{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:2P6ZXHSEQAPXHJPJOZH662EJGA","short_pith_number":"pith:2P6ZXHSE","schema_version":"1.0","canonical_sha256":"d3fd9b9e44801f73a5e9764fef688930058648acb3f988f27cbcbc679379447d","source":{"kind":"arxiv","id":"2207.09248","version":2},"attestation_state":"computed","paper":{"title":"Don't Stop Learning: Towards Continual Learning for the CLIP Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chunna Tian, Haoxuan Ding, Jingyuan Yang, Lingqiao Liu, Yuxuan Ding","submitted_at":"2022-07-19T13:03:14Z","abstract_excerpt":"The Contrastive Language-Image Pre-training (CLIP) Model is a recently proposed large-scale pre-train model which attracts increasing attention in the computer vision community. Benefiting from its gigantic image-text training set, the CLIP model has learned outstanding capabilities in zero-shot learning and image-text matching. To boost the recognition performance of CLIP on some target visual concepts, it is often desirable to further update the CLIP model by fine-tuning some classes-of-interest on extra training data. This operation, however, raises an important concern: will the update hur"},"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":"2207.09248","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-07-19T13:03:14Z","cross_cats_sorted":[],"title_canon_sha256":"dcd3d0c73db95a03e8230295279d101100944a38da79dc1ff0ce8537d6722850","abstract_canon_sha256":"b4017d0168d2e77ffc6f3cf5720306120e95c492bca3dac6bd9536ef52c654c9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:41:57.832278Z","signature_b64":"xuq4+onX20BU4YHJzDUqHR3AF923KTtit1EDDiOFou4KNlx3zj0cl2nmKZ/OdI6MztuVcBoiIcpNoZJmQ25OCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d3fd9b9e44801f73a5e9764fef688930058648acb3f988f27cbcbc679379447d","last_reissued_at":"2026-07-05T04:41:57.831944Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:41:57.831944Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Don't Stop Learning: Towards Continual Learning for the CLIP Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chunna Tian, Haoxuan Ding, Jingyuan Yang, Lingqiao Liu, Yuxuan Ding","submitted_at":"2022-07-19T13:03:14Z","abstract_excerpt":"The Contrastive Language-Image Pre-training (CLIP) Model is a recently proposed large-scale pre-train model which attracts increasing attention in the computer vision community. Benefiting from its gigantic image-text training set, the CLIP model has learned outstanding capabilities in zero-shot learning and image-text matching. To boost the recognition performance of CLIP on some target visual concepts, it is often desirable to further update the CLIP model by fine-tuning some classes-of-interest on extra training data. This operation, however, raises an important concern: will the update hur"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2207.09248","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/2207.09248/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":"2207.09248","created_at":"2026-07-05T04:41:57.831992+00:00"},{"alias_kind":"arxiv_version","alias_value":"2207.09248v2","created_at":"2026-07-05T04:41:57.831992+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2207.09248","created_at":"2026-07-05T04:41:57.831992+00:00"},{"alias_kind":"pith_short_12","alias_value":"2P6ZXHSEQAPX","created_at":"2026-07-05T04:41:57.831992+00:00"},{"alias_kind":"pith_short_16","alias_value":"2P6ZXHSEQAPXHJPJ","created_at":"2026-07-05T04:41:57.831992+00:00"},{"alias_kind":"pith_short_8","alias_value":"2P6ZXHSE","created_at":"2026-07-05T04:41:57.831992+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.25708","citing_title":"CMAP: Cross-Modal Adaptive Prompting for Multi-Domain Task-Incremental Learning","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2506.21035","citing_title":"Little by Little: Continual Learning via Incremental Mixture of Rank-1 Associative Memory Experts","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2508.04227","citing_title":"Continual Learning for VLMs: A Survey and Taxonomy Beyond Forgetting","ref_index":135,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19301","citing_title":"iGSP:Implicit Gradient Subspace Projection for Efficient Continual Learning of Vision-Language Models","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2601.20597","citing_title":"StructAlign: Structured Cross-Modal Alignment for Continual Text-to-Video Retrieval","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07494","citing_title":"DIMoE-Adapters: Dynamic Expert Evolution for Continual Learning in Vision-Language Models","ref_index":38,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA","json":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA.json","graph_json":"https://pith.science/api/pith-number/2P6ZXHSEQAPXHJPJOZH662EJGA/graph.json","events_json":"https://pith.science/api/pith-number/2P6ZXHSEQAPXHJPJOZH662EJGA/events.json","paper":"https://pith.science/paper/2P6ZXHSE"},"agent_actions":{"view_html":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA","download_json":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA.json","view_paper":"https://pith.science/paper/2P6ZXHSE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2207.09248&json=true","fetch_graph":"https://pith.science/api/pith-number/2P6ZXHSEQAPXHJPJOZH662EJGA/graph.json","fetch_events":"https://pith.science/api/pith-number/2P6ZXHSEQAPXHJPJOZH662EJGA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA/action/storage_attestation","attest_author":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA/action/author_attestation","sign_citation":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA/action/citation_signature","submit_replication":"https://pith.science/pith/2P6ZXHSEQAPXHJPJOZH662EJGA/action/replication_record"}},"created_at":"2026-07-05T04:41:57.831992+00:00","updated_at":"2026-07-05T04:41:57.831992+00:00"}