLY/action/replication_record"}},"created_at":"2026-07-05T11:52:23.693964+00:00","updated_at":"2026-07-05T11:52:23.693964+00:00"}