{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:DJELAWLL7AS66PK2IOXD522PA2","short_pith_number":"pith:DJELAWLL","schema_version":"1.0","canonical_sha256":"1a48b0596bf825ef3d5a43ae3eeb4f06b39181a3b778159f9c1834ab72e62a4b","source":{"kind":"arxiv","id":"2303.08343","version":1},"attestation_state":"computed","paper":{"title":"Sharing Low Rank Conformer Weights for Tiny Always-On Ambient Speech Recognition Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Antoine Bruguier, Ding Zhao, Ian McGraw, Rohit Prabhavalkar, Shaojin Ding, Steven M. Hernandez, Tara N. Sainath, Yanzhang He","submitted_at":"2023-03-15T03:21:38Z","abstract_excerpt":"Continued improvements in machine learning techniques offer exciting new opportunities through the use of larger models and larger training datasets. However, there is a growing need to offer these new capabilities on-board low-powered devices such as smartphones, wearables and other embedded environments where only low memory is available. Towards this, we consider methods to reduce the model size of Conformer-based speech recognition models which typically require models with greater than 100M parameters down to just $5$M parameters while minimizing impact on model quality. Such a model allo"},"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":"2303.08343","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2023-03-15T03:21:38Z","cross_cats_sorted":["cs.AI","cs.LG","cs.SD"],"title_canon_sha256":"dd197d439c57f96083d76958898f0995ca515d806612b49402847538b433eed0","abstract_canon_sha256":"4d39bf6c0a2bf52244848ff7ff94aaca695ddf26b6faf5abb0bd0edcb0ff5f7a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:51:31.802972Z","signature_b64":"U2yvB8gk0I9CYv6A82YkxrJjPVuJtJO7TIZGk3kmGo+kormZwJOQ2/iV9/sLv2bTEvPb8/KbUorSwEossYdrAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1a48b0596bf825ef3d5a43ae3eeb4f06b39181a3b778159f9c1834ab72e62a4b","last_reissued_at":"2026-07-05T05:51:31.802559Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:51:31.802559Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sharing Low Rank Conformer Weights for Tiny Always-On Ambient Speech Recognition Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Antoine Bruguier, Ding Zhao, Ian McGraw, Rohit Prabhavalkar, Shaojin Ding, Steven M. Hernandez, Tara N. Sainath, Yanzhang He","submitted_at":"2023-03-15T03:21:38Z","abstract_excerpt":"Continued improvements in machine learning techniques offer exciting new opportunities through the use of larger models and larger training datasets. However, there is a growing need to offer these new capabilities on-board low-powered devices such as smartphones, wearables and other embedded environments where only low memory is available. Towards this, we consider methods to reduce the model size of Conformer-based speech recognition models which typically require models with greater than 100M parameters down to just $5$M parameters while minimizing impact on model quality. Such a model allo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.08343","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/2303.08343/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":"2303.08343","created_at":"2026-07-05T05:51:31.802611+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.08343v1","created_at":"2026-07-05T05:51:31.802611+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.08343","created_at":"2026-07-05T05:51:31.802611+00:00"},{"alias_kind":"pith_short_12","alias_value":"DJELAWLL7AS6","created_at":"2026-07-05T05:51:31.802611+00:00"},{"alias_kind":"pith_short_16","alias_value":"DJELAWLL7AS66PK2","created_at":"2026-07-05T05:51:31.802611+00:00"},{"alias_kind":"pith_short_8","alias_value":"DJELAWLL","created_at":"2026-07-05T05:51:31.802611+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/DJELAWLL7AS66PK2IOXD522PA2","json":"https://pith.science/pith/DJELAWLL7AS66PK2IOXD522PA2.json","graph_json":"https://pith.science/api/pith-number/DJELAWLL7AS66PK2IOXD522PA2/graph.json","events_json":"https://pith.science/api/pith-number/DJELAWLL7AS66PK2IOXD522PA2/events.json","paper":"https://pith.science/paper/DJELAWLL"},"agent_actions":{"view_html":"https://pith.science/pith/DJELAWLL7AS66PK2IOXD522PA2","download_json":"https://pith.science/pith/DJELAWLL7AS66PK2IOXD522PA2.json","view_paper":"https://pith.science/paper/DJELAWLL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.08343&json=true","fetch_graph":"https://pith.science/api/pith-number/DJELAWLL7AS66PK2IOXD522PA2/graph.json","fetch_events":"https://pith.science/api/pith-number/DJELAWLL7AS66PK2IOXD522PA2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DJELAWLL7AS66PK2IOXD522PA2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DJELAWLL7AS66PK2IOXD522PA2/action/storage_attestation","attest_author":"https://pith.science/pith/DJELAWLL7AS66PK2IOXD522PA2/action/author_attestation","sign_citation":"https://pith.science/pith/DJELAWLL7AS66PK2IOXD522PA2/action/citation_signature","submit_replication":"https://pith.science/pith/DJELAWLL7AS66PK2IOXD522PA2/action/replication_record"}},"created_at":"2026-07-05T05:51:31.802611+00:00","updated_at":"2026-07-05T05:51:31.802611+00:00"}