{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:YVDUN5OATY4HAQZGIHD7P43DGG","short_pith_number":"pith:YVDUN5OA","schema_version":"1.0","canonical_sha256":"c54746f5c09e3870432641c7f7f36331910d2eb05a96c26488ee0d9f38385c31","source":{"kind":"arxiv","id":"2109.01754","version":1},"attestation_state":"computed","paper":{"title":"Error Detection in Large-Scale Natural Language Understanding Systems Using Transformer Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Darshan Fofadiya, Pradeep Natarajan, Prathap Ramachandra, Rakesh Chada","submitted_at":"2021-09-04T00:10:48Z","abstract_excerpt":"Large-scale conversational assistants like Alexa, Siri, Cortana and Google Assistant process every utterance using multiple models for domain, intent and named entity recognition. Given the decoupled nature of model development and large traffic volumes, it is extremely difficult to identify utterances processed erroneously by such systems. We address this challenge to detect domain classification errors using offline Transformer models. We combine utterance encodings from a RoBERTa model with the Nbest hypothesis produced by the production system. We then fine-tune end-to-end in a multitask s"},"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":"2109.01754","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-09-04T00:10:48Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"c768f2256a56880817598eb996af9da8b3f1bd542f053bd8ceb6b21aba961eec","abstract_canon_sha256":"2f33922f906ae85b32c8e8bcfddac2017beb283a9d905287905eadbe5fb24631"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:11:34.478842Z","signature_b64":"7HvEpUSyVfesm63K90XMyy/PZcfCUnfEq/DLaFhPdlaoP6nhpQZDMmcYyC9ZQf3gYUFWQQ4D5ErVKi4dMNMVBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c54746f5c09e3870432641c7f7f36331910d2eb05a96c26488ee0d9f38385c31","last_reissued_at":"2026-07-05T03:11:34.478503Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:11:34.478503Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Error Detection in Large-Scale Natural Language Understanding Systems Using Transformer Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Darshan Fofadiya, Pradeep Natarajan, Prathap Ramachandra, Rakesh Chada","submitted_at":"2021-09-04T00:10:48Z","abstract_excerpt":"Large-scale conversational assistants like Alexa, Siri, Cortana and Google Assistant process every utterance using multiple models for domain, intent and named entity recognition. Given the decoupled nature of model development and large traffic volumes, it is extremely difficult to identify utterances processed erroneously by such systems. We address this challenge to detect domain classification errors using offline Transformer models. We combine utterance encodings from a RoBERTa model with the Nbest hypothesis produced by the production system. We then fine-tune end-to-end in a multitask s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.01754","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/2109.01754/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":"2109.01754","created_at":"2026-07-05T03:11:34.478560+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.01754v1","created_at":"2026-07-05T03:11:34.478560+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.01754","created_at":"2026-07-05T03:11:34.478560+00:00"},{"alias_kind":"pith_short_12","alias_value":"YVDUN5OATY4H","created_at":"2026-07-05T03:11:34.478560+00:00"},{"alias_kind":"pith_short_16","alias_value":"YVDUN5OATY4HAQZG","created_at":"2026-07-05T03:11:34.478560+00:00"},{"alias_kind":"pith_short_8","alias_value":"YVDUN5OA","created_at":"2026-07-05T03:11:34.478560+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/YVDUN5OATY4HAQZGIHD7P43DGG","json":"https://pith.science/pith/YVDUN5OATY4HAQZGIHD7P43DGG.json","graph_json":"https://pith.science/api/pith-number/YVDUN5OATY4HAQZGIHD7P43DGG/graph.json","events_json":"https://pith.science/api/pith-number/YVDUN5OATY4HAQZGIHD7P43DGG/events.json","paper":"https://pith.science/paper/YVDUN5OA"},"agent_actions":{"view_html":"https://pith.science/pith/YVDUN5OATY4HAQZGIHD7P43DGG","download_json":"https://pith.science/pith/YVDUN5OATY4HAQZGIHD7P43DGG.json","view_paper":"https://pith.science/paper/YVDUN5OA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.01754&json=true","fetch_graph":"https://pith.science/api/pith-number/YVDUN5OATY4HAQZGIHD7P43DGG/graph.json","fetch_events":"https://pith.science/api/pith-number/YVDUN5OATY4HAQZGIHD7P43DGG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YVDUN5OATY4HAQZGIHD7P43DGG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YVDUN5OATY4HAQZGIHD7P43DGG/action/storage_attestation","attest_author":"https://pith.science/pith/YVDUN5OATY4HAQZGIHD7P43DGG/action/author_attestation","sign_citation":"https://pith.science/pith/YVDUN5OATY4HAQZGIHD7P43DGG/action/citation_signature","submit_replication":"https://pith.science/pith/YVDUN5OATY4HAQZGIHD7P43DGG/action/replication_record"}},"created_at":"2026-07-05T03:11:34.478560+00:00","updated_at":"2026-07-05T03:11:34.478560+00:00"}