{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:FAF4VSSGPH3EDMB566HV64UFVN","short_pith_number":"pith:FAF4VSSG","schema_version":"1.0","canonical_sha256":"280bcaca4679f641b03df78f5f7285ab4c9e8dceeac86cfec5e8e4c0979472d5","source":{"kind":"arxiv","id":"2010.10759","version":4},"attestation_state":"computed","paper":{"title":"Emformer: Efficient Memory Transformer Based Acoustic Model For Low Latency Streaming Speech Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"Ching-Feng Yeh, Chunyang Wu, Duc Le, Frank Zhang, Julian Chan, Mike Seltzer, Yangyang Shi, Yongqiang Wang","submitted_at":"2020-10-21T04:38:09Z","abstract_excerpt":"This paper proposes an efficient memory transformer Emformer for low latency streaming speech recognition. In Emformer, the long-range history context is distilled into an augmented memory bank to reduce self-attention's computation complexity. A cache mechanism saves the computation for the key and value in self-attention for the left context. Emformer applies a parallelized block processing in training to support low latency models. We carry out experiments on benchmark LibriSpeech data. Under average latency of 960 ms, Emformer gets WER $2.50\\%$ on test-clean and $5.62\\%$ on test-other. Com"},"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":"2010.10759","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SD","submitted_at":"2020-10-21T04:38:09Z","cross_cats_sorted":["cs.CL","cs.LG","eess.AS"],"title_canon_sha256":"eb863dcc5ea94d3e430c2dcf89e770cb4ec6b074e90df658a0a5e836d2d07b40","abstract_canon_sha256":"94cf8ebe991fbf9e2cdbdbc6dbd29ad33e6f8737d11f9d55fe5980c0710bc790"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:02:56.654988Z","signature_b64":"bOi1RiCYcYjceCDxf45XvwYpDvAfFq3/LwhvrJiq2c8hYNXOFRX/dlTM9roGTX0NqYyLAP47NFX1gMKIXgssAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"280bcaca4679f641b03df78f5f7285ab4c9e8dceeac86cfec5e8e4c0979472d5","last_reissued_at":"2026-07-05T02:02:56.654465Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:02:56.654465Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Emformer: Efficient Memory Transformer Based Acoustic Model For Low Latency Streaming Speech Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"Ching-Feng Yeh, Chunyang Wu, Duc Le, Frank Zhang, Julian Chan, Mike Seltzer, Yangyang Shi, Yongqiang Wang","submitted_at":"2020-10-21T04:38:09Z","abstract_excerpt":"This paper proposes an efficient memory transformer Emformer for low latency streaming speech recognition. In Emformer, the long-range history context is distilled into an augmented memory bank to reduce self-attention's computation complexity. A cache mechanism saves the computation for the key and value in self-attention for the left context. Emformer applies a parallelized block processing in training to support low latency models. We carry out experiments on benchmark LibriSpeech data. Under average latency of 960 ms, Emformer gets WER $2.50\\%$ on test-clean and $5.62\\%$ on test-other. Com"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.10759","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/2010.10759/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":"2010.10759","created_at":"2026-07-05T02:02:56.654539+00:00"},{"alias_kind":"arxiv_version","alias_value":"2010.10759v4","created_at":"2026-07-05T02:02:56.654539+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.10759","created_at":"2026-07-05T02:02:56.654539+00:00"},{"alias_kind":"pith_short_12","alias_value":"FAF4VSSGPH3E","created_at":"2026-07-05T02:02:56.654539+00:00"},{"alias_kind":"pith_short_16","alias_value":"FAF4VSSGPH3EDMB5","created_at":"2026-07-05T02:02:56.654539+00:00"},{"alias_kind":"pith_short_8","alias_value":"FAF4VSSG","created_at":"2026-07-05T02:02:56.654539+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/FAF4VSSGPH3EDMB566HV64UFVN","json":"https://pith.science/pith/FAF4VSSGPH3EDMB566HV64UFVN.json","graph_json":"https://pith.science/api/pith-number/FAF4VSSGPH3EDMB566HV64UFVN/graph.json","events_json":"https://pith.science/api/pith-number/FAF4VSSGPH3EDMB566HV64UFVN/events.json","paper":"https://pith.science/paper/FAF4VSSG"},"agent_actions":{"view_html":"https://pith.science/pith/FAF4VSSGPH3EDMB566HV64UFVN","download_json":"https://pith.science/pith/FAF4VSSGPH3EDMB566HV64UFVN.json","view_paper":"https://pith.science/paper/FAF4VSSG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2010.10759&json=true","fetch_graph":"https://pith.science/api/pith-number/FAF4VSSGPH3EDMB566HV64UFVN/graph.json","fetch_events":"https://pith.science/api/pith-number/FAF4VSSGPH3EDMB566HV64UFVN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FAF4VSSGPH3EDMB566HV64UFVN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FAF4VSSGPH3EDMB566HV64UFVN/action/storage_attestation","attest_author":"https://pith.science/pith/FAF4VSSGPH3EDMB566HV64UFVN/action/author_attestation","sign_citation":"https://pith.science/pith/FAF4VSSGPH3EDMB566HV64UFVN/action/citation_signature","submit_replication":"https://pith.science/pith/FAF4VSSGPH3EDMB566HV64UFVN/action/replication_record"}},"created_at":"2026-07-05T02:02:56.654539+00:00","updated_at":"2026-07-05T02:02:56.654539+00:00"}