{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ODAVCNAFU23JFLHUU2VXEGIG3N","short_pith_number":"pith:ODAVCNAF","schema_version":"1.0","canonical_sha256":"70c1513405a6b692acf4a6ab721906db62a977c7d49c99b9e886e31d398ad3e7","source":{"kind":"arxiv","id":"2505.07690","version":1},"attestation_state":"computed","paper":{"title":"Beyond CLIP Generalization: Against Forward&Backward Forgetting Adapter for Continual Learning of Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenhao Ding, Jiangyang Li, Jizhou Han, Qiang Wang, SongLin Dong, Yihong Gong, Yuhang He","submitted_at":"2025-05-12T15:56:23Z","abstract_excerpt":"This study aims to address the problem of multi-domain task incremental learning~(MTIL), which requires that vision-language models~(VLMs) continuously acquire new knowledge while maintaining their inherent zero-shot recognition capability. Existing paradigms delegate the testing of unseen-domain samples to the original CLIP, which only prevents the degradation of the model's zero-shot capability but fails to enhance the generalization of the VLM further. To this end, we propose a novel MTIL framework, named AFA, which comprises two core modules: (1) an against forward-forgetting adapter that "},"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":"2505.07690","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-05-12T15:56:23Z","cross_cats_sorted":[],"title_canon_sha256":"f8297f1824b7f65b653a3ea0667ba527ff7da7b1bcc440d40f7833d8d7184b5d","abstract_canon_sha256":"f4e5243698c25e7b37062029f0b5f43656f8a273f0ffbc2e2fd892fb69df69bb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:01:51.766946Z","signature_b64":"+wsYxtITZwj6ildbfVW1eWFLfilt6Hrb8azKm+XQHNQjdrDmPrKgns/2MvSFIPTi5uykQC0hQuA+fAgvniU1Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"70c1513405a6b692acf4a6ab721906db62a977c7d49c99b9e886e31d398ad3e7","last_reissued_at":"2026-07-05T11:01:51.766265Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:01:51.766265Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond CLIP Generalization: Against Forward&Backward Forgetting Adapter for Continual Learning of Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenhao Ding, Jiangyang Li, Jizhou Han, Qiang Wang, SongLin Dong, Yihong Gong, Yuhang He","submitted_at":"2025-05-12T15:56:23Z","abstract_excerpt":"This study aims to address the problem of multi-domain task incremental learning~(MTIL), which requires that vision-language models~(VLMs) continuously acquire new knowledge while maintaining their inherent zero-shot recognition capability. Existing paradigms delegate the testing of unseen-domain samples to the original CLIP, which only prevents the degradation of the model's zero-shot capability but fails to enhance the generalization of the VLM further. To this end, we propose a novel MTIL framework, named AFA, which comprises two core modules: (1) an against forward-forgetting adapter that "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.07690","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/2505.07690/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":"2505.07690","created_at":"2026-07-05T11:01:51.766356+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.07690v1","created_at":"2026-07-05T11:01:51.766356+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.07690","created_at":"2026-07-05T11:01:51.766356+00:00"},{"alias_kind":"pith_short_12","alias_value":"ODAVCNAFU23J","created_at":"2026-07-05T11:01:51.766356+00:00"},{"alias_kind":"pith_short_16","alias_value":"ODAVCNAFU23JFLHU","created_at":"2026-07-05T11:01:51.766356+00:00"},{"alias_kind":"pith_short_8","alias_value":"ODAVCNAF","created_at":"2026-07-05T11:01:51.766356+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.10804","citing_title":"BPG: Balancing Plasticity and Generalization for Domain Incremental Learning","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N","json":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N.json","graph_json":"https://pith.science/api/pith-number/ODAVCNAFU23JFLHUU2VXEGIG3N/graph.json","events_json":"https://pith.science/api/pith-number/ODAVCNAFU23JFLHUU2VXEGIG3N/events.json","paper":"https://pith.science/paper/ODAVCNAF"},"agent_actions":{"view_html":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N","download_json":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N.json","view_paper":"https://pith.science/paper/ODAVCNAF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.07690&json=true","fetch_graph":"https://pith.science/api/pith-number/ODAVCNAFU23JFLHUU2VXEGIG3N/graph.json","fetch_events":"https://pith.science/api/pith-number/ODAVCNAFU23JFLHUU2VXEGIG3N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N/action/storage_attestation","attest_author":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N/action/author_attestation","sign_citation":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N/action/citation_signature","submit_replication":"https://pith.science/pith/ODAVCNAFU23JFLHUU2VXEGIG3N/action/replication_record"}},"created_at":"2026-07-05T11:01:51.766356+00:00","updated_at":"2026-07-05T11:01:51.766356+00:00"}