{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:7BZUENFKJGAUOZA2HJPOS5GRUY","short_pith_number":"pith:7BZUENFK","schema_version":"1.0","canonical_sha256":"f8734234aa498147641a3a5ee974d1a6334321a9eb0900e471963b6645d4e09c","source":{"kind":"arxiv","id":"2005.11138","version":1},"attestation_state":"computed","paper":{"title":"TinyLSTMs: Efficient Neural Speech Enhancement for Hearing Aids","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.SD","stat.ML"],"primary_cat":"eess.AS","authors_text":"Ari Mandell, Carl Jensen, Igor Fedorov, Li-Chia Yang, Marko Stamenovic, Matthew Mattina, Paul N. Whatmough, Yiming Gan","submitted_at":"2020-05-20T20:37:47Z","abstract_excerpt":"Modern speech enhancement algorithms achieve remarkable noise suppression by means of large recurrent neural networks (RNNs). However, large RNNs limit practical deployment in hearing aid hardware (HW) form-factors, which are battery powered and run on resource-constrained microcontroller units (MCUs) with limited memory capacity and compute capability. In this work, we use model compression techniques to bridge this gap. We define the constraints imposed on the RNN by the HW and describe a method to satisfy them. Although model compression techniques are an active area of research, we are 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":"2005.11138","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.AS","submitted_at":"2020-05-20T20:37:47Z","cross_cats_sorted":["cs.LG","cs.SD","stat.ML"],"title_canon_sha256":"2d98e4ab27aa99a3f9161f5b668332709ba0e00ad33b56894c3d425bb2a63758","abstract_canon_sha256":"99376632405c8a1f20f1cbaf9fb9d747d19f2a5f4fcf5171bc930817f75bbadf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:13:47.822360Z","signature_b64":"OIxeHZj1Qoi5TAiuXwRHZ98E0JkA6GcVF05om0mnPFNlg6N5PG0tBCxZIViESgFq+GR5k/8/YlBTSJeSxgezDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f8734234aa498147641a3a5ee974d1a6334321a9eb0900e471963b6645d4e09c","last_reissued_at":"2026-07-05T03:13:47.821850Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:13:47.821850Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TinyLSTMs: Efficient Neural Speech Enhancement for Hearing Aids","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.SD","stat.ML"],"primary_cat":"eess.AS","authors_text":"Ari Mandell, Carl Jensen, Igor Fedorov, Li-Chia Yang, Marko Stamenovic, Matthew Mattina, Paul N. Whatmough, Yiming Gan","submitted_at":"2020-05-20T20:37:47Z","abstract_excerpt":"Modern speech enhancement algorithms achieve remarkable noise suppression by means of large recurrent neural networks (RNNs). However, large RNNs limit practical deployment in hearing aid hardware (HW) form-factors, which are battery powered and run on resource-constrained microcontroller units (MCUs) with limited memory capacity and compute capability. In this work, we use model compression techniques to bridge this gap. We define the constraints imposed on the RNN by the HW and describe a method to satisfy them. Although model compression techniques are an active area of research, we are the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2005.11138","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/2005.11138/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":"2005.11138","created_at":"2026-07-05T03:13:47.821908+00:00"},{"alias_kind":"arxiv_version","alias_value":"2005.11138v1","created_at":"2026-07-05T03:13:47.821908+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2005.11138","created_at":"2026-07-05T03:13:47.821908+00:00"},{"alias_kind":"pith_short_12","alias_value":"7BZUENFKJGAU","created_at":"2026-07-05T03:13:47.821908+00:00"},{"alias_kind":"pith_short_16","alias_value":"7BZUENFKJGAUOZA2","created_at":"2026-07-05T03:13:47.821908+00:00"},{"alias_kind":"pith_short_8","alias_value":"7BZUENFK","created_at":"2026-07-05T03:13:47.821908+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29575","citing_title":"TF-MoE: Time-Frequency Mixture-of-Experts for Efficient Speech Separation","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29575","citing_title":"TF-MoE: Time-Frequency Mixture-of-Experts for Efficient Speech Separation","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY","json":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY.json","graph_json":"https://pith.science/api/pith-number/7BZUENFKJGAUOZA2HJPOS5GRUY/graph.json","events_json":"https://pith.science/api/pith-number/7BZUENFKJGAUOZA2HJPOS5GRUY/events.json","paper":"https://pith.science/paper/7BZUENFK"},"agent_actions":{"view_html":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY","download_json":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY.json","view_paper":"https://pith.science/paper/7BZUENFK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2005.11138&json=true","fetch_graph":"https://pith.science/api/pith-number/7BZUENFKJGAUOZA2HJPOS5GRUY/graph.json","fetch_events":"https://pith.science/api/pith-number/7BZUENFKJGAUOZA2HJPOS5GRUY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY/action/storage_attestation","attest_author":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY/action/author_attestation","sign_citation":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY/action/citation_signature","submit_replication":"https://pith.science/pith/7BZUENFKJGAUOZA2HJPOS5GRUY/action/replication_record"}},"created_at":"2026-07-05T03:13:47.821908+00:00","updated_at":"2026-07-05T03:13:47.821908+00:00"}