{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:7ISRLJAQA2IIMXY3AWSMEBPBE7","short_pith_number":"pith:7ISRLJAQ","schema_version":"1.0","canonical_sha256":"fa2515a4100690865f1b05a4c205e127ff37ce60eedfea6cbaaa015af29809c0","source":{"kind":"arxiv","id":"2309.09546","version":2},"attestation_state":"computed","paper":{"title":"Training dynamic models using early exits for automatic speech recognition on resource-constrained devices","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.SD"],"primary_cat":"eess.AS","authors_text":"Alessio Brutti, Daniele Falavigna, Desh Raj, George August Wright, Lucas Ondel Yang, Mohamed Nabih Ali, Salah Zaiem, Umberto Cappellazzo","submitted_at":"2023-09-18T07:45:16Z","abstract_excerpt":"The ability to dynamically adjust the computational load of neural models during inference is crucial for on-device processing scenarios characterised by limited and time-varying computational resources. A promising solution is presented by early-exit architectures, in which additional exit branches are appended to intermediate layers of the encoder. In self-attention models for automatic speech recognition (ASR), early-exit architectures enable the development of dynamic models capable of adapting their size and architecture to varying levels of computational resources and ASR performance dem"},"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":"2309.09546","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2023-09-18T07:45:16Z","cross_cats_sorted":["cs.CL","cs.SD"],"title_canon_sha256":"b33c90085ee37829ba05499d824989f3c329cdf72613ac5dc98975ed4b195544","abstract_canon_sha256":"ec3df3c35240fed2f45f224530b3d0becac1c9ec9eb94ec09a750bd3caa68c25"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:47:59.339485Z","signature_b64":"o0w1br/WBNMaH2HizJR8xLh01jp4DDhFvRYWPK8nEPosXSdFi43rrmhBgCsbybQ2wsLAxR4dTBPd2Rag1/p2Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fa2515a4100690865f1b05a4c205e127ff37ce60eedfea6cbaaa015af29809c0","last_reissued_at":"2026-07-05T07:47:59.338964Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:47:59.338964Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Training dynamic models using early exits for automatic speech recognition on resource-constrained devices","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.SD"],"primary_cat":"eess.AS","authors_text":"Alessio Brutti, Daniele Falavigna, Desh Raj, George August Wright, Lucas Ondel Yang, Mohamed Nabih Ali, Salah Zaiem, Umberto Cappellazzo","submitted_at":"2023-09-18T07:45:16Z","abstract_excerpt":"The ability to dynamically adjust the computational load of neural models during inference is crucial for on-device processing scenarios characterised by limited and time-varying computational resources. A promising solution is presented by early-exit architectures, in which additional exit branches are appended to intermediate layers of the encoder. In self-attention models for automatic speech recognition (ASR), early-exit architectures enable the development of dynamic models capable of adapting their size and architecture to varying levels of computational resources and ASR performance dem"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.09546","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2309.09546/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":"2309.09546","created_at":"2026-07-05T07:47:59.339021+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.09546v2","created_at":"2026-07-05T07:47:59.339021+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.09546","created_at":"2026-07-05T07:47:59.339021+00:00"},{"alias_kind":"pith_short_12","alias_value":"7ISRLJAQA2II","created_at":"2026-07-05T07:47:59.339021+00:00"},{"alias_kind":"pith_short_16","alias_value":"7ISRLJAQA2IIMXY3","created_at":"2026-07-05T07:47:59.339021+00:00"},{"alias_kind":"pith_short_8","alias_value":"7ISRLJAQ","created_at":"2026-07-05T07:47:59.339021+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/7ISRLJAQA2IIMXY3AWSMEBPBE7","json":"https://pith.science/pith/7ISRLJAQA2IIMXY3AWSMEBPBE7.json","graph_json":"https://pith.science/api/pith-number/7ISRLJAQA2IIMXY3AWSMEBPBE7/graph.json","events_json":"https://pith.science/api/pith-number/7ISRLJAQA2IIMXY3AWSMEBPBE7/events.json","paper":"https://pith.science/paper/7ISRLJAQ"},"agent_actions":{"view_html":"https://pith.science/pith/7ISRLJAQA2IIMXY3AWSMEBPBE7","download_json":"https://pith.science/pith/7ISRLJAQA2IIMXY3AWSMEBPBE7.json","view_paper":"https://pith.science/paper/7ISRLJAQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.09546&json=true","fetch_graph":"https://pith.science/api/pith-number/7ISRLJAQA2IIMXY3AWSMEBPBE7/graph.json","fetch_events":"https://pith.science/api/pith-number/7ISRLJAQA2IIMXY3AWSMEBPBE7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7ISRLJAQA2IIMXY3AWSMEBPBE7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7ISRLJAQA2IIMXY3AWSMEBPBE7/action/storage_attestation","attest_author":"https://pith.science/pith/7ISRLJAQA2IIMXY3AWSMEBPBE7/action/author_attestation","sign_citation":"https://pith.science/pith/7ISRLJAQA2IIMXY3AWSMEBPBE7/action/citation_signature","submit_replication":"https://pith.science/pith/7ISRLJAQA2IIMXY3AWSMEBPBE7/action/replication_record"}},"created_at":"2026-07-05T07:47:59.339021+00:00","updated_at":"2026-07-05T07:47:59.339021+00:00"}