{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:I7CPU6ZVHURQEEJPEVUJFQSYAJ","short_pith_number":"pith:I7CPU6ZV","schema_version":"1.0","canonical_sha256":"47c4fa7b353d2302112f256892c258024e41daa78b3e13d54b3128f0ed02e941","source":{"kind":"arxiv","id":"2504.13407","version":1},"attestation_state":"computed","paper":{"title":"LoRA-Based Continual Learning with Constraints on Critical Parameter Changes","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Hongliang Li, Jiangwei Zhao, Liang Zhang, Lili Pan, Shimou Ling","submitted_at":"2025-04-18T02:08:19Z","abstract_excerpt":"LoRA-based continual learning represents a promising avenue for leveraging pre-trained models in downstream continual learning tasks. Recent studies have shown that orthogonal LoRA tuning effectively mitigates forgetting. However, this work unveils that under orthogonal LoRA tuning, the critical parameters for pre-tasks still change notably after learning post-tasks. To address this problem, we directly propose freezing the most critical parameter matrices in the Vision Transformer (ViT) for pre-tasks before learning post-tasks. In addition, building on orthogonal LoRA tuning, we propose ortho"},"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":"2504.13407","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-04-18T02:08:19Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"97ff5720cd8fa73a7bf19893d54d388024176144371cbae72f07f21471838007","abstract_canon_sha256":"c55c5f8b9169aa10d18e805df657333db60e92b42c6565c1c7c89d5cff4b6a6f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:50:54.361548Z","signature_b64":"Pd3405OdriYzSjRWmFBrhXHOB5Z5W0EGyGi56iD+bgE7N3m1EZ3sLdMh9jVSXt3Ao0RA7F5WRdBZir4QqxZWAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"47c4fa7b353d2302112f256892c258024e41daa78b3e13d54b3128f0ed02e941","last_reissued_at":"2026-07-05T10:50:54.361044Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:50:54.361044Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LoRA-Based Continual Learning with Constraints on Critical Parameter Changes","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Hongliang Li, Jiangwei Zhao, Liang Zhang, Lili Pan, Shimou Ling","submitted_at":"2025-04-18T02:08:19Z","abstract_excerpt":"LoRA-based continual learning represents a promising avenue for leveraging pre-trained models in downstream continual learning tasks. Recent studies have shown that orthogonal LoRA tuning effectively mitigates forgetting. However, this work unveils that under orthogonal LoRA tuning, the critical parameters for pre-tasks still change notably after learning post-tasks. To address this problem, we directly propose freezing the most critical parameter matrices in the Vision Transformer (ViT) for pre-tasks before learning post-tasks. In addition, building on orthogonal LoRA tuning, we propose ortho"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.13407","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/2504.13407/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":"2504.13407","created_at":"2026-07-05T10:50:54.361113+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.13407v1","created_at":"2026-07-05T10:50:54.361113+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.13407","created_at":"2026-07-05T10:50:54.361113+00:00"},{"alias_kind":"pith_short_12","alias_value":"I7CPU6ZVHURQ","created_at":"2026-07-05T10:50:54.361113+00:00"},{"alias_kind":"pith_short_16","alias_value":"I7CPU6ZVHURQEEJP","created_at":"2026-07-05T10:50:54.361113+00:00"},{"alias_kind":"pith_short_8","alias_value":"I7CPU6ZV","created_at":"2026-07-05T10:50:54.361113+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ","json":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ.json","graph_json":"https://pith.science/api/pith-number/I7CPU6ZVHURQEEJPEVUJFQSYAJ/graph.json","events_json":"https://pith.science/api/pith-number/I7CPU6ZVHURQEEJPEVUJFQSYAJ/events.json","paper":"https://pith.science/paper/I7CPU6ZV"},"agent_actions":{"view_html":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ","download_json":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ.json","view_paper":"https://pith.science/paper/I7CPU6ZV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.13407&json=true","fetch_graph":"https://pith.science/api/pith-number/I7CPU6ZVHURQEEJPEVUJFQSYAJ/graph.json","fetch_events":"https://pith.science/api/pith-number/I7CPU6ZVHURQEEJPEVUJFQSYAJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ/action/storage_attestation","attest_author":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ/action/author_attestation","sign_citation":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ/action/citation_signature","submit_replication":"https://pith.science/pith/I7CPU6ZVHURQEEJPEVUJFQSYAJ/action/replication_record"}},"created_at":"2026-07-05T10:50:54.361113+00:00","updated_at":"2026-07-05T10:50:54.361113+00:00"}