{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:75ADAVTPSQTRYYKAPR5LJ2LLMV","short_pith_number":"pith:75ADAVTP","schema_version":"1.0","canonical_sha256":"ff4030566f94271c61407c7ab4e96b657e168ab0b6fedf8b16fde90f19147a34","source":{"kind":"arxiv","id":"2505.24820","version":1},"attestation_state":"computed","paper":{"title":"Masked Self-distilled Transducer-based Keyword Spotting with Semi-autoregressive Decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.SD","authors_text":"Bo Zheng, Haoyu Li, Jun Song, Kai Yu, Xiaoyu Gu, Yu Xi","submitted_at":"2025-05-30T17:22:39Z","abstract_excerpt":"RNN-T-based keyword spotting (KWS) with autoregressive decoding~(AR) has gained attention due to its streaming architecture and superior performance. However, the simplicity of the prediction network in RNN-T poses an overfitting issue, especially under challenging scenarios, resulting in degraded performance. In this paper, we propose a masked self-distillation (MSD) training strategy that avoids RNN-Ts overly relying on prediction networks to alleviate overfitting. Such training enables masked non-autoregressive (NAR) decoding, which fully masks the RNN-T predictor output during KWS decoding"},"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":"2505.24820","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2025-05-30T17:22:39Z","cross_cats_sorted":["eess.AS"],"title_canon_sha256":"93cc6a08416e1230ad9ebf4b86b02b64c223e1ddfda9851bb322eb5b1845a618","abstract_canon_sha256":"fc1cb45d3a15328b432355e74fdefaa6f8431aca610b255c87d997a7e70a9de9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:57.379965Z","signature_b64":"5VDge9v0zZoXAUn3271YsWiAZFbZZU+hZiG/5o8ME82pmtqsUQpww2KZTUXe6fdKhw3N1hdA5f5SMriu1QbzAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff4030566f94271c61407c7ab4e96b657e168ab0b6fedf8b16fde90f19147a34","last_reissued_at":"2026-07-05T11:12:57.379498Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:57.379498Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Masked Self-distilled Transducer-based Keyword Spotting with Semi-autoregressive Decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.SD","authors_text":"Bo Zheng, Haoyu Li, Jun Song, Kai Yu, Xiaoyu Gu, Yu Xi","submitted_at":"2025-05-30T17:22:39Z","abstract_excerpt":"RNN-T-based keyword spotting (KWS) with autoregressive decoding~(AR) has gained attention due to its streaming architecture and superior performance. However, the simplicity of the prediction network in RNN-T poses an overfitting issue, especially under challenging scenarios, resulting in degraded performance. In this paper, we propose a masked self-distillation (MSD) training strategy that avoids RNN-Ts overly relying on prediction networks to alleviate overfitting. Such training enables masked non-autoregressive (NAR) decoding, which fully masks the RNN-T predictor output during KWS decoding"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.24820","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/2505.24820/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":"2505.24820","created_at":"2026-07-05T11:12:57.379562+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.24820v1","created_at":"2026-07-05T11:12:57.379562+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.24820","created_at":"2026-07-05T11:12:57.379562+00:00"},{"alias_kind":"pith_short_12","alias_value":"75ADAVTPSQTR","created_at":"2026-07-05T11:12:57.379562+00:00"},{"alias_kind":"pith_short_16","alias_value":"75ADAVTPSQTRYYKA","created_at":"2026-07-05T11:12:57.379562+00:00"},{"alias_kind":"pith_short_8","alias_value":"75ADAVTP","created_at":"2026-07-05T11:12:57.379562+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/75ADAVTPSQTRYYKAPR5LJ2LLMV","json":"https://pith.science/pith/75ADAVTPSQTRYYKAPR5LJ2LLMV.json","graph_json":"https://pith.science/api/pith-number/75ADAVTPSQTRYYKAPR5LJ2LLMV/graph.json","events_json":"https://pith.science/api/pith-number/75ADAVTPSQTRYYKAPR5LJ2LLMV/events.json","paper":"https://pith.science/paper/75ADAVTP"},"agent_actions":{"view_html":"https://pith.science/pith/75ADAVTPSQTRYYKAPR5LJ2LLMV","download_json":"https://pith.science/pith/75ADAVTPSQTRYYKAPR5LJ2LLMV.json","view_paper":"https://pith.science/paper/75ADAVTP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.24820&json=true","fetch_graph":"https://pith.science/api/pith-number/75ADAVTPSQTRYYKAPR5LJ2LLMV/graph.json","fetch_events":"https://pith.science/api/pith-number/75ADAVTPSQTRYYKAPR5LJ2LLMV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/75ADAVTPSQTRYYKAPR5LJ2LLMV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/75ADAVTPSQTRYYKAPR5LJ2LLMV/action/storage_attestation","attest_author":"https://pith.science/pith/75ADAVTPSQTRYYKAPR5LJ2LLMV/action/author_attestation","sign_citation":"https://pith.science/pith/75ADAVTPSQTRYYKAPR5LJ2LLMV/action/citation_signature","submit_replication":"https://pith.science/pith/75ADAVTPSQTRYYKAPR5LJ2LLMV/action/replication_record"}},"created_at":"2026-07-05T11:12:57.379562+00:00","updated_at":"2026-07-05T11:12:57.379562+00:00"}