{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:CWB5LQYWMTAOGT3JDHWPRCXTD7","short_pith_number":"pith:CWB5LQYW","schema_version":"1.0","canonical_sha256":"1583d5c31664c0e34f6919ecf88af31fc221a18f9c237c294810cf14d2d7f38b","source":{"kind":"arxiv","id":"2107.06665","version":1},"attestation_state":"computed","paper":{"title":"Disparity Between Batches as a Signal for Early Stopping","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Mahsa Forouzesh, Patrick Thiran","submitted_at":"2021-07-14T12:59:01Z","abstract_excerpt":"We propose a metric for evaluating the generalization ability of deep neural networks trained with mini-batch gradient descent. Our metric, called gradient disparity, is the $\\ell_2$ norm distance between the gradient vectors of two mini-batches drawn from the training set. It is derived from a probabilistic upper bound on the difference between the classification errors over a given mini-batch, when the network is trained on this mini-batch and when the network is trained on another mini-batch of points sampled from the same dataset. We empirically show that gradient disparity is a very promi"},"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":"2107.06665","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-07-14T12:59:01Z","cross_cats_sorted":[],"title_canon_sha256":"ac2c80ac6a037816939fec3b8c1eb2b885401ff748e58808e18aecfc2ce66e7a","abstract_canon_sha256":"4f879c5ca035b8402a6af59bf38d81ab2dece75e8c78f0a559054e204fb6193b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:58:00.116194Z","signature_b64":"WLK3eUJexEE9U/7gUXoQelhp2SZO2yvmjDrlDLG0fgeJKaXubH06u53FzxoIM+0AT9iLmQ+Rq2E/Lto9rPmTBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1583d5c31664c0e34f6919ecf88af31fc221a18f9c237c294810cf14d2d7f38b","last_reissued_at":"2026-07-05T02:58:00.115782Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:58:00.115782Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Disparity Between Batches as a Signal for Early Stopping","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Mahsa Forouzesh, Patrick Thiran","submitted_at":"2021-07-14T12:59:01Z","abstract_excerpt":"We propose a metric for evaluating the generalization ability of deep neural networks trained with mini-batch gradient descent. Our metric, called gradient disparity, is the $\\ell_2$ norm distance between the gradient vectors of two mini-batches drawn from the training set. It is derived from a probabilistic upper bound on the difference between the classification errors over a given mini-batch, when the network is trained on this mini-batch and when the network is trained on another mini-batch of points sampled from the same dataset. We empirically show that gradient disparity is a very promi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2107.06665","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/2107.06665/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":"2107.06665","created_at":"2026-07-05T02:58:00.115847+00:00"},{"alias_kind":"arxiv_version","alias_value":"2107.06665v1","created_at":"2026-07-05T02:58:00.115847+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2107.06665","created_at":"2026-07-05T02:58:00.115847+00:00"},{"alias_kind":"pith_short_12","alias_value":"CWB5LQYWMTAO","created_at":"2026-07-05T02:58:00.115847+00:00"},{"alias_kind":"pith_short_16","alias_value":"CWB5LQYWMTAOGT3J","created_at":"2026-07-05T02:58:00.115847+00:00"},{"alias_kind":"pith_short_8","alias_value":"CWB5LQYW","created_at":"2026-07-05T02:58:00.115847+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/CWB5LQYWMTAOGT3JDHWPRCXTD7","json":"https://pith.science/pith/CWB5LQYWMTAOGT3JDHWPRCXTD7.json","graph_json":"https://pith.science/api/pith-number/CWB5LQYWMTAOGT3JDHWPRCXTD7/graph.json","events_json":"https://pith.science/api/pith-number/CWB5LQYWMTAOGT3JDHWPRCXTD7/events.json","paper":"https://pith.science/paper/CWB5LQYW"},"agent_actions":{"view_html":"https://pith.science/pith/CWB5LQYWMTAOGT3JDHWPRCXTD7","download_json":"https://pith.science/pith/CWB5LQYWMTAOGT3JDHWPRCXTD7.json","view_paper":"https://pith.science/paper/CWB5LQYW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2107.06665&json=true","fetch_graph":"https://pith.science/api/pith-number/CWB5LQYWMTAOGT3JDHWPRCXTD7/graph.json","fetch_events":"https://pith.science/api/pith-number/CWB5LQYWMTAOGT3JDHWPRCXTD7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CWB5LQYWMTAOGT3JDHWPRCXTD7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CWB5LQYWMTAOGT3JDHWPRCXTD7/action/storage_attestation","attest_author":"https://pith.science/pith/CWB5LQYWMTAOGT3JDHWPRCXTD7/action/author_attestation","sign_citation":"https://pith.science/pith/CWB5LQYWMTAOGT3JDHWPRCXTD7/action/citation_signature","submit_replication":"https://pith.science/pith/CWB5LQYWMTAOGT3JDHWPRCXTD7/action/replication_record"}},"created_at":"2026-07-05T02:58:00.115847+00:00","updated_at":"2026-07-05T02:58:00.115847+00:00"}