{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FRSD7SFB6MYRF57D37MCCVSUYE","short_pith_number":"pith:FRSD7SFB","schema_version":"1.0","canonical_sha256":"2c643fc8a1f33112f7e3dfd8215654c130b4f2ed7f46c1a57b192a1c1283ab89","source":{"kind":"arxiv","id":"2506.09093","version":2},"attestation_state":"computed","paper":{"title":"Merging Smarter, Generalizing Better: Enhancing Model Merging on OOD Data","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bingjie Zhang, Changlong Shi, Dandan Guo, Dongsheng Wang, Guowei Rong, He Zhao, Hongkang Li, Meng Wang","submitted_at":"2025-06-10T11:34:23Z","abstract_excerpt":"Multi-task learning (MTL) concurrently trains a model on diverse task datasets to exploit common features, thereby improving overall performance across the tasks. Recent studies have dedicated efforts to merging multiple independent model parameters into a unified model for MTL, thus circumventing the need for training data and expanding the scope of applicable scenarios of MTL. However, current approaches to model merging predominantly concentrate on enhancing performance within in-domain (ID) datasets, often overlooking their efficacy on out-of-domain (OOD) datasets. In this work, we propose"},"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":"2506.09093","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-10T11:34:23Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3d584626f73e07dd89f329174ffddb71a767add2703e15953b2f90ab8b72e597","abstract_canon_sha256":"0e4dde8a6f51125d423e607b205daed836d2751feb7d576f953f34bf7b5c30b4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:20:56.491677Z","signature_b64":"B2d3UEfeG08QzSjbc5ycLFO1uGXqCPNZN47bJu+8I8yDBas1PafahwWzj/oj8Qq+RrvZazglnn5APiQQSrGUCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2c643fc8a1f33112f7e3dfd8215654c130b4f2ed7f46c1a57b192a1c1283ab89","last_reissued_at":"2026-07-05T11:20:56.491183Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:20:56.491183Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Merging Smarter, Generalizing Better: Enhancing Model Merging on OOD Data","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bingjie Zhang, Changlong Shi, Dandan Guo, Dongsheng Wang, Guowei Rong, He Zhao, Hongkang Li, Meng Wang","submitted_at":"2025-06-10T11:34:23Z","abstract_excerpt":"Multi-task learning (MTL) concurrently trains a model on diverse task datasets to exploit common features, thereby improving overall performance across the tasks. Recent studies have dedicated efforts to merging multiple independent model parameters into a unified model for MTL, thus circumventing the need for training data and expanding the scope of applicable scenarios of MTL. However, current approaches to model merging predominantly concentrate on enhancing performance within in-domain (ID) datasets, often overlooking their efficacy on out-of-domain (OOD) datasets. In this work, we propose"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.09093","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/2506.09093/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":"2506.09093","created_at":"2026-07-05T11:20:56.491240+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.09093v2","created_at":"2026-07-05T11:20:56.491240+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.09093","created_at":"2026-07-05T11:20:56.491240+00:00"},{"alias_kind":"pith_short_12","alias_value":"FRSD7SFB6MYR","created_at":"2026-07-05T11:20:56.491240+00:00"},{"alias_kind":"pith_short_16","alias_value":"FRSD7SFB6MYRF57D","created_at":"2026-07-05T11:20:56.491240+00:00"},{"alias_kind":"pith_short_8","alias_value":"FRSD7SFB","created_at":"2026-07-05T11:20:56.491240+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.10074","citing_title":"Transformers Learn the Optimal DDPM Denoiser for Multi-Token GMMs","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06440","citing_title":"Visual prompting reimagined: The power of the Activation Prompts","ref_index":103,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE","json":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE.json","graph_json":"https://pith.science/api/pith-number/FRSD7SFB6MYRF57D37MCCVSUYE/graph.json","events_json":"https://pith.science/api/pith-number/FRSD7SFB6MYRF57D37MCCVSUYE/events.json","paper":"https://pith.science/paper/FRSD7SFB"},"agent_actions":{"view_html":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE","download_json":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE.json","view_paper":"https://pith.science/paper/FRSD7SFB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.09093&json=true","fetch_graph":"https://pith.science/api/pith-number/FRSD7SFB6MYRF57D37MCCVSUYE/graph.json","fetch_events":"https://pith.science/api/pith-number/FRSD7SFB6MYRF57D37MCCVSUYE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE/action/storage_attestation","attest_author":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE/action/author_attestation","sign_citation":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE/action/citation_signature","submit_replication":"https://pith.science/pith/FRSD7SFB6MYRF57D37MCCVSUYE/action/replication_record"}},"created_at":"2026-07-05T11:20:56.491240+00:00","updated_at":"2026-07-05T11:20:56.491240+00:00"}