Gašper Beguš an Assistant Professor at the Department of Linguistics at UC Berkeley. His research focuses on developing deep learning models for speech data. More specifically, he trains models to learn representations of spoken words from raw audio outputs. He combines machine learning with behavioral experiments and statistical models to better understand how neural networks learn internal representations in speech and how humans learn to speak. He has worked and published on sound systems of various language families such as Indo-European, Caucasian, and Austronesian languages.
In a recent set of papers (here and here), he proposes that language acquisition can be modeled with Generative Adversarial Networks and propose a technique for exploring the relationship between learned representations and latent space in deep convolutional networks.
Beguš will be building a lab on speech and computation at Berkeley.