Unsupervised speech enhancement based on kurtosis and deep prior accepted to Acoustical Science and Technology, 2025 Visual onoma-to-wave accepted to ICASSP 2023 Wave-Trainer-Fit: Neural Vocoder with Trainable Prior and Fixed-Point Iteration towards High-Quality Speech Generation from SSL features Under review 個人サイトを google sites から github.io + al-folio に移行しました 思ったよりもすんなりできました．
