Developed efficient StyleGAN parameterizations and compact architectures for speech enhancement; introduced the HyperDomainNet model for multi-domain adaptation; explored key components of StyleGAN for effective domain adaptation; studied properties of StyleSpace directions; presented HiFi++ and FFC-SE models for speech enhancement.
Research Experience
Researched efficient parameterizations for GANs in image and speech generation, especially training on small datasets.
Education
Earned a PhD, focusing on 'Exploring efficient parameterizations for gans in image and speech generation', advisor information not provided.