{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:4F6RPHEV6MDQB5F6DDPFAYWNDG","short_pith_number":"pith:4F6RPHEV","schema_version":"1.0","canonical_sha256":"e17d179c95f30700f4be18de5062cd199765cee094eaba9e49d074ffc0dd639e","source":{"kind":"arxiv","id":"2311.02772","version":2},"attestation_state":"computed","paper":{"title":"Attention or Convolution: Transformer Encoders in Audio Language Models for Inference Efficiency","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","eess.AS"],"primary_cat":"cs.SD","authors_text":"Ching-Feng Yeh, Daniel Bikel, Hakan Inan, Rashi Rungta, Sungho Jeon, Wei-Ning Hsu, Yashar Mehdad","submitted_at":"2023-11-05T21:30:10Z","abstract_excerpt":"In this paper, we show that a simple self-supervised pre-trained audio model can achieve comparable inference efficiency to more complicated pre-trained models with speech transformer encoders. These speech transformers rely on mixing convolutional modules with self-attention modules. They achieve state-of-the-art performance on ASR with top efficiency. We first show that employing these speech transformers as an encoder significantly improves the efficiency of pre-trained audio models as well. However, our study shows that we can achieve comparable efficiency with advanced self-attention sole"},"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":"2311.02772","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2023-11-05T21:30:10Z","cross_cats_sorted":["cs.CL","eess.AS"],"title_canon_sha256":"bca6a6af68551ca9dd553775dc3371c6d625362f57f00e03ac38112116fb64ee","abstract_canon_sha256":"bde864a4394b15d2d5e49c7818777e14b6831e2fb5534afc45858a36b0357e86"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:42:43.141383Z","signature_b64":"MsuDnacsa5FY/SjbKhrhc+gmA37+ogm2hwupEsXe+JlEcZX11qkHWYU1ohyE8ZtuhtOIxHInlkCgwgsOOYc+Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e17d179c95f30700f4be18de5062cd199765cee094eaba9e49d074ffc0dd639e","last_reissued_at":"2026-07-05T07:42:43.140924Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:42:43.140924Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Attention or Convolution: Transformer Encoders in Audio Language Models for Inference Efficiency","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","eess.AS"],"primary_cat":"cs.SD","authors_text":"Ching-Feng Yeh, Daniel Bikel, Hakan Inan, Rashi Rungta, Sungho Jeon, Wei-Ning Hsu, Yashar Mehdad","submitted_at":"2023-11-05T21:30:10Z","abstract_excerpt":"In this paper, we show that a simple self-supervised pre-trained audio model can achieve comparable inference efficiency to more complicated pre-trained models with speech transformer encoders. These speech transformers rely on mixing convolutional modules with self-attention modules. They achieve state-of-the-art performance on ASR with top efficiency. We first show that employing these speech transformers as an encoder significantly improves the efficiency of pre-trained audio models as well. However, our study shows that we can achieve comparable efficiency with advanced self-attention sole"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.02772","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/2311.02772/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":"2311.02772","created_at":"2026-07-05T07:42:43.140982+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.02772v2","created_at":"2026-07-05T07:42:43.140982+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.02772","created_at":"2026-07-05T07:42:43.140982+00:00"},{"alias_kind":"pith_short_12","alias_value":"4F6RPHEV6MDQ","created_at":"2026-07-05T07:42:43.140982+00:00"},{"alias_kind":"pith_short_16","alias_value":"4F6RPHEV6MDQB5F6","created_at":"2026-07-05T07:42:43.140982+00:00"},{"alias_kind":"pith_short_8","alias_value":"4F6RPHEV","created_at":"2026-07-05T07:42:43.140982+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/4F6RPHEV6MDQB5F6DDPFAYWNDG","json":"https://pith.science/pith/4F6RPHEV6MDQB5F6DDPFAYWNDG.json","graph_json":"https://pith.science/api/pith-number/4F6RPHEV6MDQB5F6DDPFAYWNDG/graph.json","events_json":"https://pith.science/api/pith-number/4F6RPHEV6MDQB5F6DDPFAYWNDG/events.json","paper":"https://pith.science/paper/4F6RPHEV"},"agent_actions":{"view_html":"https://pith.science/pith/4F6RPHEV6MDQB5F6DDPFAYWNDG","download_json":"https://pith.science/pith/4F6RPHEV6MDQB5F6DDPFAYWNDG.json","view_paper":"https://pith.science/paper/4F6RPHEV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.02772&json=true","fetch_graph":"https://pith.science/api/pith-number/4F6RPHEV6MDQB5F6DDPFAYWNDG/graph.json","fetch_events":"https://pith.science/api/pith-number/4F6RPHEV6MDQB5F6DDPFAYWNDG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4F6RPHEV6MDQB5F6DDPFAYWNDG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4F6RPHEV6MDQB5F6DDPFAYWNDG/action/storage_attestation","attest_author":"https://pith.science/pith/4F6RPHEV6MDQB5F6DDPFAYWNDG/action/author_attestation","sign_citation":"https://pith.science/pith/4F6RPHEV6MDQB5F6DDPFAYWNDG/action/citation_signature","submit_replication":"https://pith.science/pith/4F6RPHEV6MDQB5F6DDPFAYWNDG/action/replication_record"}},"created_at":"2026-07-05T07:42:43.140982+00:00","updated_at":"2026-07-05T07:42:43.140982+00:00"}