{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FM3JGBCPRYDFHK5RXYYSCVSC7N","short_pith_number":"pith:FM3JGBCP","schema_version":"1.0","canonical_sha256":"2b3693044f8e0653abb1be31215642fb7879cee28a835f87bda684bb54054b8d","source":{"kind":"arxiv","id":"2605.17405","version":1},"attestation_state":"computed","paper":{"title":"A Distribution Matching Approach to Neural Piano Transcription with Optimal Transport","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Automatic piano transcription improves when framed as optimal transport between note distributions rather than frame-by-frame classification.","cross_cats":["cs.MM"],"primary_cat":"cs.SD","authors_text":"Dichucheng Li, Kazuyoshi Yoshii, Raynaldi Lalang, Weixing Wei","submitted_at":"2026-05-17T12:01:44Z","abstract_excerpt":"This paper describes a novel paradigm that formalizes automatic piano transcription (APT) as an optimal transport (OT) problem, not as a frame-level multi-label binary classification problem. Our method learns to minimize the cost of transporting a predicted distribution of note events to the ground-truth distribution over time and frequency. The OT loss can thus accommodate temporal misalignment, leading to perceptually relevant optimization. We also propose a convolutional recurrent neural network (CRNN) with a harmonics-aware attention mechanism to capture the spectro-temporal dependencies "},"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":"2605.17405","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SD","submitted_at":"2026-05-17T12:01:44Z","cross_cats_sorted":["cs.MM"],"title_canon_sha256":"d9359b07a5d1c4f081469e2c2141444dcd67a9adcf8a0e823f42dd16fef707b5","abstract_canon_sha256":"2db5573339328817fa5db9d2ad28b4d5de1c29c0bd62ba2de2ac333cf9bad8a7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:56.794937Z","signature_b64":"uXROsRyaSNDK9R+5uDBOFWJAfP8O3icd5l1Ng243Kb2mT0ssbtSVKkcG/jBEDc4eXrZZoCEy26EMAcL5aADtBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2b3693044f8e0653abb1be31215642fb7879cee28a835f87bda684bb54054b8d","last_reissued_at":"2026-05-20T00:03:56.794282Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:56.794282Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Distribution Matching Approach to Neural Piano Transcription with Optimal Transport","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Automatic piano transcription improves when framed as optimal transport between note distributions rather than frame-by-frame classification.","cross_cats":["cs.MM"],"primary_cat":"cs.SD","authors_text":"Dichucheng Li, Kazuyoshi Yoshii, Raynaldi Lalang, Weixing Wei","submitted_at":"2026-05-17T12:01:44Z","abstract_excerpt":"This paper describes a novel paradigm that formalizes automatic piano transcription (APT) as an optimal transport (OT) problem, not as a frame-level multi-label binary classification problem. Our method learns to minimize the cost of transporting a predicted distribution of note events to the ground-truth distribution over time and frequency. The OT loss can thus accommodate temporal misalignment, leading to perceptually relevant optimization. We also propose a convolutional recurrent neural network (CRNN) with a harmonics-aware attention mechanism to capture the spectro-temporal dependencies "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"This paper describes a novel paradigm that formalizes automatic piano transcription (APT) as an optimal transport (OT) problem, not as a frame-level multi-label binary classification problem. Our method learns to minimize the cost of transporting a predicted distribution of note events to the ground-truth distribution over time and frequency. The OT loss can thus accommodate temporal misalignment, leading to perceptually relevant optimization.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That minimizing optimal transport cost between predicted and ground-truth note distributions produces perceptually superior transcriptions and that a harmonics-aware attention mechanism in the CRNN sufficiently captures the necessary spectro-temporal dependencies in piano signals.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Introduces an optimal transport loss for neural piano transcription to accommodate temporal misalignments, achieving state-of-the-art onset detection on the MAESTRO dataset via a harmonics-aware CRNN.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Automatic piano transcription improves when framed as optimal transport between note distributions rather than frame-by-frame classification.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"b7a77a299660483ffe5de688075a16aadc2908171f46a2a29deff84cc2c3f084"},"source":{"id":"2605.17405","kind":"arxiv","version":1},"verdict":{"id":"0ad74f8e-023e-4f80-85fe-a048e5acf83c","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T22:42:40.037720Z","strongest_claim":"This paper describes a novel paradigm that formalizes automatic piano transcription (APT) as an optimal transport (OT) problem, not as a frame-level multi-label binary classification problem. Our method learns to minimize the cost of transporting a predicted distribution of note events to the ground-truth distribution over time and frequency. The OT loss can thus accommodate temporal misalignment, leading to perceptually relevant optimization.","one_line_summary":"Introduces an optimal transport loss for neural piano transcription to accommodate temporal misalignments, achieving state-of-the-art onset detection on the MAESTRO dataset via a harmonics-aware CRNN.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That minimizing optimal transport cost between predicted and ground-truth note distributions produces perceptually superior transcriptions and that a harmonics-aware attention mechanism in the CRNN sufficiently captures the necessary spectro-temporal dependencies in piano signals.","pith_extraction_headline":"Automatic piano transcription improves when framed as optimal transport between note distributions rather than frame-by-frame classification."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.17405/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T23:01:19.642234Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T22:52:12.681033Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T21:41:57.750608Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T21:33:23.693799Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"b9ff83c3b5a93095127292e401f1e1eedc447ae7b9d91d60c9b688b65d2f7e09"},"references":{"count":35,"sample":[{"doi":"","year":null,"title":"INTRODUCTION In automatic piano transcription (APT) that aims to estimate a piano-roll representation (MIDI data) from a music record- ing [1], deep learning models play a central role. The standard p","work_id":"a406d2db-dbf2-4194-8be0-b55735b597c2","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"On- sets&Frames model [2] predicted note onsets and frame-wise activations separately and then combining them afterwards","work_id":"d6a44adb-c0c5-4ebb-865d-69a7d7064011","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"PROPOSED METHOD This section introduces the problem formulation of APT with OT and the proposed model architecture. 3.1. Problem formulation LetX∈R T×F be the time-frequency representation of a targe ","work_id":"e0232347-75d3-421d-b3e3-b128587e8ccf","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Each layer uses a7×7kernel and strides of (1,2), (1,2), and (2,1), progressively downsampling the tem- poral and frequency dimensions by factors of 2 and 4","work_id":"96e1fe56-c94b-4cf7-8485-a58f3006f54d","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"EV ALUA TION This section reports comparative evaluation conducted for val- idating the effectiveness of the OT loss in APT. 4.1. Experimental Conditions The MAESTRO [26] dataset was used for evaluati","work_id":"59f99d1b-cc4e-44f5-980e-457a6d077c9b","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":35,"snapshot_sha256":"f37732044f5e5664563c0872f7dd754145189c2060fd9580273cbd32b20f9175","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"7cdd3499034c5028c4af2ada94df4914b21bc80340742ab2f0fff69cc7ab09f0"},"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":"2605.17405","created_at":"2026-05-20T00:03:56.794409+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.17405v1","created_at":"2026-05-20T00:03:56.794409+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17405","created_at":"2026-05-20T00:03:56.794409+00:00"},{"alias_kind":"pith_short_12","alias_value":"FM3JGBCPRYDF","created_at":"2026-05-20T00:03:56.794409+00:00"},{"alias_kind":"pith_short_16","alias_value":"FM3JGBCPRYDFHK5R","created_at":"2026-05-20T00:03:56.794409+00:00"},{"alias_kind":"pith_short_8","alias_value":"FM3JGBCP","created_at":"2026-05-20T00:03:56.794409+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N","json":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N.json","graph_json":"https://pith.science/api/pith-number/FM3JGBCPRYDFHK5RXYYSCVSC7N/graph.json","events_json":"https://pith.science/api/pith-number/FM3JGBCPRYDFHK5RXYYSCVSC7N/events.json","paper":"https://pith.science/paper/FM3JGBCP"},"agent_actions":{"view_html":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N","download_json":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N.json","view_paper":"https://pith.science/paper/FM3JGBCP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.17405&json=true","fetch_graph":"https://pith.science/api/pith-number/FM3JGBCPRYDFHK5RXYYSCVSC7N/graph.json","fetch_events":"https://pith.science/api/pith-number/FM3JGBCPRYDFHK5RXYYSCVSC7N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N/action/storage_attestation","attest_author":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N/action/author_attestation","sign_citation":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N/action/citation_signature","submit_replication":"https://pith.science/pith/FM3JGBCPRYDFHK5RXYYSCVSC7N/action/replication_record"}},"created_at":"2026-05-20T00:03:56.794409+00:00","updated_at":"2026-05-20T00:03:56.794409+00:00"}