{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CA2MS7I437KBZXM2EHEI2LE2OQ","short_pith_number":"pith:CA2MS7I4","schema_version":"1.0","canonical_sha256":"1034c97d1cdfd41cdd9a21c88d2c9a7426681da9f77ef4e343bf8067fb4e3662","source":{"kind":"arxiv","id":"2504.12436","version":2},"attestation_state":"computed","paper":{"title":"Sparsity Outperforms Low-Rank Projections in Few-Shot Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"\\'Eric Granger, Ismail Ben Ayed, Nairouz Mrabah, Nicolas Richet","submitted_at":"2025-04-16T19:10:34Z","abstract_excerpt":"Adapting Vision-Language Models (VLMs) to new domains with few labeled samples remains a significant challenge due to severe overfitting and computational constraints. State-of-the-art solutions, such as low-rank reparameterization, mitigate these issues but often struggle with generalization and require extensive hyperparameter tuning. In this paper, a novel Sparse Optimization (SO) framework is proposed. Unlike low-rank approaches that typically constrain updates to a fixed subspace, our SO method leverages high sparsity to dynamically adjust very few parameters. We introduce two key paradig"},"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":"2504.12436","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-04-16T19:10:34Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b0d206dbbb67257e874f4139c6cae05dcc5a0a7ce428b0254c68916915814ef0","abstract_canon_sha256":"a190d622707635c7066693a71a5a63745e8acf01be756339734a988527c296f6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:01.950134Z","signature_b64":"1vzVt87OqW9uxC2gT+YYkln2DhZt5UGHDkXd3Y+hse2XWdPMsyne46cHz9tPhM43EGdJzltVgC+7VC1vGJmJDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1034c97d1cdfd41cdd9a21c88d2c9a7426681da9f77ef4e343bf8067fb4e3662","last_reissued_at":"2026-07-05T11:52:01.949685Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:01.949685Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sparsity Outperforms Low-Rank Projections in Few-Shot Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"\\'Eric Granger, Ismail Ben Ayed, Nairouz Mrabah, Nicolas Richet","submitted_at":"2025-04-16T19:10:34Z","abstract_excerpt":"Adapting Vision-Language Models (VLMs) to new domains with few labeled samples remains a significant challenge due to severe overfitting and computational constraints. State-of-the-art solutions, such as low-rank reparameterization, mitigate these issues but often struggle with generalization and require extensive hyperparameter tuning. In this paper, a novel Sparse Optimization (SO) framework is proposed. Unlike low-rank approaches that typically constrain updates to a fixed subspace, our SO method leverages high sparsity to dynamically adjust very few parameters. We introduce two key paradig"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.12436","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/2504.12436/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":"2504.12436","created_at":"2026-07-05T11:52:01.949744+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.12436v2","created_at":"2026-07-05T11:52:01.949744+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.12436","created_at":"2026-07-05T11:52:01.949744+00:00"},{"alias_kind":"pith_short_12","alias_value":"CA2MS7I437KB","created_at":"2026-07-05T11:52:01.949744+00:00"},{"alias_kind":"pith_short_16","alias_value":"CA2MS7I437KBZXM2","created_at":"2026-07-05T11:52:01.949744+00:00"},{"alias_kind":"pith_short_8","alias_value":"CA2MS7I4","created_at":"2026-07-05T11:52:01.949744+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/CA2MS7I437KBZXM2EHEI2LE2OQ","json":"https://pith.science/pith/CA2MS7I437KBZXM2EHEI2LE2OQ.json","graph_json":"https://pith.science/api/pith-number/CA2MS7I437KBZXM2EHEI2LE2OQ/graph.json","events_json":"https://pith.science/api/pith-number/CA2MS7I437KBZXM2EHEI2LE2OQ/events.json","paper":"https://pith.science/paper/CA2MS7I4"},"agent_actions":{"view_html":"https://pith.science/pith/CA2MS7I437KBZXM2EHEI2LE2OQ","download_json":"https://pith.science/pith/CA2MS7I437KBZXM2EHEI2LE2OQ.json","view_paper":"https://pith.science/paper/CA2MS7I4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.12436&json=true","fetch_graph":"https://pith.science/api/pith-number/CA2MS7I437KBZXM2EHEI2LE2OQ/graph.json","fetch_events":"https://pith.science/api/pith-number/CA2MS7I437KBZXM2EHEI2LE2OQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CA2MS7I437KBZXM2EHEI2LE2OQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CA2MS7I437KBZXM2EHEI2LE2OQ/action/storage_attestation","attest_author":"https://pith.science/pith/CA2MS7I437KBZXM2EHEI2LE2OQ/action/author_attestation","sign_citation":"https://pith.science/pith/CA2MS7I437KBZXM2EHEI2LE2OQ/action/citation_signature","submit_replication":"https://pith.science/pith/CA2MS7I437KBZXM2EHEI2LE2OQ/action/replication_record"}},"created_at":"2026-07-05T11:52:01.949744+00:00","updated_at":"2026-07-05T11:52:01.949744+00:00"}