{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:BPROKP4IQW2PZDBWPAAUIMDQJC","short_pith_number":"pith:BPROKP4I","schema_version":"1.0","canonical_sha256":"0be2e53f8885b4fc8c36780144307048a64ce0a88ee0d59f4dfeed66325fb2fb","source":{"kind":"arxiv","id":"2008.09371","version":3},"attestation_state":"computed","paper":{"title":"Towards Improving Selective Prediction Ability of NLP Systems","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Chitta Baral, Neeraj Varshney, Swaroop Mishra","submitted_at":"2020-08-21T08:46:36Z","abstract_excerpt":"It's better to say \"I can't answer\" than to answer incorrectly. This selective prediction ability is crucial for NLP systems to be reliably deployed in real-world applications. Prior work has shown that existing selective prediction techniques fail to perform well, especially in the out-of-domain setting. In this work, we propose a method that improves probability estimates of models by calibrating them using prediction confidence and difficulty score of instances. Using these two signals, we first annotate held-out instances and then train a calibrator to predict the likelihood of correctness"},"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":"2008.09371","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2020-08-21T08:46:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"fd6d4624b6d7006f58ccc1c64d31a4baa433d722ab46ecd9114d0d5f0a82b482","abstract_canon_sha256":"3a138a970f30f09142decececffdb0d783c533def121354e077cf769e03eef44"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:12:14.571286Z","signature_b64":"Ee3Hz+6oWd1tul6yzSmEAK49Qrbxqo8GE+PK/4gxt+GqXR+BihE1KzSb0ioOTnI8KvdAFb5b1bab7NKzBJfwAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0be2e53f8885b4fc8c36780144307048a64ce0a88ee0d59f4dfeed66325fb2fb","last_reissued_at":"2026-07-05T04:12:14.570754Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:12:14.570754Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Improving Selective Prediction Ability of NLP Systems","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Chitta Baral, Neeraj Varshney, Swaroop Mishra","submitted_at":"2020-08-21T08:46:36Z","abstract_excerpt":"It's better to say \"I can't answer\" than to answer incorrectly. This selective prediction ability is crucial for NLP systems to be reliably deployed in real-world applications. Prior work has shown that existing selective prediction techniques fail to perform well, especially in the out-of-domain setting. In this work, we propose a method that improves probability estimates of models by calibrating them using prediction confidence and difficulty score of instances. Using these two signals, we first annotate held-out instances and then train a calibrator to predict the likelihood of correctness"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2008.09371","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2008.09371/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":"2008.09371","created_at":"2026-07-05T04:12:14.570812+00:00"},{"alias_kind":"arxiv_version","alias_value":"2008.09371v3","created_at":"2026-07-05T04:12:14.570812+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2008.09371","created_at":"2026-07-05T04:12:14.570812+00:00"},{"alias_kind":"pith_short_12","alias_value":"BPROKP4IQW2P","created_at":"2026-07-05T04:12:14.570812+00:00"},{"alias_kind":"pith_short_16","alias_value":"BPROKP4IQW2PZDBW","created_at":"2026-07-05T04:12:14.570812+00:00"},{"alias_kind":"pith_short_8","alias_value":"BPROKP4I","created_at":"2026-07-05T04:12:14.570812+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.20490","citing_title":"ECUAS$_n$: A family of metrics for principled evaluation of uncertainty-augmented systems","ref_index":87,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20490","citing_title":"ECUAS$_n$: A family of metrics for principled evaluation of uncertainty-augmented systems","ref_index":87,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20490","citing_title":"ECUAS$_n$: A family of metrics for principled evaluation of uncertainty-augmented systems","ref_index":87,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC","json":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC.json","graph_json":"https://pith.science/api/pith-number/BPROKP4IQW2PZDBWPAAUIMDQJC/graph.json","events_json":"https://pith.science/api/pith-number/BPROKP4IQW2PZDBWPAAUIMDQJC/events.json","paper":"https://pith.science/paper/BPROKP4I"},"agent_actions":{"view_html":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC","download_json":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC.json","view_paper":"https://pith.science/paper/BPROKP4I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2008.09371&json=true","fetch_graph":"https://pith.science/api/pith-number/BPROKP4IQW2PZDBWPAAUIMDQJC/graph.json","fetch_events":"https://pith.science/api/pith-number/BPROKP4IQW2PZDBWPAAUIMDQJC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC/action/storage_attestation","attest_author":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC/action/author_attestation","sign_citation":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC/action/citation_signature","submit_replication":"https://pith.science/pith/BPROKP4IQW2PZDBWPAAUIMDQJC/action/replication_record"}},"created_at":"2026-07-05T04:12:14.570812+00:00","updated_at":"2026-07-05T04:12:14.570812+00:00"}