{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:5QV2IXETCG2TOSNMTH3SCQKUMD","short_pith_number":"pith:5QV2IXET","canonical_record":{"source":{"id":"2205.09615","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-05-19T15:13:00Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"f160fe8baf543d915c490a2a73595dbd985892e25ed34a9e16effe4ecb9acfca","abstract_canon_sha256":"0c67379b473becc5443bd56c5a5a87c70400fc9fe61b7a8f5a9e2ec53ce79f70"},"schema_version":"1.0"},"canonical_sha256":"ec2ba45c9311b53749ac99f721415460f22952fcb0048ec33a4eeeb4b8f09e59","source":{"kind":"arxiv","id":"2205.09615","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2205.09615","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"arxiv_version","alias_value":"2205.09615v5","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.09615","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"pith_short_12","alias_value":"5QV2IXETCG2T","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"pith_short_16","alias_value":"5QV2IXETCG2TOSNM","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"pith_short_8","alias_value":"5QV2IXET","created_at":"2026-07-05T08:47:44Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:5QV2IXETCG2TOSNMTH3SCQKUMD","target":"record","payload":{"canonical_record":{"source":{"id":"2205.09615","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-05-19T15:13:00Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"f160fe8baf543d915c490a2a73595dbd985892e25ed34a9e16effe4ecb9acfca","abstract_canon_sha256":"0c67379b473becc5443bd56c5a5a87c70400fc9fe61b7a8f5a9e2ec53ce79f70"},"schema_version":"1.0"},"canonical_sha256":"ec2ba45c9311b53749ac99f721415460f22952fcb0048ec33a4eeeb4b8f09e59","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:47:44.001097Z","signature_b64":"f2fmQzFfGBc3MGd2ctOWRYn7InDT/kah4isZxHPMvxp5+hk/wMtSVd1gtTENAc2OcgXeSG0439kwuf6Rs83OAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ec2ba45c9311b53749ac99f721415460f22952fcb0048ec33a4eeeb4b8f09e59","last_reissued_at":"2026-07-05T08:47:44.000641Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:47:44.000641Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2205.09615","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:47:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2C+0arLCeChRvzApq8685coctRSanhKI5i6tow1w/3GOaabxEwf8eHpA5GfB0zZcntGncWqDdp9Nn7VH66iQBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-09T02:50:12.137503Z"},"content_sha256":"af38ea9757dee12fbb8f58ec08fcc184011700c7da277d48e2b5f2905d2e422c","schema_version":"1.0","event_id":"sha256:af38ea9757dee12fbb8f58ec08fcc184011700c7da277d48e2b5f2905d2e422c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:5QV2IXETCG2TOSNMTH3SCQKUMD","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"EXACT: How to Train Your Accuracy","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Ivan Karpukhin, Sergey Kolesnikov, Stanislav Dereka","submitted_at":"2022-05-19T15:13:00Z","abstract_excerpt":"Classification tasks are usually evaluated in terms of accuracy. However, accuracy is discontinuous and cannot be directly optimized using gradient ascent. Popular methods minimize cross-entropy, hinge loss, or other surrogate losses, which can lead to suboptimal results. In this paper, we propose a new optimization framework by introducing stochasticity to a model's output and optimizing expected accuracy, i.e. accuracy of the stochastic model. Extensive experiments on linear models and deep image classification show that the proposed optimization method is a powerful alternative to widely us"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.09615","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2205.09615/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:47:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oZjvaQr1fnsKRkLPvuQ8/wMiffs5tTILTKXUlGjT4PwreeIQVp3pMgePiDJObw22CzQXk+jMKYdKdv2zzT5vDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-09T02:50:12.137877Z"},"content_sha256":"4d426ac03d6ab7a04a94b8d7328706af902f06c9fea6e7f5c4911a813325a4b5","schema_version":"1.0","event_id":"sha256:4d426ac03d6ab7a04a94b8d7328706af902f06c9fea6e7f5c4911a813325a4b5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5QV2IXETCG2TOSNMTH3SCQKUMD/bundle.json","state_url":"https://pith.science/pith/5QV2IXETCG2TOSNMTH3SCQKUMD/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5QV2IXETCG2TOSNMTH3SCQKUMD/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-09T02:50:12Z","links":{"resolver":"https://pith.science/pith/5QV2IXETCG2TOSNMTH3SCQKUMD","bundle":"https://pith.science/pith/5QV2IXETCG2TOSNMTH3SCQKUMD/bundle.json","state":"https://pith.science/pith/5QV2IXETCG2TOSNMTH3SCQKUMD/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5QV2IXETCG2TOSNMTH3SCQKUMD/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:5QV2IXETCG2TOSNMTH3SCQKUMD","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"0c67379b473becc5443bd56c5a5a87c70400fc9fe61b7a8f5a9e2ec53ce79f70","cross_cats_sorted":["cs.CV"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-05-19T15:13:00Z","title_canon_sha256":"f160fe8baf543d915c490a2a73595dbd985892e25ed34a9e16effe4ecb9acfca"},"schema_version":"1.0","source":{"id":"2205.09615","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2205.09615","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"arxiv_version","alias_value":"2205.09615v5","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.09615","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"pith_short_12","alias_value":"5QV2IXETCG2T","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"pith_short_16","alias_value":"5QV2IXETCG2TOSNM","created_at":"2026-07-05T08:47:44Z"},{"alias_kind":"pith_short_8","alias_value":"5QV2IXET","created_at":"2026-07-05T08:47:44Z"}],"graph_snapshots":[{"event_id":"sha256:4d426ac03d6ab7a04a94b8d7328706af902f06c9fea6e7f5c4911a813325a4b5","target":"graph","created_at":"2026-07-05T08:47:44Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2205.09615/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Classification tasks are usually evaluated in terms of accuracy. However, accuracy is discontinuous and cannot be directly optimized using gradient ascent. Popular methods minimize cross-entropy, hinge loss, or other surrogate losses, which can lead to suboptimal results. In this paper, we propose a new optimization framework by introducing stochasticity to a model's output and optimizing expected accuracy, i.e. accuracy of the stochastic model. Extensive experiments on linear models and deep image classification show that the proposed optimization method is a powerful alternative to widely us","authors_text":"Ivan Karpukhin, Sergey Kolesnikov, Stanislav Dereka","cross_cats":["cs.CV"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-05-19T15:13:00Z","title":"EXACT: How to Train Your Accuracy"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.09615","kind":"arxiv","version":5},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:af38ea9757dee12fbb8f58ec08fcc184011700c7da277d48e2b5f2905d2e422c","target":"record","created_at":"2026-07-05T08:47:44Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"0c67379b473becc5443bd56c5a5a87c70400fc9fe61b7a8f5a9e2ec53ce79f70","cross_cats_sorted":["cs.CV"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-05-19T15:13:00Z","title_canon_sha256":"f160fe8baf543d915c490a2a73595dbd985892e25ed34a9e16effe4ecb9acfca"},"schema_version":"1.0","source":{"id":"2205.09615","kind":"arxiv","version":5}},"canonical_sha256":"ec2ba45c9311b53749ac99f721415460f22952fcb0048ec33a4eeeb4b8f09e59","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ec2ba45c9311b53749ac99f721415460f22952fcb0048ec33a4eeeb4b8f09e59","first_computed_at":"2026-07-05T08:47:44.000641Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:47:44.000641Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"f2fmQzFfGBc3MGd2ctOWRYn7InDT/kah4isZxHPMvxp5+hk/wMtSVd1gtTENAc2OcgXeSG0439kwuf6Rs83OAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:47:44.001097Z","signed_message":"canonical_sha256_bytes"},"source_id":"2205.09615","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:af38ea9757dee12fbb8f58ec08fcc184011700c7da277d48e2b5f2905d2e422c","sha256:4d426ac03d6ab7a04a94b8d7328706af902f06c9fea6e7f5c4911a813325a4b5"],"state_sha256":"6946ab9d9e95040947013857b207b9beca9ffb272acfe88febfe72965cd8b2fd"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CUfALCf+KfPokbRvUAWxuNBKydvQF6/InCGcV4nE7ZMsX8bHPkdF1fQ0t16dHnrDyI6Q23c3Ttq2zvYjPZqZCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-09T02:50:12.140115Z","bundle_sha256":"b75ffebb4bbfe327c0d17fcd6dc9a69bc7cb1afb9df30b12e8ac93dbd8f52767"}}