{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QYUGG377F4HOOKGFAIMKSGDIS5","short_pith_number":"pith:QYUGG377","schema_version":"1.0","canonical_sha256":"8628636fff2f0ee728c50218a9186897453540641735a973c43d02014616875a","source":{"kind":"arxiv","id":"2602.23171","version":2},"attestation_state":"computed","paper":{"title":"Align-Consistency: Improving Non-autoregressive and Semi-supervised ASR with Consistency Regularization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"eess.AS","authors_text":"Wanting Huang, Weiran Wang","submitted_at":"2026-02-26T16:34:30Z","abstract_excerpt":"Consistency regularization (CR) improves the robustness and accuracy of Connectionist Temporal Classification (CTC) by ensuring predictions remain stable across input perturbations. In this work, we propose Align-Consistency, an extension of CR designed for Align-Refine -- a non-autoregressive (non-AR) model that performs iterative refinement of frame-level hypotheses. This method leverages the speed of parallel inference while significantly boosting recognition performance. The effectiveness of Align-Consistency is demonstrated in two settings. First, in the fully supervised setting, our resu"},"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":"2602.23171","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.AS","submitted_at":"2026-02-26T16:34:30Z","cross_cats_sorted":[],"title_canon_sha256":"c160270a227d0c1789596b18b9fbae3ac9144e0c183383024000a6e4a744556e","abstract_canon_sha256":"cb2f153f83f4fc98181825c9b0d00206138ad0ef16465f5871b9bf1664d592a3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T01:16:28.612627Z","signature_b64":"qJt5KUIyKQVTg3e/rvPWQHBGWsBFzCvZtMDJNjFu8Iq9dtxUU0k0kqP85Eh/U1Y70PJsrtaE7DwfRVTXx53KAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8628636fff2f0ee728c50218a9186897453540641735a973c43d02014616875a","last_reissued_at":"2026-06-30T01:16:28.611927Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T01:16:28.611927Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Align-Consistency: Improving Non-autoregressive and Semi-supervised ASR with Consistency Regularization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"eess.AS","authors_text":"Wanting Huang, Weiran Wang","submitted_at":"2026-02-26T16:34:30Z","abstract_excerpt":"Consistency regularization (CR) improves the robustness and accuracy of Connectionist Temporal Classification (CTC) by ensuring predictions remain stable across input perturbations. In this work, we propose Align-Consistency, an extension of CR designed for Align-Refine -- a non-autoregressive (non-AR) model that performs iterative refinement of frame-level hypotheses. This method leverages the speed of parallel inference while significantly boosting recognition performance. The effectiveness of Align-Consistency is demonstrated in two settings. First, in the fully supervised setting, our resu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.23171","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/2602.23171/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":"2602.23171","created_at":"2026-06-30T01:16:28.612009+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.23171v2","created_at":"2026-06-30T01:16:28.612009+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.23171","created_at":"2026-06-30T01:16:28.612009+00:00"},{"alias_kind":"pith_short_12","alias_value":"QYUGG377F4HO","created_at":"2026-06-30T01:16:28.612009+00:00"},{"alias_kind":"pith_short_16","alias_value":"QYUGG377F4HOOKGF","created_at":"2026-06-30T01:16:28.612009+00:00"},{"alias_kind":"pith_short_8","alias_value":"QYUGG377","created_at":"2026-06-30T01:16:28.612009+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/QYUGG377F4HOOKGFAIMKSGDIS5","json":"https://pith.science/pith/QYUGG377F4HOOKGFAIMKSGDIS5.json","graph_json":"https://pith.science/api/pith-number/QYUGG377F4HOOKGFAIMKSGDIS5/graph.json","events_json":"https://pith.science/api/pith-number/QYUGG377F4HOOKGFAIMKSGDIS5/events.json","paper":"https://pith.science/paper/QYUGG377"},"agent_actions":{"view_html":"https://pith.science/pith/QYUGG377F4HOOKGFAIMKSGDIS5","download_json":"https://pith.science/pith/QYUGG377F4HOOKGFAIMKSGDIS5.json","view_paper":"https://pith.science/paper/QYUGG377","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.23171&json=true","fetch_graph":"https://pith.science/api/pith-number/QYUGG377F4HOOKGFAIMKSGDIS5/graph.json","fetch_events":"https://pith.science/api/pith-number/QYUGG377F4HOOKGFAIMKSGDIS5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QYUGG377F4HOOKGFAIMKSGDIS5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QYUGG377F4HOOKGFAIMKSGDIS5/action/storage_attestation","attest_author":"https://pith.science/pith/QYUGG377F4HOOKGFAIMKSGDIS5/action/author_attestation","sign_citation":"https://pith.science/pith/QYUGG377F4HOOKGFAIMKSGDIS5/action/citation_signature","submit_replication":"https://pith.science/pith/QYUGG377F4HOOKGFAIMKSGDIS5/action/replication_record"}},"created_at":"2026-06-30T01:16:28.612009+00:00","updated_at":"2026-06-30T01:16:28.612009+00:00"}