{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NVBI3BK5XWACSISDJT6I2LMSN3","short_pith_number":"pith:NVBI3BK5","schema_version":"1.0","canonical_sha256":"6d428d855dbd802922434cfc8d2d926ed4a57a6032805f4c67373b892ca0b6d9","source":{"kind":"arxiv","id":"2605.28139","version":1},"attestation_state":"computed","paper":{"title":"Data-Efficient On-Policy Distillation for Automatic Speech Recognition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Runyuan Cai, Xiaodong Zeng, Yiming Wang, Yu Lin","submitted_at":"2026-05-27T08:22:47Z","abstract_excerpt":"Building competitive automatic speech recognition (ASR) models usually requires large-scale au- dio supervision, which makes reproduction and specialization expensive. We study Ark-ASR, a 0.6B- parameter audio-conditioned language model trained with 100k hours of speech, and examine whether a strong Qwen-ASR teacher can transfer additional recognition capability through on-policy distillation. Across Mandarin and English ASR benchmarks, the proposed training recipe consistently improves over supervised fine-tuning alone and outperforms the same-scale Qwen3-ASR-0.6B baseline on four of five eva"},"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":"2605.28139","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-27T08:22:47Z","cross_cats_sorted":[],"title_canon_sha256":"dd7d13eb9b6fe36c6d52d45e70f6937ccd6e5c36d36818ef39b8bf24ba9b4e8c","abstract_canon_sha256":"5bdc92333b8638b9e652a0f7d07365eb75ce8df057de26cb9a537aeb52829a96"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:05:00.189091Z","signature_b64":"DPeUZi39Slsqmg8zNxyTmCxJIV5GZSGKAJMJAvEGlyrgFN2NUCdh/eChiVRXX0/C+KPKo07p0tb8UUga1wk/AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6d428d855dbd802922434cfc8d2d926ed4a57a6032805f4c67373b892ca0b6d9","last_reissued_at":"2026-05-28T01:05:00.188679Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:05:00.188679Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Data-Efficient On-Policy Distillation for Automatic Speech Recognition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Runyuan Cai, Xiaodong Zeng, Yiming Wang, Yu Lin","submitted_at":"2026-05-27T08:22:47Z","abstract_excerpt":"Building competitive automatic speech recognition (ASR) models usually requires large-scale au- dio supervision, which makes reproduction and specialization expensive. We study Ark-ASR, a 0.6B- parameter audio-conditioned language model trained with 100k hours of speech, and examine whether a strong Qwen-ASR teacher can transfer additional recognition capability through on-policy distillation. Across Mandarin and English ASR benchmarks, the proposed training recipe consistently improves over supervised fine-tuning alone and outperforms the same-scale Qwen3-ASR-0.6B baseline on four of five eva"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28139","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/2605.28139/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":"2605.28139","created_at":"2026-05-28T01:05:00.188738+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28139v1","created_at":"2026-05-28T01:05:00.188738+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28139","created_at":"2026-05-28T01:05:00.188738+00:00"},{"alias_kind":"pith_short_12","alias_value":"NVBI3BK5XWAC","created_at":"2026-05-28T01:05:00.188738+00:00"},{"alias_kind":"pith_short_16","alias_value":"NVBI3BK5XWACSISD","created_at":"2026-05-28T01:05:00.188738+00:00"},{"alias_kind":"pith_short_8","alias_value":"NVBI3BK5","created_at":"2026-05-28T01:05:00.188738+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/NVBI3BK5XWACSISDJT6I2LMSN3","json":"https://pith.science/pith/NVBI3BK5XWACSISDJT6I2LMSN3.json","graph_json":"https://pith.science/api/pith-number/NVBI3BK5XWACSISDJT6I2LMSN3/graph.json","events_json":"https://pith.science/api/pith-number/NVBI3BK5XWACSISDJT6I2LMSN3/events.json","paper":"https://pith.science/paper/NVBI3BK5"},"agent_actions":{"view_html":"https://pith.science/pith/NVBI3BK5XWACSISDJT6I2LMSN3","download_json":"https://pith.science/pith/NVBI3BK5XWACSISDJT6I2LMSN3.json","view_paper":"https://pith.science/paper/NVBI3BK5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28139&json=true","fetch_graph":"https://pith.science/api/pith-number/NVBI3BK5XWACSISDJT6I2LMSN3/graph.json","fetch_events":"https://pith.science/api/pith-number/NVBI3BK5XWACSISDJT6I2LMSN3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NVBI3BK5XWACSISDJT6I2LMSN3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NVBI3BK5XWACSISDJT6I2LMSN3/action/storage_attestation","attest_author":"https://pith.science/pith/NVBI3BK5XWACSISDJT6I2LMSN3/action/author_attestation","sign_citation":"https://pith.science/pith/NVBI3BK5XWACSISDJT6I2LMSN3/action/citation_signature","submit_replication":"https://pith.science/pith/NVBI3BK5XWACSISDJT6I2LMSN3/action/replication_record"}},"created_at":"2026-05-28T01:05:00.188738+00:00","updated_at":"2026-05-28T01:05:00.188738+00:00"}