{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:TIAI4T4MHMLJZRYHLUECXG7RV3","short_pith_number":"pith:TIAI4T4M","schema_version":"1.0","canonical_sha256":"9a008e4f8c3b169cc7075d082b9bf1aef252f65e2b4b221c9a82d9b4f3ccea04","source":{"kind":"arxiv","id":"2406.16232","version":3},"attestation_state":"computed","paper":{"title":"Jacobian Descent for Multi-Objective Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","math.OC"],"primary_cat":"cs.LG","authors_text":"Pierre Quinton, Val\\'erian Rey","submitted_at":"2024-06-23T22:06:25Z","abstract_excerpt":"Many optimization problems require balancing multiple conflicting objectives. As gradient descent is limited to single-objective optimization, we introduce its direct generalization: Jacobian descent (JD). This algorithm iteratively updates parameters using the Jacobian matrix of a vector-valued objective function, in which each row is the gradient of an individual objective. While several methods to combine gradients already exist in the literature, they are generally hindered when the objectives conflict. In contrast, we propose projecting gradients to fully resolve conflict while ensuring 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":"2406.16232","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-23T22:06:25Z","cross_cats_sorted":["cs.AI","math.OC"],"title_canon_sha256":"6f8096ba2b1d92341f745d995ebf108698d57a8d8c6e145fa23926f520b77e2c","abstract_canon_sha256":"5865de3fd2a0ee3ee8ff603fd2965a286ca109292991799176e18b0d5ca9642f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:08:18.691070Z","signature_b64":"SaoMcNulJ7U8XBJLUC3RcoAK6shNYERUnYnuoA1VumxtMMyqzxPkE+vX61DF1XIT0Jr7hlHLkl6bMtTl0RzCDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9a008e4f8c3b169cc7075d082b9bf1aef252f65e2b4b221c9a82d9b4f3ccea04","last_reissued_at":"2026-07-05T10:08:18.690621Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:08:18.690621Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Jacobian Descent for Multi-Objective Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","math.OC"],"primary_cat":"cs.LG","authors_text":"Pierre Quinton, Val\\'erian Rey","submitted_at":"2024-06-23T22:06:25Z","abstract_excerpt":"Many optimization problems require balancing multiple conflicting objectives. As gradient descent is limited to single-objective optimization, we introduce its direct generalization: Jacobian descent (JD). This algorithm iteratively updates parameters using the Jacobian matrix of a vector-valued objective function, in which each row is the gradient of an individual objective. While several methods to combine gradients already exist in the literature, they are generally hindered when the objectives conflict. In contrast, we propose projecting gradients to fully resolve conflict while ensuring t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.16232","kind":"arxiv","version":3},"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/2406.16232/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":"2406.16232","created_at":"2026-07-05T10:08:18.690680+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.16232v3","created_at":"2026-07-05T10:08:18.690680+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.16232","created_at":"2026-07-05T10:08:18.690680+00:00"},{"alias_kind":"pith_short_12","alias_value":"TIAI4T4MHMLJ","created_at":"2026-07-05T10:08:18.690680+00:00"},{"alias_kind":"pith_short_16","alias_value":"TIAI4T4MHMLJZRYH","created_at":"2026-07-05T10:08:18.690680+00:00"},{"alias_kind":"pith_short_8","alias_value":"TIAI4T4M","created_at":"2026-07-05T10:08:18.690680+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.20160","citing_title":"Multi-objective design of photon blockade for bright single-photon sources","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27734","citing_title":"Learn from your own latents and not from tokens: A sample-complexity theory","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30452","citing_title":"A Unified Framework for Gradient Aggregation in Multi-Objective Optimization","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00399","citing_title":"Multi-Objective Reference-Aligned Machine Unlearning","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08939","citing_title":"Delve into the Applicability of Advanced Optimizers for Multi-Task Learning","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07756","citing_title":"When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05734","citing_title":"Accessing the performance of CC2 for excited state dynamics: a benchmark study with pyrazine","ref_index":85,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3","json":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3.json","graph_json":"https://pith.science/api/pith-number/TIAI4T4MHMLJZRYHLUECXG7RV3/graph.json","events_json":"https://pith.science/api/pith-number/TIAI4T4MHMLJZRYHLUECXG7RV3/events.json","paper":"https://pith.science/paper/TIAI4T4M"},"agent_actions":{"view_html":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3","download_json":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3.json","view_paper":"https://pith.science/paper/TIAI4T4M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.16232&json=true","fetch_graph":"https://pith.science/api/pith-number/TIAI4T4MHMLJZRYHLUECXG7RV3/graph.json","fetch_events":"https://pith.science/api/pith-number/TIAI4T4MHMLJZRYHLUECXG7RV3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3/action/storage_attestation","attest_author":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3/action/author_attestation","sign_citation":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3/action/citation_signature","submit_replication":"https://pith.science/pith/TIAI4T4MHMLJZRYHLUECXG7RV3/action/replication_record"}},"created_at":"2026-07-05T10:08:18.690680+00:00","updated_at":"2026-07-05T10:08:18.690680+00:00"}