{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:AXOTLYSSDC3YUKPSZQUB5ZPXZA","short_pith_number":"pith:AXOTLYSS","schema_version":"1.0","canonical_sha256":"05dd35e25218b78a29f2cc281ee5f7c83d0d7e41723b287cbf9edb910fb3da06","source":{"kind":"arxiv","id":"1803.10986","version":3},"attestation_state":"computed","paper":{"title":"Error Analysis and Improving the Accuracy of Winograd Convolution for Deep Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.NA","authors_text":"Andrew Anderson, Barbara Barabasz, David Gregg, Kirk M. Soodhalter","submitted_at":"2018-03-29T09:48:02Z","abstract_excerpt":"Popular deep neural networks (DNNs) spend the majority of their execution time computing convolutions. The Winograd family of algorithms can greatly reduce the number of arithmetic operations required and is present in many DNN software frameworks. However, the performance gain is at the expense of a reduction in floating point (FP) numerical accuracy. In this paper, we analyse the worst case FP error and prove the estimation of norm and conditioning of the algorithm. We show that the bound grows exponentially with the size of the convolution, but the error bound of the \\textit{modified} algor"},"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":"1803.10986","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.NA","submitted_at":"2018-03-29T09:48:02Z","cross_cats_sorted":["cs.LG","stat.ML"],"title_canon_sha256":"d52529657d5669472de08d690005b88d04b4cea42c109c2660189a177b8b08c5","abstract_canon_sha256":"d66b80d5fb72f16d00475941a6a35dfabab732e139d2c1f5ae0cbb67e787912d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:47:12.693262Z","signature_b64":"AsFSaTYJR23HhrSA8AJrPyOWNHV/Iwdyd93/KdeBsexsH1kKrcW05/L2W6+yIfCYvLbWAfs8A3RpaKM5IkAvDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"05dd35e25218b78a29f2cc281ee5f7c83d0d7e41723b287cbf9edb910fb3da06","last_reissued_at":"2026-05-17T23:47:12.692787Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:47:12.692787Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Error Analysis and Improving the Accuracy of Winograd Convolution for Deep Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.NA","authors_text":"Andrew Anderson, Barbara Barabasz, David Gregg, Kirk M. Soodhalter","submitted_at":"2018-03-29T09:48:02Z","abstract_excerpt":"Popular deep neural networks (DNNs) spend the majority of their execution time computing convolutions. The Winograd family of algorithms can greatly reduce the number of arithmetic operations required and is present in many DNN software frameworks. However, the performance gain is at the expense of a reduction in floating point (FP) numerical accuracy. In this paper, we analyse the worst case FP error and prove the estimation of norm and conditioning of the algorithm. We show that the bound grows exponentially with the size of the convolution, but the error bound of the \\textit{modified} algor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1803.10986","kind":"arxiv","version":3},"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":"1803.10986","created_at":"2026-05-17T23:47:12.692868+00:00"},{"alias_kind":"arxiv_version","alias_value":"1803.10986v3","created_at":"2026-05-17T23:47:12.692868+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1803.10986","created_at":"2026-05-17T23:47:12.692868+00:00"},{"alias_kind":"pith_short_12","alias_value":"AXOTLYSSDC3Y","created_at":"2026-05-18T12:32:13.499390+00:00"},{"alias_kind":"pith_short_16","alias_value":"AXOTLYSSDC3YUKPS","created_at":"2026-05-18T12:32:13.499390+00:00"},{"alias_kind":"pith_short_8","alias_value":"AXOTLYSS","created_at":"2026-05-18T12:32:13.499390+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/AXOTLYSSDC3YUKPSZQUB5ZPXZA","json":"https://pith.science/pith/AXOTLYSSDC3YUKPSZQUB5ZPXZA.json","graph_json":"https://pith.science/api/pith-number/AXOTLYSSDC3YUKPSZQUB5ZPXZA/graph.json","events_json":"https://pith.science/api/pith-number/AXOTLYSSDC3YUKPSZQUB5ZPXZA/events.json","paper":"https://pith.science/paper/AXOTLYSS"},"agent_actions":{"view_html":"https://pith.science/pith/AXOTLYSSDC3YUKPSZQUB5ZPXZA","download_json":"https://pith.science/pith/AXOTLYSSDC3YUKPSZQUB5ZPXZA.json","view_paper":"https://pith.science/paper/AXOTLYSS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1803.10986&json=true","fetch_graph":"https://pith.science/api/pith-number/AXOTLYSSDC3YUKPSZQUB5ZPXZA/graph.json","fetch_events":"https://pith.science/api/pith-number/AXOTLYSSDC3YUKPSZQUB5ZPXZA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AXOTLYSSDC3YUKPSZQUB5ZPXZA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AXOTLYSSDC3YUKPSZQUB5ZPXZA/action/storage_attestation","attest_author":"https://pith.science/pith/AXOTLYSSDC3YUKPSZQUB5ZPXZA/action/author_attestation","sign_citation":"https://pith.science/pith/AXOTLYSSDC3YUKPSZQUB5ZPXZA/action/citation_signature","submit_replication":"https://pith.science/pith/AXOTLYSSDC3YUKPSZQUB5ZPXZA/action/replication_record"}},"created_at":"2026-05-17T23:47:12.692868+00:00","updated_at":"2026-05-17T23:47:12.692868+00:00"}