{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:4AU7VBULLZEFFPDWA25T7IHJ2Y","short_pith_number":"pith:4AU7VBUL","schema_version":"1.0","canonical_sha256":"e029fa868b5e4852bc7606bb3fa0e9d609462b9cf18b93bdabbde55fe609e8b1","source":{"kind":"arxiv","id":"2104.02133","version":3},"attestation_state":"computed","paper":{"title":"SpeechStew: Simply Mix All Available Speech Recognition Data to Train One Large Neural Network","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Chris Lee, Daniel Park, Mohammad Norouzi, Quoc Le, William Chan, Yu Zhang","submitted_at":"2021-04-05T20:13:36Z","abstract_excerpt":"We present SpeechStew, a speech recognition model that is trained on a combination of various publicly available speech recognition datasets: AMI, Broadcast News, Common Voice, LibriSpeech, Switchboard/Fisher, Tedlium, and Wall Street Journal. SpeechStew simply mixes all of these datasets together, without any special re-weighting or re-balancing of the datasets. SpeechStew achieves SoTA or near SoTA results across a variety of tasks, without the use of an external language model. Our results include 9.0\\% WER on AMI-IHM, 4.7\\% WER on Switchboard, 8.3\\% WER on CallHome, and 1.3\\% on WSJ, which"},"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":"2104.02133","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-04-05T20:13:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"ce9df96df3d9c4a5bde772f60172d240d6c31afeee8d02531db728d9f7ac0a05","abstract_canon_sha256":"f54ac7b43d62743f566c3d28dc64102f3de8f57848edcaef805e1a8e1d5842b8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:35:41.733429Z","signature_b64":"3TvS24N5yPVDwUyGgH4f0Y7Imv0N88+HW+GutYBXrAPOtdhOz4axdDlkL5QIGUp/RtUaXCT6TTpioPwRSNjVDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e029fa868b5e4852bc7606bb3fa0e9d609462b9cf18b93bdabbde55fe609e8b1","last_reissued_at":"2026-07-05T02:35:41.733032Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:35:41.733032Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SpeechStew: Simply Mix All Available Speech Recognition Data to Train One Large Neural Network","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Chris Lee, Daniel Park, Mohammad Norouzi, Quoc Le, William Chan, Yu Zhang","submitted_at":"2021-04-05T20:13:36Z","abstract_excerpt":"We present SpeechStew, a speech recognition model that is trained on a combination of various publicly available speech recognition datasets: AMI, Broadcast News, Common Voice, LibriSpeech, Switchboard/Fisher, Tedlium, and Wall Street Journal. SpeechStew simply mixes all of these datasets together, without any special re-weighting or re-balancing of the datasets. SpeechStew achieves SoTA or near SoTA results across a variety of tasks, without the use of an external language model. Our results include 9.0\\% WER on AMI-IHM, 4.7\\% WER on Switchboard, 8.3\\% WER on CallHome, and 1.3\\% on WSJ, which"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2104.02133","kind":"arxiv","version":3},"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/2104.02133/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":"2104.02133","created_at":"2026-07-05T02:35:41.733090+00:00"},{"alias_kind":"arxiv_version","alias_value":"2104.02133v3","created_at":"2026-07-05T02:35:41.733090+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2104.02133","created_at":"2026-07-05T02:35:41.733090+00:00"},{"alias_kind":"pith_short_12","alias_value":"4AU7VBULLZEF","created_at":"2026-07-05T02:35:41.733090+00:00"},{"alias_kind":"pith_short_16","alias_value":"4AU7VBULLZEFFPDW","created_at":"2026-07-05T02:35:41.733090+00:00"},{"alias_kind":"pith_short_8","alias_value":"4AU7VBUL","created_at":"2026-07-05T02:35:41.733090+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.20869","citing_title":"OLMoASR: Open Models and Data for Training Robust Speech Recognition Models","ref_index":6,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y","json":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y.json","graph_json":"https://pith.science/api/pith-number/4AU7VBULLZEFFPDWA25T7IHJ2Y/graph.json","events_json":"https://pith.science/api/pith-number/4AU7VBULLZEFFPDWA25T7IHJ2Y/events.json","paper":"https://pith.science/paper/4AU7VBUL"},"agent_actions":{"view_html":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y","download_json":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y.json","view_paper":"https://pith.science/paper/4AU7VBUL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2104.02133&json=true","fetch_graph":"https://pith.science/api/pith-number/4AU7VBULLZEFFPDWA25T7IHJ2Y/graph.json","fetch_events":"https://pith.science/api/pith-number/4AU7VBULLZEFFPDWA25T7IHJ2Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y/action/storage_attestation","attest_author":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y/action/author_attestation","sign_citation":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y/action/citation_signature","submit_replication":"https://pith.science/pith/4AU7VBULLZEFFPDWA25T7IHJ2Y/action/replication_record"}},"created_at":"2026-07-05T02:35:41.733090+00:00","updated_at":"2026-07-05T02:35:41.733090+00:00"}