{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:5SUEQIWRMC7RXGVMCIBK23Z32X","short_pith_number":"pith:5SUEQIWR","canonical_record":{"source":{"id":"2210.05153","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-10-11T05:18:47Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f05a55a03c5493b4a365b72999759dc5fb526af7fdf1e62a06f578af279b8306","abstract_canon_sha256":"7b9c0899103b26df0c8e618b4ead79a8b2b7dc7e3636eea889596204d8d0d1de"},"schema_version":"1.0"},"canonical_sha256":"eca84822d160bf1b9aac1202ad6f3bd5e9b66bf7581ef4b00342d0d3bd661089","source":{"kind":"arxiv","id":"2210.05153","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2210.05153","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"arxiv_version","alias_value":"2210.05153v1","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.05153","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"pith_short_12","alias_value":"5SUEQIWRMC7R","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"pith_short_16","alias_value":"5SUEQIWRMC7RXGVM","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"pith_short_8","alias_value":"5SUEQIWR","created_at":"2026-07-05T05:06:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:5SUEQIWRMC7RXGVMCIBK23Z32X","target":"record","payload":{"canonical_record":{"source":{"id":"2210.05153","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-10-11T05:18:47Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f05a55a03c5493b4a365b72999759dc5fb526af7fdf1e62a06f578af279b8306","abstract_canon_sha256":"7b9c0899103b26df0c8e618b4ead79a8b2b7dc7e3636eea889596204d8d0d1de"},"schema_version":"1.0"},"canonical_sha256":"eca84822d160bf1b9aac1202ad6f3bd5e9b66bf7581ef4b00342d0d3bd661089","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:06:15.486234Z","signature_b64":"QXh8ZODPwCagazI+WJZLzsBkxztYPbQbHN3QFlwTQFhQg/H1MLnyHRqCqof8B8DCGXIDZxpIj7xH2CEBmP6JCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eca84822d160bf1b9aac1202ad6f3bd5e9b66bf7581ef4b00342d0d3bd661089","last_reissued_at":"2026-07-05T05:06:15.485646Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:06:15.485646Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2210.05153","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T05:06:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Nbkn9ypd0B3VRWP+MzdITTs9a0/j1fEBIIjMow9TmLCQZuubizgYZGmeGNTZC2QoE0ob3HLSIqXJsC6wpjOlBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-16T06:14:36.361244Z"},"content_sha256":"320d8b20351f6034adc106ed760b2d155c2432b52d33c5d5f9f0eee8dd779b5a","schema_version":"1.0","event_id":"sha256:320d8b20351f6034adc106ed760b2d155c2432b52d33c5d5f9f0eee8dd779b5a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:5SUEQIWRMC7RXGVMCIBK23Z32X","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Understanding the Failure of Batch Normalization for Transformers in NLP","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Jiaxi Wang, Ji Wu, Lei Huang","submitted_at":"2022-10-11T05:18:47Z","abstract_excerpt":"Batch Normalization (BN) is a core and prevalent technique in accelerating the training of deep neural networks and improving the generalization on Computer Vision (CV) tasks. However, it fails to defend its position in Natural Language Processing (NLP), which is dominated by Layer Normalization (LN). In this paper, we are trying to answer why BN usually performs worse than LN in NLP tasks with Transformer models. We find that the inconsistency between training and inference of BN is the leading cause that results in the failure of BN in NLP. We define Training Inference Discrepancy (TID) to q"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.05153","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/2210.05153/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":"2026-07-05T05:06:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/WQHuUeO67yRv+d0ZLq8FSMV5OCN+0CMcHON8zBdSDLWS8QVc9C6DfQmt6oS8flUYqXqjt9wfglBxKYmSeluCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-16T06:14:36.361810Z"},"content_sha256":"0450f1fa6e3ed940141175936a3d9a32b33e2ca451158f1fa83e4ec085b47bd9","schema_version":"1.0","event_id":"sha256:0450f1fa6e3ed940141175936a3d9a32b33e2ca451158f1fa83e4ec085b47bd9"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5SUEQIWRMC7RXGVMCIBK23Z32X/bundle.json","state_url":"https://pith.science/pith/5SUEQIWRMC7RXGVMCIBK23Z32X/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5SUEQIWRMC7RXGVMCIBK23Z32X/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":"2026-08-16T06:14:36Z","links":{"resolver":"https://pith.science/pith/5SUEQIWRMC7RXGVMCIBK23Z32X","bundle":"https://pith.science/pith/5SUEQIWRMC7RXGVMCIBK23Z32X/bundle.json","state":"https://pith.science/pith/5SUEQIWRMC7RXGVMCIBK23Z32X/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5SUEQIWRMC7RXGVMCIBK23Z32X/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:5SUEQIWRMC7RXGVMCIBK23Z32X","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":"7b9c0899103b26df0c8e618b4ead79a8b2b7dc7e3636eea889596204d8d0d1de","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-10-11T05:18:47Z","title_canon_sha256":"f05a55a03c5493b4a365b72999759dc5fb526af7fdf1e62a06f578af279b8306"},"schema_version":"1.0","source":{"id":"2210.05153","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2210.05153","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"arxiv_version","alias_value":"2210.05153v1","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.05153","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"pith_short_12","alias_value":"5SUEQIWRMC7R","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"pith_short_16","alias_value":"5SUEQIWRMC7RXGVM","created_at":"2026-07-05T05:06:15Z"},{"alias_kind":"pith_short_8","alias_value":"5SUEQIWR","created_at":"2026-07-05T05:06:15Z"}],"graph_snapshots":[{"event_id":"sha256:0450f1fa6e3ed940141175936a3d9a32b33e2ca451158f1fa83e4ec085b47bd9","target":"graph","created_at":"2026-07-05T05:06:15Z","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/2210.05153/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Batch Normalization (BN) is a core and prevalent technique in accelerating the training of deep neural networks and improving the generalization on Computer Vision (CV) tasks. However, it fails to defend its position in Natural Language Processing (NLP), which is dominated by Layer Normalization (LN). In this paper, we are trying to answer why BN usually performs worse than LN in NLP tasks with Transformer models. We find that the inconsistency between training and inference of BN is the leading cause that results in the failure of BN in NLP. We define Training Inference Discrepancy (TID) to q","authors_text":"Jiaxi Wang, Ji Wu, Lei Huang","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-10-11T05:18:47Z","title":"Understanding the Failure of Batch Normalization for Transformers in NLP"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.05153","kind":"arxiv","version":1},"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":"sha256:320d8b20351f6034adc106ed760b2d155c2432b52d33c5d5f9f0eee8dd779b5a","target":"record","created_at":"2026-07-05T05:06:15Z","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":"7b9c0899103b26df0c8e618b4ead79a8b2b7dc7e3636eea889596204d8d0d1de","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-10-11T05:18:47Z","title_canon_sha256":"f05a55a03c5493b4a365b72999759dc5fb526af7fdf1e62a06f578af279b8306"},"schema_version":"1.0","source":{"id":"2210.05153","kind":"arxiv","version":1}},"canonical_sha256":"eca84822d160bf1b9aac1202ad6f3bd5e9b66bf7581ef4b00342d0d3bd661089","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"eca84822d160bf1b9aac1202ad6f3bd5e9b66bf7581ef4b00342d0d3bd661089","first_computed_at":"2026-07-05T05:06:15.485646Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:06:15.485646Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"QXh8ZODPwCagazI+WJZLzsBkxztYPbQbHN3QFlwTQFhQg/H1MLnyHRqCqof8B8DCGXIDZxpIj7xH2CEBmP6JCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T05:06:15.486234Z","signed_message":"canonical_sha256_bytes"},"source_id":"2210.05153","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:320d8b20351f6034adc106ed760b2d155c2432b52d33c5d5f9f0eee8dd779b5a","sha256:0450f1fa6e3ed940141175936a3d9a32b33e2ca451158f1fa83e4ec085b47bd9"],"state_sha256":"0029b4106f7971544446b44fb30b4f4da40827e4af19b117cf4394ba3e512e3f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5fvaxoVM3MeWgOH/DwsvecpVoSbAHeIRzFVQ7LKVJGykWfltySAVradQOl5K+gRAbCUB15F/MUplH+Dvmt3GDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-16T06:14:36.368604Z","bundle_sha256":"cb098221937849337c3fed082a21862d8e0486c87091e91f39b451b9c8154405"}}