{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:DQ5GTTDCNVBVTKKFS6FIV257UP","short_pith_number":"pith:DQ5GTTDC","schema_version":"1.0","canonical_sha256":"1c3a69cc626d4359a945978a8aebbfa3d5f646c0fc137ed849d3029085e9eb50","source":{"kind":"arxiv","id":"2506.06130","version":1},"attestation_state":"computed","paper":{"title":"Gradient Similarity Surgery in Multi-Task Deep Learning","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Andrea Rosani, Giuseppe Di Fatta, Giuseppe Nicosia, Thomas Borsani","submitted_at":"2025-06-06T14:40:50Z","abstract_excerpt":"The multi-task learning ($MTL$) paradigm aims to simultaneously learn multiple tasks within a single model capturing higher-level, more general hidden patterns that are shared by the tasks. In deep learning, a significant challenge in the backpropagation training process is the design of advanced optimisers to improve the convergence speed and stability of the gradient descent learning rule. In particular, in multi-task deep learning ($MTDL$) the multitude of tasks may generate potentially conflicting gradients that would hinder the concurrent convergence of the diverse loss functions. This ch"},"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":"2506.06130","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-06T14:40:50Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"a2be743757ebaccbdf0a0f88917130c7ab50fb0cec76e5b2d425b07a4bdc6089","abstract_canon_sha256":"d6e67584102dffa3b3f6293dde5c921db6e1c6dc5702c209aa04c60681b57f24"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:17:26.200222Z","signature_b64":"Kpx3QkmiULIoGlgSkQs0GyylPe8HAN6jv67RH2r4YScCWtea6Ue1amo4txB9zLPszG4ejksxYS4JFOEaW6AtAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1c3a69cc626d4359a945978a8aebbfa3d5f646c0fc137ed849d3029085e9eb50","last_reissued_at":"2026-07-05T11:17:26.199696Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:17:26.199696Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Gradient Similarity Surgery in Multi-Task Deep Learning","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Andrea Rosani, Giuseppe Di Fatta, Giuseppe Nicosia, Thomas Borsani","submitted_at":"2025-06-06T14:40:50Z","abstract_excerpt":"The multi-task learning ($MTL$) paradigm aims to simultaneously learn multiple tasks within a single model capturing higher-level, more general hidden patterns that are shared by the tasks. In deep learning, a significant challenge in the backpropagation training process is the design of advanced optimisers to improve the convergence speed and stability of the gradient descent learning rule. In particular, in multi-task deep learning ($MTDL$) the multitude of tasks may generate potentially conflicting gradients that would hinder the concurrent convergence of the diverse loss functions. This ch"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.06130","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/2506.06130/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":"2506.06130","created_at":"2026-07-05T11:17:26.199757+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.06130v1","created_at":"2026-07-05T11:17:26.199757+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.06130","created_at":"2026-07-05T11:17:26.199757+00:00"},{"alias_kind":"pith_short_12","alias_value":"DQ5GTTDCNVBV","created_at":"2026-07-05T11:17:26.199757+00:00"},{"alias_kind":"pith_short_16","alias_value":"DQ5GTTDCNVBVTKKF","created_at":"2026-07-05T11:17:26.199757+00:00"},{"alias_kind":"pith_short_8","alias_value":"DQ5GTTDC","created_at":"2026-07-05T11:17:26.199757+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.00072","citing_title":"XekRung Technical Report","ref_index":132,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP","json":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP.json","graph_json":"https://pith.science/api/pith-number/DQ5GTTDCNVBVTKKFS6FIV257UP/graph.json","events_json":"https://pith.science/api/pith-number/DQ5GTTDCNVBVTKKFS6FIV257UP/events.json","paper":"https://pith.science/paper/DQ5GTTDC"},"agent_actions":{"view_html":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP","download_json":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP.json","view_paper":"https://pith.science/paper/DQ5GTTDC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.06130&json=true","fetch_graph":"https://pith.science/api/pith-number/DQ5GTTDCNVBVTKKFS6FIV257UP/graph.json","fetch_events":"https://pith.science/api/pith-number/DQ5GTTDCNVBVTKKFS6FIV257UP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP/action/storage_attestation","attest_author":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP/action/author_attestation","sign_citation":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP/action/citation_signature","submit_replication":"https://pith.science/pith/DQ5GTTDCNVBVTKKFS6FIV257UP/action/replication_record"}},"created_at":"2026-07-05T11:17:26.199757+00:00","updated_at":"2026-07-05T11:17:26.199757+00:00"}