{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:2VHVNAZT6UHTT5E2Y7SXQHVPMW","short_pith_number":"pith:2VHVNAZT","schema_version":"1.0","canonical_sha256":"d54f568333f50f39f49ac7e5781eaf65a6b512e45fb845c8ad0d212788c81c01","source":{"kind":"arxiv","id":"2209.15200","version":5},"attestation_state":"computed","paper":{"title":"An efficient encoder-decoder architecture with top-down attention for speech separation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"Kai Li, Runxuan Yang, Xiaolin Hu","submitted_at":"2022-09-30T03:09:53Z","abstract_excerpt":"Deep neural networks have shown excellent prospects in speech separation tasks. However, obtaining good results while keeping a low model complexity remains challenging in real-world applications. In this paper, we provide a bio-inspired efficient encoder-decoder architecture by mimicking the brain's top-down attention, called TDANet, with decreased model complexity without sacrificing performance. The top-down attention in TDANet is extracted by the global attention (GA) module and the cascaded local attention (LA) layers. The GA module takes multi-scale acoustic features as input to extract "},"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":"2209.15200","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SD","submitted_at":"2022-09-30T03:09:53Z","cross_cats_sorted":["cs.LG","eess.AS"],"title_canon_sha256":"90a4048bf65c5a168f2de038c755a526996600f17d429be37012c64213c2d8a8","abstract_canon_sha256":"9fe1cea02d5431cd981360efa82bf9d2729f8717aa201133a7581598daded3c6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:56:14.664942Z","signature_b64":"OOwRwtF2Sd/VF70XRQJqFyLCJkdq9b24CffRGlXXLV5SPaPvlEq/BaLPyfGTqIAQ5EC9Gguy0P3aR5JJnlAoCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d54f568333f50f39f49ac7e5781eaf65a6b512e45fb845c8ad0d212788c81c01","last_reissued_at":"2026-07-05T05:56:14.664369Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:56:14.664369Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An efficient encoder-decoder architecture with top-down attention for speech separation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"Kai Li, Runxuan Yang, Xiaolin Hu","submitted_at":"2022-09-30T03:09:53Z","abstract_excerpt":"Deep neural networks have shown excellent prospects in speech separation tasks. However, obtaining good results while keeping a low model complexity remains challenging in real-world applications. In this paper, we provide a bio-inspired efficient encoder-decoder architecture by mimicking the brain's top-down attention, called TDANet, with decreased model complexity without sacrificing performance. The top-down attention in TDANet is extracted by the global attention (GA) module and the cascaded local attention (LA) layers. The GA module takes multi-scale acoustic features as input to extract "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2209.15200","kind":"arxiv","version":5},"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/2209.15200/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":"2209.15200","created_at":"2026-07-05T05:56:14.664429+00:00"},{"alias_kind":"arxiv_version","alias_value":"2209.15200v5","created_at":"2026-07-05T05:56:14.664429+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2209.15200","created_at":"2026-07-05T05:56:14.664429+00:00"},{"alias_kind":"pith_short_12","alias_value":"2VHVNAZT6UHT","created_at":"2026-07-05T05:56:14.664429+00:00"},{"alias_kind":"pith_short_16","alias_value":"2VHVNAZT6UHTT5E2","created_at":"2026-07-05T05:56:14.664429+00:00"},{"alias_kind":"pith_short_8","alias_value":"2VHVNAZT","created_at":"2026-07-05T05:56:14.664429+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29575","citing_title":"TF-MoE: Time-Frequency Mixture-of-Experts for Efficient Speech Separation","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29575","citing_title":"TF-MoE: Time-Frequency Mixture-of-Experts for Efficient Speech Separation","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2601.06006","citing_title":"Discriminative-Generative Target Speaker Extraction with Decoder-Only Language Models","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2509.11717","citing_title":"CodecSep: Prompt-Driven Universal Sound Separation on Neural Audio Codec Latents","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW","json":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW.json","graph_json":"https://pith.science/api/pith-number/2VHVNAZT6UHTT5E2Y7SXQHVPMW/graph.json","events_json":"https://pith.science/api/pith-number/2VHVNAZT6UHTT5E2Y7SXQHVPMW/events.json","paper":"https://pith.science/paper/2VHVNAZT"},"agent_actions":{"view_html":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW","download_json":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW.json","view_paper":"https://pith.science/paper/2VHVNAZT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2209.15200&json=true","fetch_graph":"https://pith.science/api/pith-number/2VHVNAZT6UHTT5E2Y7SXQHVPMW/graph.json","fetch_events":"https://pith.science/api/pith-number/2VHVNAZT6UHTT5E2Y7SXQHVPMW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW/action/storage_attestation","attest_author":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW/action/author_attestation","sign_citation":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW/action/citation_signature","submit_replication":"https://pith.science/pith/2VHVNAZT6UHTT5E2Y7SXQHVPMW/action/replication_record"}},"created_at":"2026-07-05T05:56:14.664429+00:00","updated_at":"2026-07-05T05:56:14.664429+00:00"}