arXiv subjects

Makoto Sakai

Publications and source records attributed to Makoto Sakai.

3 recordsLinked to original sources

arXiv2026

QC-GAN: A Parameter-Efficient Quaternion Conformer GAN for High-Fidelity Speech Enhancement

We propose a parameter-efficient speech enhancement framework, Quaternion Conformer GAN (QC-GAN), which combines a Quaternion Conformer generator with MetricGAN-based training. The Hamilton product encodes the magnitude and phase via structured weight sharing, reducing the number of layer parameters while preserving their interdependencies. A metric-learning discriminator was employed to maximize perceptual quality by optimizing the approximate perceptual evaluation scores. On the VoiceBank+DEMAND dataset, QC-GAN achieved a Perceptual Evaluation of Speech Quality (PESQ) score of 3.48 with only 0.89M parameters, delivering a performance comparable to state-of-the-art models at less than half their size. A 35K-parameter variant achieved a PESQ score of 3.23, surpassing conventional methods with significantly fewer parameters. Evaluation on the DNS-Challenge 3 dataset further confirmed generalization to real-world conditions.

cs.SD↗

arXiv2024

NAIST Simultaneous Speech Translation System for IWSLT 2024

This paper describes NAIST's submission to the simultaneous track of the IWSLT 2024 Evaluation Campaign: English-to-{German, Japanese, Chinese} speech-to-text translation and English-to-Japanese speech-to-speech translation. We develop a multilingual end-to-end speech-to-text translation model combining two pre-trained language models, HuBERT and mBART. We trained this model with two decoding policies, Local Agreement (LA) and AlignAtt. The submitted models employ the LA policy because it outperformed the AlignAtt policy in previous models. Our speech-to-speech translation method is a cascade of the above speech-to-text model and an incremental text-to-speech (TTS) module that incorporates a phoneme estimation model, a parallel acoustic model, and a parallel WaveGAN vocoder. We improved our incremental TTS by applying the Transformer architecture with the AlignAtt policy for the estimation model. The results show that our upgraded TTS module contributed to improving the system performance.

cs.CL↗

arXiv2011

On the curvature of some free boundaries in higher dimensions

It is known that any subharmonic quadrature domain in two dimensions satisfies a natural inner ball condition, in other words there is a specific upper bound on the curvature of the boundary. This result directly applies to free boundaries appearing in obstacle type problems and in Hele-Shaw flow. In the present paper we make partial progress on the corresponding question in higher dimensions. Specifically, we prove the equivalence between several different ways to formulate the inner ball condition, and we compute the Brouwer degree for a geometrically important mapping related to the Schwarz potential of the boundary. The latter gives in particular a new proof in the two dimensional case.

math.AP↗