{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:AXJ64V6PBV23I33KBUKJXBX25E","short_pith_number":"pith:AXJ64V6P","schema_version":"1.0","canonical_sha256":"05d3ee57cf0d75b46f6a0d149b86fae936de96176b6ba346982437e566d1dbcc","source":{"kind":"arxiv","id":"2103.14152","version":2},"attestation_state":"computed","paper":{"title":"Residual Energy-Based Models for End-to-End Speech Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Bo Li, Liangliang Cao, Philip C. Woodland, Qiujia Li, Yu Zhang","submitted_at":"2021-03-25T22:08:00Z","abstract_excerpt":"End-to-end models with auto-regressive decoders have shown impressive results for automatic speech recognition (ASR). These models formulate the sequence-level probability as a product of the conditional probabilities of all individual tokens given their histories. However, the performance of locally normalised models can be sub-optimal because of factors such as exposure bias. Consequently, the model distribution differs from the underlying data distribution. In this paper, the residual energy-based model (R-EBM) is proposed to complement the auto-regressive ASR model to close the gap between"},"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":"2103.14152","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.AS","submitted_at":"2021-03-25T22:08:00Z","cross_cats_sorted":["cs.CL","cs.LG","cs.SD"],"title_canon_sha256":"8a6b4d1645eda9862ecb2644acd2a9fe1da968bc36971678cb81d94ce376b594","abstract_canon_sha256":"4b7d81ac3ffae04b391b247b4197fdec319ca545b9a59fe0132425bcc552b10c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:51:54.214303Z","signature_b64":"ZMVgPaEzRUQ1gZ5MF1tP8u4pBVabPVhNVO6lzrw0YZdo9/OgUn1z9UCS78oR/T3KgyY43lDrsM941llApCxzCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"05d3ee57cf0d75b46f6a0d149b86fae936de96176b6ba346982437e566d1dbcc","last_reissued_at":"2026-07-05T02:51:54.213815Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:51:54.213815Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Residual Energy-Based Models for End-to-End Speech Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Bo Li, Liangliang Cao, Philip C. Woodland, Qiujia Li, Yu Zhang","submitted_at":"2021-03-25T22:08:00Z","abstract_excerpt":"End-to-end models with auto-regressive decoders have shown impressive results for automatic speech recognition (ASR). These models formulate the sequence-level probability as a product of the conditional probabilities of all individual tokens given their histories. However, the performance of locally normalised models can be sub-optimal because of factors such as exposure bias. Consequently, the model distribution differs from the underlying data distribution. In this paper, the residual energy-based model (R-EBM) is proposed to complement the auto-regressive ASR model to close the gap between"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.14152","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/2103.14152/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":"2103.14152","created_at":"2026-07-05T02:51:54.213872+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.14152v2","created_at":"2026-07-05T02:51:54.213872+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.14152","created_at":"2026-07-05T02:51:54.213872+00:00"},{"alias_kind":"pith_short_12","alias_value":"AXJ64V6PBV23","created_at":"2026-07-05T02:51:54.213872+00:00"},{"alias_kind":"pith_short_16","alias_value":"AXJ64V6PBV23I33K","created_at":"2026-07-05T02:51:54.213872+00:00"},{"alias_kind":"pith_short_8","alias_value":"AXJ64V6P","created_at":"2026-07-05T02:51:54.213872+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.13771","citing_title":"Score-Based Training for Energy-Based TTS Models","ref_index":12,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E","json":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E.json","graph_json":"https://pith.science/api/pith-number/AXJ64V6PBV23I33KBUKJXBX25E/graph.json","events_json":"https://pith.science/api/pith-number/AXJ64V6PBV23I33KBUKJXBX25E/events.json","paper":"https://pith.science/paper/AXJ64V6P"},"agent_actions":{"view_html":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E","download_json":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E.json","view_paper":"https://pith.science/paper/AXJ64V6P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.14152&json=true","fetch_graph":"https://pith.science/api/pith-number/AXJ64V6PBV23I33KBUKJXBX25E/graph.json","fetch_events":"https://pith.science/api/pith-number/AXJ64V6PBV23I33KBUKJXBX25E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E/action/storage_attestation","attest_author":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E/action/author_attestation","sign_citation":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E/action/citation_signature","submit_replication":"https://pith.science/pith/AXJ64V6PBV23I33KBUKJXBX25E/action/replication_record"}},"created_at":"2026-07-05T02:51:54.213872+00:00","updated_at":"2026-07-05T02:51:54.213872+00:00"}