{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:PWZJFKS5H54WFOTLR7BPMRXGOS","short_pith_number":"pith:PWZJFKS5","schema_version":"1.0","canonical_sha256":"7db292aa5d3f7962ba6b8fc2f646e674b070823c08611f0f0686311a49791fdd","source":{"kind":"arxiv","id":"2511.21740","version":5},"attestation_state":"computed","paper":{"title":"A cross-species neural foundation model for end-to-end speech decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A cross-species pretrained neural encoder enables end-to-end decoding of brain activity into sentences at 10.22 percent word error rate.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chaofei Fan, Francis R Willett, Han Yu, Jingyuan Li, Lea Duncker, Liam Paninski, Linyang He, Nima Mesgarani, Scott Linderman, Tingkai Liu, Trung Le, Yizi Zhang","submitted_at":"2025-11-21T21:25:54Z","abstract_excerpt":"Speech brain-computer interfaces (BCIs) aim to restore communication for people with paralysis by translating neural activity into text. Most systems use cascaded frameworks that decode phonemes before assembling sentences with an n-gram language model (LM), preventing joint optimization of all stages simultaneously. Here, we introduce an end-to-end BraIn-to-Text (BIT) framework that translates neural activity into coherent sentences using a single differentiable neural network. Central to our approach is a cross-task, cross-species pretrained neural encoder, whose representations transfer to "},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2511.21740","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-11-21T21:25:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c121b2147897fb3756c9351b2dd1f12314a87174fb6eabc56739d505caac29cf","abstract_canon_sha256":"e5f815a0ec10f45c4c748ca0849fbca13846f2101c21b7a47c051e26bdf7c708"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:17.017532Z","signature_b64":"Hrtv6EokgmUhkAqtJb+nQZ31qbL+ZjTYCxwKlBGjQWb62HlMG4OE/+1YFKj0Cw3XCDmVzX1ihwI7emW+ZShACw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7db292aa5d3f7962ba6b8fc2f646e674b070823c08611f0f0686311a49791fdd","last_reissued_at":"2026-05-17T23:39:17.016781Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:17.016781Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A cross-species neural foundation model for end-to-end speech decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A cross-species pretrained neural encoder enables end-to-end decoding of brain activity into sentences at 10.22 percent word error rate.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chaofei Fan, Francis R Willett, Han Yu, Jingyuan Li, Lea Duncker, Liam Paninski, Linyang He, Nima Mesgarani, Scott Linderman, Tingkai Liu, Trung Le, Yizi Zhang","submitted_at":"2025-11-21T21:25:54Z","abstract_excerpt":"Speech brain-computer interfaces (BCIs) aim to restore communication for people with paralysis by translating neural activity into text. Most systems use cascaded frameworks that decode phonemes before assembling sentences with an n-gram language model (LM), preventing joint optimization of all stages simultaneously. Here, we introduce an end-to-end BraIn-to-Text (BIT) framework that translates neural activity into coherent sentences using a single differentiable neural network. Central to our approach is a cross-task, cross-species pretrained neural encoder, whose representations transfer to "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Integrated end-to-end with audio large language models and trained with contrastive learning for cross-modal alignment, BIT reduces the word error rate (WER) of the prior end-to-end method from 24.69% to 10.22%.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The representations learned by the cross-species pretrained neural encoder transfer effectively and without major domain shift to human attempted and imagined speech recordings used in the Brain-to-Text benchmarks.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A cross-species pretrained neural encoder combined with end-to-end training and audio LLMs reduces word error rate in neural speech decoding from 24.69% to 10.22% while aligning attempted and imagined speech.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A cross-species pretrained neural encoder enables end-to-end decoding of brain activity into sentences at 10.22 percent word error rate.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"41767f63eea341d2f6ee2b0cd93a47a69053f83fc59f15f12d4c97f79098e62b"},"source":{"id":"2511.21740","kind":"arxiv","version":5},"verdict":{"id":"fe5e8dba-2c17-4d88-b0fd-b585e708d394","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T19:56:23.299472Z","strongest_claim":"Integrated end-to-end with audio large language models and trained with contrastive learning for cross-modal alignment, BIT reduces the word error rate (WER) of the prior end-to-end method from 24.69% to 10.22%.","one_line_summary":"A cross-species pretrained neural encoder combined with end-to-end training and audio LLMs reduces word error rate in neural speech decoding from 24.69% to 10.22% while aligning attempted and imagined speech.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The representations learned by the cross-species pretrained neural encoder transfer effectively and without major domain shift to human attempted and imagined speech recordings used in the Brain-to-Text benchmarks.","pith_extraction_headline":"A cross-species pretrained neural encoder enables end-to-end decoding of brain activity into sentences at 10.22 percent word error rate."},"references":{"count":25,"sample":[{"doi":"","year":null,"title":"Qwen2-Audio Technical Report","work_id":"c249e63c-cf40-408f-a4ff-fdf68e8cbeb8","ref_index":1,"cited_arxiv_id":"2407.10759","is_internal_anchor":true},{"doi":"","year":null,"title":"Time-masked trans- formers with lightweight test-time adaptation for neural speech decoding.arXiv preprint arXiv:2507.02800,","work_id":"2391c167-698f-488d-9bc8-a6662d489b71","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Towards an end-to-end framework for invasive brain signal decoding with large language models.arXiv preprint arXiv:2406.11568,","work_id":"af8471c4-0b90-4b01-a172-e7abcdf7c548","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1904,"title":"The Curious Case of Neural Text Degeneration","work_id":"1ef2ec4a-db7b-42b2-9416-0f1bb628c3c8","ref_index":4,"cited_arxiv_id":"1904.09751","is_internal_anchor":true},{"doi":"10.1101/2022.04","year":2022,"title":"Stabilizing brain- computer interfaces through alignment of latent dynamics","work_id":"42f27d99-f1ac-42e6-8684-07beea8891c1","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":25,"snapshot_sha256":"84fa834ea3609f2b716136e5e4fe01ff8fd85c6636e61068642d51313a710e02","internal_anchors":8},"formal_canon":{"evidence_count":2,"snapshot_sha256":"e952c07c7b922b9384a8363b223a8a6b4f3f67eff391543636779e235a365d04"},"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":"2511.21740","created_at":"2026-05-17T23:39:17.016909+00:00"},{"alias_kind":"arxiv_version","alias_value":"2511.21740v5","created_at":"2026-05-17T23:39:17.016909+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.21740","created_at":"2026-05-17T23:39:17.016909+00:00"},{"alias_kind":"pith_short_12","alias_value":"PWZJFKS5H54W","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"PWZJFKS5H54WFOTL","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"PWZJFKS5","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2605.10688","citing_title":"DANCE: Detect and Classify Events in EEG","ref_index":123,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00025","citing_title":"MoDAl: Self-Supervised Neural Modality Discovery via Decorrelation for Speech Neuroprosthesis","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS","json":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS.json","graph_json":"https://pith.science/api/pith-number/PWZJFKS5H54WFOTLR7BPMRXGOS/graph.json","events_json":"https://pith.science/api/pith-number/PWZJFKS5H54WFOTLR7BPMRXGOS/events.json","paper":"https://pith.science/paper/PWZJFKS5"},"agent_actions":{"view_html":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS","download_json":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS.json","view_paper":"https://pith.science/paper/PWZJFKS5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2511.21740&json=true","fetch_graph":"https://pith.science/api/pith-number/PWZJFKS5H54WFOTLR7BPMRXGOS/graph.json","fetch_events":"https://pith.science/api/pith-number/PWZJFKS5H54WFOTLR7BPMRXGOS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS/action/storage_attestation","attest_author":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS/action/author_attestation","sign_citation":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS/action/citation_signature","submit_replication":"https://pith.science/pith/PWZJFKS5H54WFOTLR7BPMRXGOS/action/replication_record"}},"created_at":"2026-05-17T23:39:17.016909+00:00","updated_at":"2026-05-17T23:39:17.016909+00:00"}