{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:OCXKLV426LDXCCJME7JSY36CQW","short_pith_number":"pith:OCXKLV42","schema_version":"1.0","canonical_sha256":"70aea5d79af2c771092c27d32c6fc285a5148dd5ea3f5b3d479333e27092d377","source":{"kind":"arxiv","id":"1611.01838","version":5},"attestation_state":"computed","paper":{"title":"Entropy-SGD: Biasing Gradient Descent Into Wide Valleys","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Anna Choromanska, Carlo Baldassi, Christian Borgs, Jennifer Chayes, Levent Sagun, Pratik Chaudhari, Riccardo Zecchina, Stefano Soatto, Yann LeCun","submitted_at":"2016-11-06T20:22:49Z","abstract_excerpt":"This paper proposes a new optimization algorithm called Entropy-SGD for training deep neural networks that is motivated by the local geometry of the energy landscape. Local extrema with low generalization error have a large proportion of almost-zero eigenvalues in the Hessian with very few positive or negative eigenvalues. We leverage upon this observation to construct a local-entropy-based objective function that favors well-generalizable solutions lying in large flat regions of the energy landscape, while avoiding poorly-generalizable solutions located in the sharp valleys. Conceptually, our"},"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":"1611.01838","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2016-11-06T20:22:49Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"98edffd0d0debf11130af9ab391f0ae517df463424820ad98268b62bc1ad299c","abstract_canon_sha256":"4b36b9ae1b94b7e17adc82c70bf48fb68b45ba8c4d9f9ebf21f09ffaac022add"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:46:02.390449Z","signature_b64":"opk87gotDDbWNKYWrp4gaKUx2dcQb5CEfIz1iej1gxV8SKSKssQdXBXjhQteGQFAZD/tNhOoLS/sb6A7qTBJCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"70aea5d79af2c771092c27d32c6fc285a5148dd5ea3f5b3d479333e27092d377","last_reissued_at":"2026-05-18T00:46:02.389976Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:46:02.389976Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Entropy-SGD: Biasing Gradient Descent Into Wide Valleys","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Anna Choromanska, Carlo Baldassi, Christian Borgs, Jennifer Chayes, Levent Sagun, Pratik Chaudhari, Riccardo Zecchina, Stefano Soatto, Yann LeCun","submitted_at":"2016-11-06T20:22:49Z","abstract_excerpt":"This paper proposes a new optimization algorithm called Entropy-SGD for training deep neural networks that is motivated by the local geometry of the energy landscape. Local extrema with low generalization error have a large proportion of almost-zero eigenvalues in the Hessian with very few positive or negative eigenvalues. We leverage upon this observation to construct a local-entropy-based objective function that favors well-generalizable solutions lying in large flat regions of the energy landscape, while avoiding poorly-generalizable solutions located in the sharp valleys. Conceptually, our"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1611.01838","kind":"arxiv","version":5},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"1611.01838","created_at":"2026-05-18T00:46:02.390051+00:00"},{"alias_kind":"arxiv_version","alias_value":"1611.01838v5","created_at":"2026-05-18T00:46:02.390051+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1611.01838","created_at":"2026-05-18T00:46:02.390051+00:00"},{"alias_kind":"pith_short_12","alias_value":"OCXKLV426LDX","created_at":"2026-05-18T12:30:36.002864+00:00"},{"alias_kind":"pith_short_16","alias_value":"OCXKLV426LDXCCJM","created_at":"2026-05-18T12:30:36.002864+00:00"},{"alias_kind":"pith_short_8","alias_value":"OCXKLV42","created_at":"2026-05-18T12:30:36.002864+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2010.01412","citing_title":"Sharpness-Aware Minimization for Efficiently Improving Generalization","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"1609.04836","citing_title":"On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05436","citing_title":"Estimating Implicit Regularization in Deep Learning","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02105","citing_title":"Sharpness-Aware Pretraining Mitigates Catastrophic Forgetting","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW","json":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW.json","graph_json":"https://pith.science/api/pith-number/OCXKLV426LDXCCJME7JSY36CQW/graph.json","events_json":"https://pith.science/api/pith-number/OCXKLV426LDXCCJME7JSY36CQW/events.json","paper":"https://pith.science/paper/OCXKLV42"},"agent_actions":{"view_html":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW","download_json":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW.json","view_paper":"https://pith.science/paper/OCXKLV42","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1611.01838&json=true","fetch_graph":"https://pith.science/api/pith-number/OCXKLV426LDXCCJME7JSY36CQW/graph.json","fetch_events":"https://pith.science/api/pith-number/OCXKLV426LDXCCJME7JSY36CQW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW/action/storage_attestation","attest_author":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW/action/author_attestation","sign_citation":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW/action/citation_signature","submit_replication":"https://pith.science/pith/OCXKLV426LDXCCJME7JSY36CQW/action/replication_record"}},"created_at":"2026-05-18T00:46:02.390051+00:00","updated_at":"2026-05-18T00:46:02.390051+00:00"}