{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:ETGZN7JDX2L6ECCDBBO74BB7FX","short_pith_number":"pith:ETGZN7JD","schema_version":"1.0","canonical_sha256":"24cd96fd23be97e20843085dfe043f2de211eeb2528e2afc0091edfc5a42711c","source":{"kind":"arxiv","id":"1710.03743","version":1},"attestation_state":"computed","paper":{"title":"Confidence through Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Mark Fishel, Mat\\=iss Rikters","submitted_at":"2017-10-10T17:47:41Z","abstract_excerpt":"Attention distributions of the generated translations are a useful bi-product of attention-based recurrent neural network translation models and can be treated as soft alignments between the input and output tokens. In this work, we use attention distributions as a confidence metric for output translations. We present two strategies of using the attention distributions: filtering out bad translations from a large back-translated corpus, and selecting the best translation in a hybrid setup of two different translation systems. While manual evaluation indicated only a weak correlation between ou"},"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":"1710.03743","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2017-10-10T17:47:41Z","cross_cats_sorted":[],"title_canon_sha256":"cf30c8a754612467dfcf860c1c1e128cd93ceb5a458f9385e8549f65b8a51007","abstract_canon_sha256":"58a79bd0d1a3436d199f7575dfbcad9ef779748912c6b145c08b5327ea93c2b3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:33:10.823973Z","signature_b64":"AYx1vLFchI8837gKW2fko5WkKzKmk08WNrU+bVFbwqtCSezKfTEjptcxA6S6xdJZK5fLqWOSAR5OTuXU8wUDAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"24cd96fd23be97e20843085dfe043f2de211eeb2528e2afc0091edfc5a42711c","last_reissued_at":"2026-05-18T00:33:10.823480Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:33:10.823480Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Confidence through Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Mark Fishel, Mat\\=iss Rikters","submitted_at":"2017-10-10T17:47:41Z","abstract_excerpt":"Attention distributions of the generated translations are a useful bi-product of attention-based recurrent neural network translation models and can be treated as soft alignments between the input and output tokens. In this work, we use attention distributions as a confidence metric for output translations. We present two strategies of using the attention distributions: filtering out bad translations from a large back-translated corpus, and selecting the best translation in a hybrid setup of two different translation systems. While manual evaluation indicated only a weak correlation between ou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1710.03743","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":""},"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":"1710.03743","created_at":"2026-05-18T00:33:10.823555+00:00"},{"alias_kind":"arxiv_version","alias_value":"1710.03743v1","created_at":"2026-05-18T00:33:10.823555+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1710.03743","created_at":"2026-05-18T00:33:10.823555+00:00"},{"alias_kind":"pith_short_12","alias_value":"ETGZN7JDX2L6","created_at":"2026-05-18T12:31:12.930513+00:00"},{"alias_kind":"pith_short_16","alias_value":"ETGZN7JDX2L6ECCD","created_at":"2026-05-18T12:31:12.930513+00:00"},{"alias_kind":"pith_short_8","alias_value":"ETGZN7JD","created_at":"2026-05-18T12:31:12.930513+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"1909.00157","citing_title":"Improving Back-Translation with Uncertainty-based Confidence Estimation","ref_index":41,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX","json":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX.json","graph_json":"https://pith.science/api/pith-number/ETGZN7JDX2L6ECCDBBO74BB7FX/graph.json","events_json":"https://pith.science/api/pith-number/ETGZN7JDX2L6ECCDBBO74BB7FX/events.json","paper":"https://pith.science/paper/ETGZN7JD"},"agent_actions":{"view_html":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX","download_json":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX.json","view_paper":"https://pith.science/paper/ETGZN7JD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1710.03743&json=true","fetch_graph":"https://pith.science/api/pith-number/ETGZN7JDX2L6ECCDBBO74BB7FX/graph.json","fetch_events":"https://pith.science/api/pith-number/ETGZN7JDX2L6ECCDBBO74BB7FX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX/action/storage_attestation","attest_author":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX/action/author_attestation","sign_citation":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX/action/citation_signature","submit_replication":"https://pith.science/pith/ETGZN7JDX2L6ECCDBBO74BB7FX/action/replication_record"}},"created_at":"2026-05-18T00:33:10.823555+00:00","updated_at":"2026-05-18T00:33:10.823555+00:00"}