{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:QHLBWBFSGS5TYTJLOTWEOFPJV6","short_pith_number":"pith:QHLBWBFS","schema_version":"1.0","canonical_sha256":"81d61b04b234bb3c4d2b74ec4715e9afba83e9655671a4c150a28ca1927220a5","source":{"kind":"arxiv","id":"2110.01900","version":4},"attestation_state":"computed","paper":{"title":"DistilHuBERT: Speech Representation Learning by Layer-wise Distillation of Hidden-unit BERT","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.CL","authors_text":"Heng-Jui Chang, Hung-yi Lee, Shu-wen Yang","submitted_at":"2021-10-05T09:34:44Z","abstract_excerpt":"Self-supervised speech representation learning methods like wav2vec 2.0 and Hidden-unit BERT (HuBERT) leverage unlabeled speech data for pre-training and offer good representations for numerous speech processing tasks. Despite the success of these methods, they require large memory and high pre-training costs, making them inaccessible for researchers in academia and small companies. Therefore, this paper introduces DistilHuBERT, a novel multi-task learning framework to distill hidden representations from a HuBERT model directly. This method reduces HuBERT's size by 75% and 73% faster while ret"},"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":"2110.01900","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-10-05T09:34:44Z","cross_cats_sorted":["eess.AS"],"title_canon_sha256":"0cd3c23c074ea850f38f32f94ada545b081beebc9dda7e6c6b9c475ee18ac902","abstract_canon_sha256":"bc5a347dcefff3371f20afc48ca667b703f69462c606205bc01f36fbfcb3fc2a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:18:28.381133Z","signature_b64":"WYapdevqXMZoG9hQtIyHRzNrnoZjEkbSW4HwEcIok8iJa84h+P21OnqrMBSOzmG2/keSoIkXZkk8mixH9BtKDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"81d61b04b234bb3c4d2b74ec4715e9afba83e9655671a4c150a28ca1927220a5","last_reissued_at":"2026-07-05T04:18:28.380653Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:18:28.380653Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DistilHuBERT: Speech Representation Learning by Layer-wise Distillation of Hidden-unit BERT","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.CL","authors_text":"Heng-Jui Chang, Hung-yi Lee, Shu-wen Yang","submitted_at":"2021-10-05T09:34:44Z","abstract_excerpt":"Self-supervised speech representation learning methods like wav2vec 2.0 and Hidden-unit BERT (HuBERT) leverage unlabeled speech data for pre-training and offer good representations for numerous speech processing tasks. Despite the success of these methods, they require large memory and high pre-training costs, making them inaccessible for researchers in academia and small companies. Therefore, this paper introduces DistilHuBERT, a novel multi-task learning framework to distill hidden representations from a HuBERT model directly. This method reduces HuBERT's size by 75% and 73% faster while ret"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.01900","kind":"arxiv","version":4},"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/2110.01900/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":"2110.01900","created_at":"2026-07-05T04:18:28.380705+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.01900v4","created_at":"2026-07-05T04:18:28.380705+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.01900","created_at":"2026-07-05T04:18:28.380705+00:00"},{"alias_kind":"pith_short_12","alias_value":"QHLBWBFSGS5T","created_at":"2026-07-05T04:18:28.380705+00:00"},{"alias_kind":"pith_short_16","alias_value":"QHLBWBFSGS5TYTJL","created_at":"2026-07-05T04:18:28.380705+00:00"},{"alias_kind":"pith_short_8","alias_value":"QHLBWBFS","created_at":"2026-07-05T04:18:28.380705+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24910","citing_title":"End-to-End Voice Intent Recognition for Spontaneous Human-Drone Interaction with Naive Users","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06837","citing_title":"SEAM: Shortcut-Aware Real-Time Detection of Scripted vs. Spontaneous Speech for Interview Guardrails","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6","json":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6.json","graph_json":"https://pith.science/api/pith-number/QHLBWBFSGS5TYTJLOTWEOFPJV6/graph.json","events_json":"https://pith.science/api/pith-number/QHLBWBFSGS5TYTJLOTWEOFPJV6/events.json","paper":"https://pith.science/paper/QHLBWBFS"},"agent_actions":{"view_html":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6","download_json":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6.json","view_paper":"https://pith.science/paper/QHLBWBFS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.01900&json=true","fetch_graph":"https://pith.science/api/pith-number/QHLBWBFSGS5TYTJLOTWEOFPJV6/graph.json","fetch_events":"https://pith.science/api/pith-number/QHLBWBFSGS5TYTJLOTWEOFPJV6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6/action/storage_attestation","attest_author":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6/action/author_attestation","sign_citation":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6/action/citation_signature","submit_replication":"https://pith.science/pith/QHLBWBFSGS5TYTJLOTWEOFPJV6/action/replication_record"}},"created_at":"2026-07-05T04:18:28.380705+00:00","updated_at":"2026-07-05T04:18:28.380705+00:00"}