{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:4XA3FU654LMXCPMQK4QGC4RQG7","short_pith_number":"pith:4XA3FU65","schema_version":"1.0","canonical_sha256":"e5c1b2d3dde2d9713d90572061723037e2c980d7a9ed30a334dd42e9d9f993d9","source":{"kind":"arxiv","id":"2110.08611","version":2},"attestation_state":"computed","paper":{"title":"Deep Active Learning by Leveraging Training Dynamics","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andrew Margenot, Hanghang Tong, Haonan Wang, Jingrui He, Wei Huang, Ziwei Wu","submitted_at":"2021-10-16T16:51:05Z","abstract_excerpt":"Active learning theories and methods have been extensively studied in classical statistical learning settings. However, deep active learning, i.e., active learning with deep learning models, is usually based on empirical criteria without solid theoretical justification, thus suffering from heavy doubts when some of those fail to provide benefits in real applications. In this paper, by exploring the connection between the generalization performance and the training dynamics, we propose a theory-driven deep active learning method (dynamicAL) which selects samples to maximize training dynamics. I"},"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":"2110.08611","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-10-16T16:51:05Z","cross_cats_sorted":[],"title_canon_sha256":"d6c4019cbe0bf1fa9751ee71e224914209068e99f596f0d45c8fca879a5f00be","abstract_canon_sha256":"26f918bfd741e7bb3b503ed92f27c75fe2decbb06b55794f28e9f74695e13d98"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:17:23.947539Z","signature_b64":"MIlZQduFvCbpGy4bYuIRsR5+mmWoNaseThTen0rYhmMM4lEW4P/MTkrZkaoYn/b1qpkKNE/gk8N62k0duEQ2Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e5c1b2d3dde2d9713d90572061723037e2c980d7a9ed30a334dd42e9d9f993d9","last_reissued_at":"2026-07-05T05:17:23.947040Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:17:23.947040Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Deep Active Learning by Leveraging Training Dynamics","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andrew Margenot, Hanghang Tong, Haonan Wang, Jingrui He, Wei Huang, Ziwei Wu","submitted_at":"2021-10-16T16:51:05Z","abstract_excerpt":"Active learning theories and methods have been extensively studied in classical statistical learning settings. However, deep active learning, i.e., active learning with deep learning models, is usually based on empirical criteria without solid theoretical justification, thus suffering from heavy doubts when some of those fail to provide benefits in real applications. In this paper, by exploring the connection between the generalization performance and the training dynamics, we propose a theory-driven deep active learning method (dynamicAL) which selects samples to maximize training dynamics. I"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.08611","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/2110.08611/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":"2110.08611","created_at":"2026-07-05T05:17:23.947096+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.08611v2","created_at":"2026-07-05T05:17:23.947096+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.08611","created_at":"2026-07-05T05:17:23.947096+00:00"},{"alias_kind":"pith_short_12","alias_value":"4XA3FU654LMX","created_at":"2026-07-05T05:17:23.947096+00:00"},{"alias_kind":"pith_short_16","alias_value":"4XA3FU654LMXCPMQ","created_at":"2026-07-05T05:17:23.947096+00:00"},{"alias_kind":"pith_short_8","alias_value":"4XA3FU65","created_at":"2026-07-05T05:17:23.947096+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/4XA3FU654LMXCPMQK4QGC4RQG7","json":"https://pith.science/pith/4XA3FU654LMXCPMQK4QGC4RQG7.json","graph_json":"https://pith.science/api/pith-number/4XA3FU654LMXCPMQK4QGC4RQG7/graph.json","events_json":"https://pith.science/api/pith-number/4XA3FU654LMXCPMQK4QGC4RQG7/events.json","paper":"https://pith.science/paper/4XA3FU65"},"agent_actions":{"view_html":"https://pith.science/pith/4XA3FU654LMXCPMQK4QGC4RQG7","download_json":"https://pith.science/pith/4XA3FU654LMXCPMQK4QGC4RQG7.json","view_paper":"https://pith.science/paper/4XA3FU65","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.08611&json=true","fetch_graph":"https://pith.science/api/pith-number/4XA3FU654LMXCPMQK4QGC4RQG7/graph.json","fetch_events":"https://pith.science/api/pith-number/4XA3FU654LMXCPMQK4QGC4RQG7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4XA3FU654LMXCPMQK4QGC4RQG7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4XA3FU654LMXCPMQK4QGC4RQG7/action/storage_attestation","attest_author":"https://pith.science/pith/4XA3FU654LMXCPMQK4QGC4RQG7/action/author_attestation","sign_citation":"https://pith.science/pith/4XA3FU654LMXCPMQK4QGC4RQG7/action/citation_signature","submit_replication":"https://pith.science/pith/4XA3FU654LMXCPMQK4QGC4RQG7/action/replication_record"}},"created_at":"2026-07-05T05:17:23.947096+00:00","updated_at":"2026-07-05T05:17:23.947096+00:00"}