{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:OUZQKAKT632EFST5JBLYSUDRFY","short_pith_number":"pith:OUZQKAKT","schema_version":"1.0","canonical_sha256":"7533050153f6f442ca7d48578950712e268e44bb3702269aacf91b8fccb51668","source":{"kind":"arxiv","id":"2302.09178","version":2},"attestation_state":"computed","paper":{"title":"Improving Training Stability for Multitask Ranking Models in Recommender Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Daryl Chang, Ed H. Chi, Jiaxi Tang, Justin Gilmer, Lichan Hong, Li Wei, Maheswaran Sathiamoorthy, Xinyang Yi, Yoel Drori","submitted_at":"2023-02-17T23:04:56Z","abstract_excerpt":"Recommender systems play an important role in many content platforms. While most recommendation research is dedicated to designing better models to improve user experience, we found that research on stabilizing the training for such models is severely under-explored. As recommendation models become larger and more sophisticated, they are more susceptible to training instability issues, i.e., loss divergence, which can make the model unusable, waste significant resources and block model developments. In this paper, we share our findings and best practices we learned for improving the training s"},"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":"2302.09178","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-02-17T23:04:56Z","cross_cats_sorted":[],"title_canon_sha256":"57828b13755a9cb711d741eb076bf86d89e8baac54497c79e2f17f7cfbef323c","abstract_canon_sha256":"6ba26058872f4387aa9d9c897edcfdf54df3ff858e22e1f7b11e1826bf16e79b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:21:05.602365Z","signature_b64":"mmNLpHdbqqOMXppVVSp4XwJnowdd7jW0T2T9Mr5TFpfXsozGiLTL1vC2isEWnXsqm3vCUHO9okaNxjcx7fAqDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7533050153f6f442ca7d48578950712e268e44bb3702269aacf91b8fccb51668","last_reissued_at":"2026-07-05T06:21:05.601844Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:21:05.601844Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Training Stability for Multitask Ranking Models in Recommender Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Daryl Chang, Ed H. Chi, Jiaxi Tang, Justin Gilmer, Lichan Hong, Li Wei, Maheswaran Sathiamoorthy, Xinyang Yi, Yoel Drori","submitted_at":"2023-02-17T23:04:56Z","abstract_excerpt":"Recommender systems play an important role in many content platforms. While most recommendation research is dedicated to designing better models to improve user experience, we found that research on stabilizing the training for such models is severely under-explored. As recommendation models become larger and more sophisticated, they are more susceptible to training instability issues, i.e., loss divergence, which can make the model unusable, waste significant resources and block model developments. In this paper, we share our findings and best practices we learned for improving the training s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.09178","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/2302.09178/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":"2302.09178","created_at":"2026-07-05T06:21:05.601923+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.09178v2","created_at":"2026-07-05T06:21:05.601923+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.09178","created_at":"2026-07-05T06:21:05.601923+00:00"},{"alias_kind":"pith_short_12","alias_value":"OUZQKAKT632E","created_at":"2026-07-05T06:21:05.601923+00:00"},{"alias_kind":"pith_short_16","alias_value":"OUZQKAKT632EFST5","created_at":"2026-07-05T06:21:05.601923+00:00"},{"alias_kind":"pith_short_8","alias_value":"OUZQKAKT","created_at":"2026-07-05T06:21:05.601923+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.13336","citing_title":"SGCL: Unifying Self-Supervised and Supervised Learning for Graph Recommendation","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY","json":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY.json","graph_json":"https://pith.science/api/pith-number/OUZQKAKT632EFST5JBLYSUDRFY/graph.json","events_json":"https://pith.science/api/pith-number/OUZQKAKT632EFST5JBLYSUDRFY/events.json","paper":"https://pith.science/paper/OUZQKAKT"},"agent_actions":{"view_html":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY","download_json":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY.json","view_paper":"https://pith.science/paper/OUZQKAKT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.09178&json=true","fetch_graph":"https://pith.science/api/pith-number/OUZQKAKT632EFST5JBLYSUDRFY/graph.json","fetch_events":"https://pith.science/api/pith-number/OUZQKAKT632EFST5JBLYSUDRFY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY/action/storage_attestation","attest_author":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY/action/author_attestation","sign_citation":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY/action/citation_signature","submit_replication":"https://pith.science/pith/OUZQKAKT632EFST5JBLYSUDRFY/action/replication_record"}},"created_at":"2026-07-05T06:21:05.601923+00:00","updated_at":"2026-07-05T06:21:05.601923+00:00"}