{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:CQT3PEMA75I22VZJ3C2JNLQOCC","short_pith_number":"pith:CQT3PEMA","schema_version":"1.0","canonical_sha256":"1427b79180ff51ad5729d8b496ae0e10bab456cef7373cb03b2f0f360b07037f","source":{"kind":"arxiv","id":"1904.09324","version":2},"attestation_state":"computed","paper":{"title":"Mask-Predict: Parallel Decoding of Conditional Masked Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Luke Zettlemoyer, Marjan Ghazvininejad, Omer Levy, Yinhan Liu","submitted_at":"2019-04-19T19:53:01Z","abstract_excerpt":"Most machine translation systems generate text autoregressively from left to right. We, instead, use a masked language modeling objective to train a model to predict any subset of the target words, conditioned on both the input text and a partially masked target translation. This approach allows for efficient iterative decoding, where we first predict all of the target words non-autoregressively, and then repeatedly mask out and regenerate the subset of words that the model is least confident about. By applying this strategy for a constant number of iterations, our model improves state-of-the-"},"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":"1904.09324","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-04-19T19:53:01Z","cross_cats_sorted":["cs.AI","cs.LG","stat.ML"],"title_canon_sha256":"fef38f3f17051fddbf9b301a37da4004a58d5bd2bd188aa0bbb7f815205b7045","abstract_canon_sha256":"e8496a28946971d90709e34928893d6fcf96d5d329f3faa26370ef488d4e27bc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:02:13.200274Z","signature_b64":"Oyri98UcMOYWWDODbsArSX/6URU9kYLw116t5S/DlJ2qvgbR70/JhpYUex/bbNwKmbzFHOUeBsGXpyplCII+Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1427b79180ff51ad5729d8b496ae0e10bab456cef7373cb03b2f0f360b07037f","last_reissued_at":"2026-07-05T00:02:13.199812Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:02:13.199812Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mask-Predict: Parallel Decoding of Conditional Masked Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Luke Zettlemoyer, Marjan Ghazvininejad, Omer Levy, Yinhan Liu","submitted_at":"2019-04-19T19:53:01Z","abstract_excerpt":"Most machine translation systems generate text autoregressively from left to right. We, instead, use a masked language modeling objective to train a model to predict any subset of the target words, conditioned on both the input text and a partially masked target translation. This approach allows for efficient iterative decoding, where we first predict all of the target words non-autoregressively, and then repeatedly mask out and regenerate the subset of words that the model is least confident about. By applying this strategy for a constant number of iterations, our model improves state-of-the-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1904.09324","kind":"arxiv","version":2},"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/1904.09324/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":"1904.09324","created_at":"2026-07-05T00:02:13.199868+00:00"},{"alias_kind":"arxiv_version","alias_value":"1904.09324v2","created_at":"2026-07-05T00:02:13.199868+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1904.09324","created_at":"2026-07-05T00:02:13.199868+00:00"},{"alias_kind":"pith_short_12","alias_value":"CQT3PEMA75I2","created_at":"2026-07-05T00:02:13.199868+00:00"},{"alias_kind":"pith_short_16","alias_value":"CQT3PEMA75I22VZJ","created_at":"2026-07-05T00:02:13.199868+00:00"},{"alias_kind":"pith_short_8","alias_value":"CQT3PEMA","created_at":"2026-07-05T00:02:13.199868+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.06547","citing_title":"FAIR-Calib: Frontier-Aware Instability-Reweighted Calibration for Post-Training Quantization of Diffusion Large Language Models","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2602.16169","citing_title":"Discrete Stochastic Localization for Non-autoregressive Generation","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2508.02193","citing_title":"Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14567","citing_title":"Scaling Laws from Sequential Feature Recovery: A Solvable Hierarchical Model","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22906","citing_title":"Network Edge Inference for Large Language Models: Principles, Techniques, and Opportunities","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15009","citing_title":"Towards Faster Language Model Inference Using Mixture-of-Experts Flow Matching","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC","json":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC.json","graph_json":"https://pith.science/api/pith-number/CQT3PEMA75I22VZJ3C2JNLQOCC/graph.json","events_json":"https://pith.science/api/pith-number/CQT3PEMA75I22VZJ3C2JNLQOCC/events.json","paper":"https://pith.science/paper/CQT3PEMA"},"agent_actions":{"view_html":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC","download_json":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC.json","view_paper":"https://pith.science/paper/CQT3PEMA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1904.09324&json=true","fetch_graph":"https://pith.science/api/pith-number/CQT3PEMA75I22VZJ3C2JNLQOCC/graph.json","fetch_events":"https://pith.science/api/pith-number/CQT3PEMA75I22VZJ3C2JNLQOCC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC/action/storage_attestation","attest_author":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC/action/author_attestation","sign_citation":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC/action/citation_signature","submit_replication":"https://pith.science/pith/CQT3PEMA75I22VZJ3C2JNLQOCC/action/replication_record"}},"created_at":"2026-07-05T00:02:13.199868+00:00","updated_at":"2026-07-05T00:02:13.199868+00:00"}