{"paper":{"title":"LMU-Based Sequential Learning and Posterior Ensemble Fusion for Cross-Domain Infant Cry Classification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A framework combining multi-branch CNNs, Legendre Memory Units, and entropy-gated fusion improves cross-domain classification of infant cry causes.","cross_cats":["cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Hilmi R. Dajani, Marco Janeczek, Martin Bouchard, Niloofar Jazaeri","submitted_at":"2026-02-24T23:44:41Z","abstract_excerpt":"Decoding infant cry causes remains challenging for healthcare monitoring due to short nonstationary signals, limited annotations, and strong domain shifts across infants and datasets. We propose a compact acoustic framework that fuses mel-frequency cepstral coefficients (MFCCs), short-time Fourier transform (STFT) features, and fundamental-frequency (F0) contours within a multi-branch convolutional neural network (CNN) encoder, and models temporal dynamics using an enhanced Legendre Memory Unit (LMU). Compared to LSTMs, the LMU backbone provides stable sequence modeling with substantially fewe"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on Baby2020 and Baby Crying demonstrate improved macro-F1 under cross-domain evaluation, along with leakage aware splits and real-time feasibility for on-device monitoring.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the combination of MFCC/STFT/F0 features, LMU temporal modeling, and entropy-gated posterior fusion will reliably mitigate dataset bias and generalize to unseen infants without post-hoc tuning or data leakage.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LMU-based CNN with calibrated posterior ensemble fusion reports improved macro-F1 for cross-domain infant cry classification on Baby2020 and Baby Crying datasets.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A framework combining multi-branch CNNs, Legendre Memory Units, and entropy-gated fusion improves cross-domain classification of infant cry causes.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"5f2a57f43a466d5540d8871aef3912ed3efa0184d0b35ee9ac06aff022f4a4e7"},"source":{"id":"2603.02245","kind":"arxiv","version":3},"verdict":{"id":"19a675c6-2416-4ce5-b277-a33d63aa5781","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T19:29:14.725852Z","strongest_claim":"Experiments on Baby2020 and Baby Crying demonstrate improved macro-F1 under cross-domain evaluation, along with leakage aware splits and real-time feasibility for on-device monitoring.","one_line_summary":"LMU-based CNN with calibrated posterior ensemble fusion reports improved macro-F1 for cross-domain infant cry classification on Baby2020 and Baby Crying datasets.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the combination of MFCC/STFT/F0 features, LMU temporal modeling, and entropy-gated posterior fusion will reliably mitigate dataset bias and generalize to unseen infants without post-hoc tuning or data leakage.","pith_extraction_headline":"A framework combining multi-branch CNNs, Legendre Memory Units, and entropy-gated fusion improves cross-domain classification of infant cry causes."},"references":{"count":23,"sample":[{"doi":"","year":2022,"title":"Automated newborn cry diagnostic system using machine learning,","work_id":"b2881f5f-a580-4222-97df-1e47f6f73ce5","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Pain cues override identity cues in baby cries,","work_id":"2fca1239-4fbc-4fd6-8744-8ed8488d302d","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Machine learning-based cry diagnostic system for identifying septic newborns,","work_id":"d3fd7935-2fef-4b8d-bee8-cf72c1c5c083","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Infant crying detection in real-world environments,","work_id":"c88ef53c-2cd1-47e3-9359-eb2ae853f111","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Design and implementation of infant crying monitoring and analysis system,","work_id":"ab73bcb8-2858-4755-ad49-162cd4fc9e80","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":23,"snapshot_sha256":"7345d3e5d963fd2bdbae4291720f86f34bee304b65adcf5c50020cb261021749","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"}