{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:NB3G7NLCQEENVEU2IXELOVWSNW","short_pith_number":"pith:NB3G7NLC","schema_version":"1.0","canonical_sha256":"68766fb5628108da929a45c8b756d26d9d7877c9a4cea3111cef6f16e487f99e","source":{"kind":"arxiv","id":"2111.10603","version":2},"attestation_state":"computed","paper":{"title":"Reasonable Effectiveness of Random Weighting: A Litmus Test for Multi-Task Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Baijiong Lin, Feiyang Ye, Ivor W. Tsang, Yu Zhang","submitted_at":"2021-11-20T14:28:32Z","abstract_excerpt":"Multi-Task Learning (MTL) has achieved success in various fields. However, how to balance different tasks to achieve good performance is a key problem. To achieve the task balancing, there are many works to carefully design dynamical loss/gradient weighting strategies but the basic random experiments are ignored to examine their effectiveness. In this paper, we propose the Random Weighting (RW) methods, including Random Loss Weighting (RLW) and Random Gradient Weighting (RGW), where an MTL model is trained with random loss/gradient weights sampled from a distribution. To show the effectiveness"},"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":"2111.10603","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-11-20T14:28:32Z","cross_cats_sorted":[],"title_canon_sha256":"d84f8826e37729f33fd3a0a7736fbb962691ce6ba581abe0efb3aefe39e75613","abstract_canon_sha256":"e042dfca00a9ac0051390bfbc204eb2cdc523ee194301db437b382c04464ae61"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:43:58.328090Z","signature_b64":"F/eOZoNJuj2BuEtpQvUd9sG/LBspDiLPlSb9aSgNT4bQo8jmdE9OEjwTpZ0Qkk62DCXRJXOXekTSQCk6AHu4BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"68766fb5628108da929a45c8b756d26d9d7877c9a4cea3111cef6f16e487f99e","last_reissued_at":"2026-07-05T04:43:58.327536Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:43:58.327536Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reasonable Effectiveness of Random Weighting: A Litmus Test for Multi-Task Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Baijiong Lin, Feiyang Ye, Ivor W. Tsang, Yu Zhang","submitted_at":"2021-11-20T14:28:32Z","abstract_excerpt":"Multi-Task Learning (MTL) has achieved success in various fields. However, how to balance different tasks to achieve good performance is a key problem. To achieve the task balancing, there are many works to carefully design dynamical loss/gradient weighting strategies but the basic random experiments are ignored to examine their effectiveness. In this paper, we propose the Random Weighting (RW) methods, including Random Loss Weighting (RLW) and Random Gradient Weighting (RGW), where an MTL model is trained with random loss/gradient weights sampled from a distribution. To show the effectiveness"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2111.10603","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/2111.10603/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":"2111.10603","created_at":"2026-07-05T04:43:58.327603+00:00"},{"alias_kind":"arxiv_version","alias_value":"2111.10603v2","created_at":"2026-07-05T04:43:58.327603+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2111.10603","created_at":"2026-07-05T04:43:58.327603+00:00"},{"alias_kind":"pith_short_12","alias_value":"NB3G7NLCQEEN","created_at":"2026-07-05T04:43:58.327603+00:00"},{"alias_kind":"pith_short_16","alias_value":"NB3G7NLCQEENVEU2","created_at":"2026-07-05T04:43:58.327603+00:00"},{"alias_kind":"pith_short_8","alias_value":"NB3G7NLC","created_at":"2026-07-05T04:43:58.327603+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.00078","citing_title":"Exploring Line Bundle Standard Models with Transformers","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31201","citing_title":"ExPLoRe: Expert Patch-Level Loss Routing for Multi-Objective Masked Image Modeling","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08939","citing_title":"Delve into the Applicability of Advanced Optimizers for Multi-Task Learning","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07914","citing_title":"Flatness and Gradient Alignment Are Both Necessary: Spectral-Aware Gradient-Aligned Exploration for Multi-Distribution Learning","ref_index":66,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW","json":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW.json","graph_json":"https://pith.science/api/pith-number/NB3G7NLCQEENVEU2IXELOVWSNW/graph.json","events_json":"https://pith.science/api/pith-number/NB3G7NLCQEENVEU2IXELOVWSNW/events.json","paper":"https://pith.science/paper/NB3G7NLC"},"agent_actions":{"view_html":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW","download_json":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW.json","view_paper":"https://pith.science/paper/NB3G7NLC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2111.10603&json=true","fetch_graph":"https://pith.science/api/pith-number/NB3G7NLCQEENVEU2IXELOVWSNW/graph.json","fetch_events":"https://pith.science/api/pith-number/NB3G7NLCQEENVEU2IXELOVWSNW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW/action/storage_attestation","attest_author":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW/action/author_attestation","sign_citation":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW/action/citation_signature","submit_replication":"https://pith.science/pith/NB3G7NLCQEENVEU2IXELOVWSNW/action/replication_record"}},"created_at":"2026-07-05T04:43:58.327603+00:00","updated_at":"2026-07-05T04:43:58.327603+00:00"}