{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:77XMKQSJBOPKY4LJSLNFMIH3UU","short_pith_number":"pith:77XMKQSJ","schema_version":"1.0","canonical_sha256":"ffeec542490b9eac716992da5620fba53582ad1ef04007eb841f7abd5f23072e","source":{"kind":"arxiv","id":"2402.04087","version":1},"attestation_state":"computed","paper":{"title":"A Hard-to-Beat Baseline for Training-free CLIP-based Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Jian Liang, Lijun Sheng, Ran He, Tieniu Tan, Zhengbo Wang, Zilei Wang","submitted_at":"2024-02-06T15:45:27Z","abstract_excerpt":"Contrastive Language-Image Pretraining (CLIP) has gained popularity for its remarkable zero-shot capacity. Recent research has focused on developing efficient fine-tuning methods, such as prompt learning and adapter, to enhance CLIP's performance in downstream tasks. However, these methods still require additional training time and computational resources, which is undesirable for devices with limited resources. In this paper, we revisit a classical algorithm, Gaussian Discriminant Analysis (GDA), and apply it to the downstream classification of CLIP. Typically, GDA assumes that features of ea"},"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":"2402.04087","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-02-06T15:45:27Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"42133b0c17fb8920e37c4bae742708ee94e38652360d8d36b68ab532ae23ecb8","abstract_canon_sha256":"a44f6c617981dfb57b74c19df31ebba2dc183d0829bff5b086cfeec4652caf23"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:42:02.821726Z","signature_b64":"swyRTPZXjDG6k45dpQc/zsFz5Mw2r1wqBGhwMJbwSiXd79Mj1Ssy2gW3+Q4wWqqC7bH8gelu1XcVDWQVZ/32Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ffeec542490b9eac716992da5620fba53582ad1ef04007eb841f7abd5f23072e","last_reissued_at":"2026-07-05T07:42:02.821226Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:42:02.821226Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Hard-to-Beat Baseline for Training-free CLIP-based Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Jian Liang, Lijun Sheng, Ran He, Tieniu Tan, Zhengbo Wang, Zilei Wang","submitted_at":"2024-02-06T15:45:27Z","abstract_excerpt":"Contrastive Language-Image Pretraining (CLIP) has gained popularity for its remarkable zero-shot capacity. Recent research has focused on developing efficient fine-tuning methods, such as prompt learning and adapter, to enhance CLIP's performance in downstream tasks. However, these methods still require additional training time and computational resources, which is undesirable for devices with limited resources. In this paper, we revisit a classical algorithm, Gaussian Discriminant Analysis (GDA), and apply it to the downstream classification of CLIP. Typically, GDA assumes that features of ea"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.04087","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/2402.04087/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":"2402.04087","created_at":"2026-07-05T07:42:02.821286+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.04087v1","created_at":"2026-07-05T07:42:02.821286+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.04087","created_at":"2026-07-05T07:42:02.821286+00:00"},{"alias_kind":"pith_short_12","alias_value":"77XMKQSJBOPK","created_at":"2026-07-05T07:42:02.821286+00:00"},{"alias_kind":"pith_short_16","alias_value":"77XMKQSJBOPKY4LJ","created_at":"2026-07-05T07:42:02.821286+00:00"},{"alias_kind":"pith_short_8","alias_value":"77XMKQSJ","created_at":"2026-07-05T07:42:02.821286+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03730","citing_title":"Beyond False Stability: High-Noise Drift Gating for Test-Time Adversarial Defenses in Vision-Language Models","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25821","citing_title":"[CLS] is Not Enough: Multi-Label Recognition via Patch-Level Inference and Adaptive Aggregation","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2509.03740","citing_title":"CLIP-SVD: Efficient and Interpretable Vision-Language Adaptation via Singular Values","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23977","citing_title":"Multi-View Synergistic Learning with Vision-Language Adaption for Low-Resource Biomedical Image Classification","ref_index":51,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU","json":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU.json","graph_json":"https://pith.science/api/pith-number/77XMKQSJBOPKY4LJSLNFMIH3UU/graph.json","events_json":"https://pith.science/api/pith-number/77XMKQSJBOPKY4LJSLNFMIH3UU/events.json","paper":"https://pith.science/paper/77XMKQSJ"},"agent_actions":{"view_html":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU","download_json":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU.json","view_paper":"https://pith.science/paper/77XMKQSJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.04087&json=true","fetch_graph":"https://pith.science/api/pith-number/77XMKQSJBOPKY4LJSLNFMIH3UU/graph.json","fetch_events":"https://pith.science/api/pith-number/77XMKQSJBOPKY4LJSLNFMIH3UU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU/action/storage_attestation","attest_author":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU/action/author_attestation","sign_citation":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU/action/citation_signature","submit_replication":"https://pith.science/pith/77XMKQSJBOPKY4LJSLNFMIH3UU/action/replication_record"}},"created_at":"2026-07-05T07:42:02.821286+00:00","updated_at":"2026-07-05T07:42:02.821286+00:00"}