{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:CWJJEWKKSZ6OLMW5G3CDSVLOZK","short_pith_number":"pith:CWJJEWKK","schema_version":"1.0","canonical_sha256":"159292594a967ce5b2dd36c439556ecaab399a3ad722a61c1b5f59b2abd85b04","source":{"kind":"arxiv","id":"2310.02575","version":2},"attestation_state":"computed","paper":{"title":"AdaMerging: Adaptive Model Merging for Multi-Task Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Dacheng Tao, Enneng Yang, Guibing Guo, Li Shen, Shiwei Liu, Xingwei Wang, Zhenyi Wang","submitted_at":"2023-10-04T04:26:33Z","abstract_excerpt":"Multi-task learning (MTL) aims to empower a model to tackle multiple tasks simultaneously. A recent development known as task arithmetic has revealed that several models, each fine-tuned for distinct tasks, can be directly merged into a single model to execute MTL without necessitating a retraining process using the initial training data. Nevertheless, this direct addition of models often leads to a significant deterioration in the overall performance of the merged model. This decline occurs due to potential conflicts and intricate correlations among the multiple tasks. Consequently, the chall"},"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":"2310.02575","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-10-04T04:26:33Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"dce26c644963ff51f11e93bbc2914f339eede188ff9009f4971a8754bb036f24","abstract_canon_sha256":"e638643ebbdc02b7fe51de91cef6439fedca594ebe80146a4b8d1dd328e317de"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:23:56.065517Z","signature_b64":"xI5hWv5WNzg5qzoOjpgXOXM8dBHhK2GvP4gp7+AGE6Sw01rHjohAnakfeOU+hMtyuKegkv/M9ySIQ7bncd7qCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"159292594a967ce5b2dd36c439556ecaab399a3ad722a61c1b5f59b2abd85b04","last_reissued_at":"2026-07-05T08:23:56.065075Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:23:56.065075Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AdaMerging: Adaptive Model Merging for Multi-Task Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Dacheng Tao, Enneng Yang, Guibing Guo, Li Shen, Shiwei Liu, Xingwei Wang, Zhenyi Wang","submitted_at":"2023-10-04T04:26:33Z","abstract_excerpt":"Multi-task learning (MTL) aims to empower a model to tackle multiple tasks simultaneously. A recent development known as task arithmetic has revealed that several models, each fine-tuned for distinct tasks, can be directly merged into a single model to execute MTL without necessitating a retraining process using the initial training data. Nevertheless, this direct addition of models often leads to a significant deterioration in the overall performance of the merged model. This decline occurs due to potential conflicts and intricate correlations among the multiple tasks. Consequently, the chall"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.02575","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/2310.02575/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":"2310.02575","created_at":"2026-07-05T08:23:56.065129+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.02575v2","created_at":"2026-07-05T08:23:56.065129+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.02575","created_at":"2026-07-05T08:23:56.065129+00:00"},{"alias_kind":"pith_short_12","alias_value":"CWJJEWKKSZ6O","created_at":"2026-07-05T08:23:56.065129+00:00"},{"alias_kind":"pith_short_16","alias_value":"CWJJEWKKSZ6OLMW5","created_at":"2026-07-05T08:23:56.065129+00:00"},{"alias_kind":"pith_short_8","alias_value":"CWJJEWKK","created_at":"2026-07-05T08:23:56.065129+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":10,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03391","citing_title":"When Model Merging Breaks Routing: Training-Free Calibration for MoE","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14084","citing_title":"CRANE: Constrained Reasoning Injection for Code Agents via Nullspace Editing","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02606","citing_title":"ReLoRA: Knowledge-Reusing Adaptation for Fast Rollout of Evolving LLM Services","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2509.24244","citing_title":"Model Merging Scaling Laws in Large Language Models","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2511.11439","citing_title":"Retrofit: Continual Learning with Controlled Forgetting for Binary Security Detection and Analysis","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14084","citing_title":"CRANE: Constrained Reasoning Injection for Code Agents via Nullspace Editing","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14546","citing_title":"Discovering Physical Directions in Weight Space: Composing Neural PDE Experts","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09879","citing_title":"M2A: Synergizing Mathematical and Agentic Reasoning in Large Language Models","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20985","citing_title":"Differentially Private Model Merging","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22823","citing_title":"PivotMerge: Bridging Heterogeneous Multimodal Pre-training via Post-Alignment Model Merging","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK","json":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK.json","graph_json":"https://pith.science/api/pith-number/CWJJEWKKSZ6OLMW5G3CDSVLOZK/graph.json","events_json":"https://pith.science/api/pith-number/CWJJEWKKSZ6OLMW5G3CDSVLOZK/events.json","paper":"https://pith.science/paper/CWJJEWKK"},"agent_actions":{"view_html":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK","download_json":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK.json","view_paper":"https://pith.science/paper/CWJJEWKK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.02575&json=true","fetch_graph":"https://pith.science/api/pith-number/CWJJEWKKSZ6OLMW5G3CDSVLOZK/graph.json","fetch_events":"https://pith.science/api/pith-number/CWJJEWKKSZ6OLMW5G3CDSVLOZK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK/action/storage_attestation","attest_author":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK/action/author_attestation","sign_citation":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK/action/citation_signature","submit_replication":"https://pith.science/pith/CWJJEWKKSZ6OLMW5G3CDSVLOZK/action/replication_record"}},"created_at":"2026-07-05T08:23:56.065129+00:00","updated_at":"2026-07-05T08:23:56.065129+00:00"}