{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:XRDDJ2XBJMRYLH6IRWOKIVLTKA","short_pith_number":"pith:XRDDJ2XB","schema_version":"1.0","canonical_sha256":"bc4634eae14b23859fc88d9ca4557350134906b001f7fe3e17b1be7bfa559738","source":{"kind":"arxiv","id":"2411.11871","version":1},"attestation_state":"computed","paper":{"title":"MultiBalance: Multi-Objective Gradient Balancing in Industrial-Scale Multi-Task Recommendation System","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"cs.IR","authors_text":"Bo Long, Jennifer Cao, Jiayi Xu, Liu Yang, Minhui Huang, Renqin Cai, Rong Jin, Sem Park, Xiaoyi Liu, Xue Feng, Xuxing Chen, Yiling You, Yiqun Liu, Yun He","submitted_at":"2024-11-03T21:20:21Z","abstract_excerpt":"In industrial recommendation systems, multi-task learning (learning multiple tasks simultaneously on a single model) is a predominant approach to save training/serving resources and improve recommendation performance via knowledge transfer between the joint learning tasks. However, multi-task learning often suffers from negative transfer: one or several tasks are less optimized than training them separately. To carefully balance the optimization, we propose a gradient balancing approach called MultiBalance, which is suitable for industrial-scale multi-task recommendation systems. It balances t"},"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":"2411.11871","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2024-11-03T21:20:21Z","cross_cats_sorted":["cs.LG","math.OC"],"title_canon_sha256":"625b1810efe275d24e8d24b288111e22a8809bf1cc8098ec354102e323c701a4","abstract_canon_sha256":"42944fb28db12eb4704cfe5a67ff840f47062f5de11d3ec2e326a0d4fa6efc7f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:37:12.830989Z","signature_b64":"DOmMVQNLlqABFn78L1nTcn0y15mnJPLATzrmp8F9e3X6gspD6lWCNm2xTW/7lKwtZwqaZO0YGxms0uvrfRRFAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bc4634eae14b23859fc88d9ca4557350134906b001f7fe3e17b1be7bfa559738","last_reissued_at":"2026-07-05T09:37:12.830531Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:37:12.830531Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MultiBalance: Multi-Objective Gradient Balancing in Industrial-Scale Multi-Task Recommendation System","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"cs.IR","authors_text":"Bo Long, Jennifer Cao, Jiayi Xu, Liu Yang, Minhui Huang, Renqin Cai, Rong Jin, Sem Park, Xiaoyi Liu, Xue Feng, Xuxing Chen, Yiling You, Yiqun Liu, Yun He","submitted_at":"2024-11-03T21:20:21Z","abstract_excerpt":"In industrial recommendation systems, multi-task learning (learning multiple tasks simultaneously on a single model) is a predominant approach to save training/serving resources and improve recommendation performance via knowledge transfer between the joint learning tasks. However, multi-task learning often suffers from negative transfer: one or several tasks are less optimized than training them separately. To carefully balance the optimization, we propose a gradient balancing approach called MultiBalance, which is suitable for industrial-scale multi-task recommendation systems. It balances t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.11871","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/2411.11871/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":"2411.11871","created_at":"2026-07-05T09:37:12.830589+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.11871v1","created_at":"2026-07-05T09:37:12.830589+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.11871","created_at":"2026-07-05T09:37:12.830589+00:00"},{"alias_kind":"pith_short_12","alias_value":"XRDDJ2XBJMRY","created_at":"2026-07-05T09:37:12.830589+00:00"},{"alias_kind":"pith_short_16","alias_value":"XRDDJ2XBJMRYLH6I","created_at":"2026-07-05T09:37:12.830589+00:00"},{"alias_kind":"pith_short_8","alias_value":"XRDDJ2XB","created_at":"2026-07-05T09:37:12.830589+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.20900","citing_title":"Embed Progressive Implicit Preference in Unified Space for Deep Collaborative Filtering","ref_index":20,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA","json":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA.json","graph_json":"https://pith.science/api/pith-number/XRDDJ2XBJMRYLH6IRWOKIVLTKA/graph.json","events_json":"https://pith.science/api/pith-number/XRDDJ2XBJMRYLH6IRWOKIVLTKA/events.json","paper":"https://pith.science/paper/XRDDJ2XB"},"agent_actions":{"view_html":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA","download_json":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA.json","view_paper":"https://pith.science/paper/XRDDJ2XB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.11871&json=true","fetch_graph":"https://pith.science/api/pith-number/XRDDJ2XBJMRYLH6IRWOKIVLTKA/graph.json","fetch_events":"https://pith.science/api/pith-number/XRDDJ2XBJMRYLH6IRWOKIVLTKA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA/action/storage_attestation","attest_author":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA/action/author_attestation","sign_citation":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA/action/citation_signature","submit_replication":"https://pith.science/pith/XRDDJ2XBJMRYLH6IRWOKIVLTKA/action/replication_record"}},"created_at":"2026-07-05T09:37:12.830589+00:00","updated_at":"2026-07-05T09:37:12.830589+00:00"}