2026-10-10T20:01:25Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33975704 2026-09-23T17:31:46Z category_7 category_45 category_734 portal_5 item_type_3 month_year_09_2026

<p>Evaluation of acoustic reconstruction on high-loss emotional samples.</p> Yan Zhu (112437) Yu Wang (12152) Yijin Zhou (25099121) Bomin Liu (23019004) Rui Zhou (38109) Medicine Sociology Biological Sciences not elsewhere classified word error rate uniformly averaging per trainable lookup embedding shot speech synthesis may introduce interference continuous representations using closely coupled pathways assigns larger weights achieved substantial progress grained emotion space batch loss statistics improve content preservation means clustering based grained emotional scenarios proposed method achieves div >< p loss emotional samples threshold reweighting estimates adaptive gated fusion proposed method adaptive loss semantic content grained zero emotional expression emotional conditions based zero two conditions study proposes results suggest resulting labels respective information processes semantic parallel cross length expansion fixed subset feature decoupling evaluated conditions energy features disjoint evaluation deep learning controlled text classifier filtering cepstral distortion attention streams <p>Evaluation of acoustic reconstruction on high-loss emotional samples.</p> 2026-09-23T17:31:30Z Dataset Dataset 10.1371/journal.pone.0358777.t007 https://figshare.com/articles/dataset/_p_Evaluation_of_acoustic_reconstruction_on_high-loss_emotional_samples_p_/33975704 CC BY 4.0