{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:6KT2Z6JAEH35SWV5ORIMXNL4FF","short_pith_number":"pith:6KT2Z6JA","schema_version":"1.0","canonical_sha256":"f2a7acf92021f7d95abd7450cbb57c29527caa0267de97aec8e87542c074a601","source":{"kind":"arxiv","id":"1603.03185","version":2},"attestation_state":"computed","paper":{"title":"Personalized Speech recognition on mobile devices","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.SD"],"primary_cat":"cs.CL","authors_text":"Alexander Gruenstein, Carolina Parada, David Rybach, Francoise Beaufays, Hasim Sak, Ian McGraw, Kanishka Rao, Montse Gonzalez Arenas, Ouais Alsharif, Raziel Alvarez, Rohit Prabhavalkar","submitted_at":"2016-03-10T08:51:51Z","abstract_excerpt":"We describe a large vocabulary speech recognition system that is accurate, has low latency, and yet has a small enough memory and computational footprint to run faster than real-time on a Nexus 5 Android smartphone. We employ a quantized Long Short-Term Memory (LSTM) acoustic model trained with connectionist temporal classification (CTC) to directly predict phoneme targets, and further reduce its memory footprint using an SVD-based compression scheme. Additionally, we minimize our memory footprint by using a single language model for both dictation and voice command domains, constructed using "},"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":"1603.03185","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2016-03-10T08:51:51Z","cross_cats_sorted":["cs.LG","cs.SD"],"title_canon_sha256":"839b00debdfcbc854357d3130723cafa0f7985bbacb8f0069194f126c156dc31","abstract_canon_sha256":"f3faec933fa65185cf7c38145462a4cbc6d3108bc0b5524d862b6781c7825104"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T01:19:10.593690Z","signature_b64":"XuYbXKXB0w/m9oJWR5iC4FxZ7pUn//uKIw9F8v9x/cvt0mzDNkZ/CKUFLXL28/c+90S+cPvDEKBXLLIcvia+AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f2a7acf92021f7d95abd7450cbb57c29527caa0267de97aec8e87542c074a601","last_reissued_at":"2026-05-18T01:19:10.593049Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T01:19:10.593049Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Personalized Speech recognition on mobile devices","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.SD"],"primary_cat":"cs.CL","authors_text":"Alexander Gruenstein, Carolina Parada, David Rybach, Francoise Beaufays, Hasim Sak, Ian McGraw, Kanishka Rao, Montse Gonzalez Arenas, Ouais Alsharif, Raziel Alvarez, Rohit Prabhavalkar","submitted_at":"2016-03-10T08:51:51Z","abstract_excerpt":"We describe a large vocabulary speech recognition system that is accurate, has low latency, and yet has a small enough memory and computational footprint to run faster than real-time on a Nexus 5 Android smartphone. We employ a quantized Long Short-Term Memory (LSTM) acoustic model trained with connectionist temporal classification (CTC) to directly predict phoneme targets, and further reduce its memory footprint using an SVD-based compression scheme. Additionally, we minimize our memory footprint by using a single language model for both dictation and voice command domains, constructed using "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1603.03185","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":""},"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":"1603.03185","created_at":"2026-05-18T01:19:10.593138+00:00"},{"alias_kind":"arxiv_version","alias_value":"1603.03185v2","created_at":"2026-05-18T01:19:10.593138+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1603.03185","created_at":"2026-05-18T01:19:10.593138+00:00"},{"alias_kind":"pith_short_12","alias_value":"6KT2Z6JAEH35","created_at":"2026-05-18T12:30:01.593930+00:00"},{"alias_kind":"pith_short_16","alias_value":"6KT2Z6JAEH35SWV5","created_at":"2026-05-18T12:30:01.593930+00:00"},{"alias_kind":"pith_short_8","alias_value":"6KT2Z6JA","created_at":"2026-05-18T12:30:01.593930+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/6KT2Z6JAEH35SWV5ORIMXNL4FF","json":"https://pith.science/pith/6KT2Z6JAEH35SWV5ORIMXNL4FF.json","graph_json":"https://pith.science/api/pith-number/6KT2Z6JAEH35SWV5ORIMXNL4FF/graph.json","events_json":"https://pith.science/api/pith-number/6KT2Z6JAEH35SWV5ORIMXNL4FF/events.json","paper":"https://pith.science/paper/6KT2Z6JA"},"agent_actions":{"view_html":"https://pith.science/pith/6KT2Z6JAEH35SWV5ORIMXNL4FF","download_json":"https://pith.science/pith/6KT2Z6JAEH35SWV5ORIMXNL4FF.json","view_paper":"https://pith.science/paper/6KT2Z6JA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1603.03185&json=true","fetch_graph":"https://pith.science/api/pith-number/6KT2Z6JAEH35SWV5ORIMXNL4FF/graph.json","fetch_events":"https://pith.science/api/pith-number/6KT2Z6JAEH35SWV5ORIMXNL4FF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6KT2Z6JAEH35SWV5ORIMXNL4FF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6KT2Z6JAEH35SWV5ORIMXNL4FF/action/storage_attestation","attest_author":"https://pith.science/pith/6KT2Z6JAEH35SWV5ORIMXNL4FF/action/author_attestation","sign_citation":"https://pith.science/pith/6KT2Z6JAEH35SWV5ORIMXNL4FF/action/citation_signature","submit_replication":"https://pith.science/pith/6KT2Z6JAEH35SWV5ORIMXNL4FF/action/replication_record"}},"created_at":"2026-05-18T01:19:10.593138+00:00","updated_at":"2026-05-18T01:19:10.593138+00:00"}