{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:ERQEPOA43UF36HFPG6J4DLEV67","short_pith_number":"pith:ERQEPOA4","schema_version":"1.0","canonical_sha256":"246047b81cdd0bbf1caf3793c1ac95f7e3ff0f63aa04460baaff5089ade290f1","source":{"kind":"arxiv","id":"2103.11795","version":1},"attestation_state":"computed","paper":{"title":"Simpson's Bias in NLP Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Fei Yuan, Huang Bojun, Longtu Zhang, Yaobo Liang","submitted_at":"2021-03-13T06:19:37Z","abstract_excerpt":"In most machine learning tasks, we evaluate a model $M$ on a given data population $S$ by measuring a population-level metric $F(S;M)$. Examples of such evaluation metric $F$ include precision/recall for (binary) recognition, the F1 score for multi-class classification, and the BLEU metric for language generation. On the other hand, the model $M$ is trained by optimizing a sample-level loss $G(S_t;M)$ at each learning step $t$, where $S_t$ is a subset of $S$ (a.k.a. the mini-batch). Popular choices of $G$ include cross-entropy loss, the Dice loss, and sentence-level BLEU scores. A fundamental "},"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":"2103.11795","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-03-13T06:19:37Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"ec88cc780366541610111b7115056f9fbcb93f5110af8e9c7ee7b498897ba4b4","abstract_canon_sha256":"4db3374832dd23ca4d3da25db27db849af6d769ebc7e43f3c450aa2f0c75efaa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:25:12.337049Z","signature_b64":"loNGeLvbMfYVEYBnJuAilhOAN+18tTJNRjYd/mGmYGHe7SxFqSLV7ivEl6JyANH2snsdLNZmXAPIQ44gVZ/uBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"246047b81cdd0bbf1caf3793c1ac95f7e3ff0f63aa04460baaff5089ade290f1","last_reissued_at":"2026-07-05T02:25:12.336656Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:25:12.336656Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Simpson's Bias in NLP Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Fei Yuan, Huang Bojun, Longtu Zhang, Yaobo Liang","submitted_at":"2021-03-13T06:19:37Z","abstract_excerpt":"In most machine learning tasks, we evaluate a model $M$ on a given data population $S$ by measuring a population-level metric $F(S;M)$. Examples of such evaluation metric $F$ include precision/recall for (binary) recognition, the F1 score for multi-class classification, and the BLEU metric for language generation. On the other hand, the model $M$ is trained by optimizing a sample-level loss $G(S_t;M)$ at each learning step $t$, where $S_t$ is a subset of $S$ (a.k.a. the mini-batch). Popular choices of $G$ include cross-entropy loss, the Dice loss, and sentence-level BLEU scores. A fundamental "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.11795","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/2103.11795/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":"2103.11795","created_at":"2026-07-05T02:25:12.336716+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.11795v1","created_at":"2026-07-05T02:25:12.336716+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.11795","created_at":"2026-07-05T02:25:12.336716+00:00"},{"alias_kind":"pith_short_12","alias_value":"ERQEPOA43UF3","created_at":"2026-07-05T02:25:12.336716+00:00"},{"alias_kind":"pith_short_16","alias_value":"ERQEPOA43UF36HFP","created_at":"2026-07-05T02:25:12.336716+00:00"},{"alias_kind":"pith_short_8","alias_value":"ERQEPOA4","created_at":"2026-07-05T02:25:12.336716+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/ERQEPOA43UF36HFPG6J4DLEV67","json":"https://pith.science/pith/ERQEPOA43UF36HFPG6J4DLEV67.json","graph_json":"https://pith.science/api/pith-number/ERQEPOA43UF36HFPG6J4DLEV67/graph.json","events_json":"https://pith.science/api/pith-number/ERQEPOA43UF36HFPG6J4DLEV67/events.json","paper":"https://pith.science/paper/ERQEPOA4"},"agent_actions":{"view_html":"https://pith.science/pith/ERQEPOA43UF36HFPG6J4DLEV67","download_json":"https://pith.science/pith/ERQEPOA43UF36HFPG6J4DLEV67.json","view_paper":"https://pith.science/paper/ERQEPOA4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.11795&json=true","fetch_graph":"https://pith.science/api/pith-number/ERQEPOA43UF36HFPG6J4DLEV67/graph.json","fetch_events":"https://pith.science/api/pith-number/ERQEPOA43UF36HFPG6J4DLEV67/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ERQEPOA43UF36HFPG6J4DLEV67/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ERQEPOA43UF36HFPG6J4DLEV67/action/storage_attestation","attest_author":"https://pith.science/pith/ERQEPOA43UF36HFPG6J4DLEV67/action/author_attestation","sign_citation":"https://pith.science/pith/ERQEPOA43UF36HFPG6J4DLEV67/action/citation_signature","submit_replication":"https://pith.science/pith/ERQEPOA43UF36HFPG6J4DLEV67/action/replication_record"}},"created_at":"2026-07-05T02:25:12.336716+00:00","updated_at":"2026-07-05T02:25:12.336716+00:00"}