Gašper Beguš

Gašper Beguš is an Associate Professor of Linguistics in the Department of Linguistics at UC Berkeley. He obtained his undergraduate degree from the University of Ljubljana and his M.A. and Ph.D. from Harvard University. During his time at Harvard, he served as a Resident Tutor and Sophomore Advising Coordinator. He received several honors, including the Certificate of Teaching Excellence Award, the Mind Brain Behavior Graduate Student Award, and the Harvard Merit/Term-Time Award. He was also nominated for the Star Family Prize in Excellence in Advising. The central question of his research is how humans acquire speech and how this process can be computationally simulated. He combines experimental, statistical, and computational models to address this topic. He is currently exploring how the well-understood dependencies in speech data can help us understand what and how neural networks learn, as well as how they encode learning representations. His work has appeared in academic journals such as the Journal of Linguistics, the Journal of the Acoustical Society of America, and the Journal of the American Oriental Society.

Research projects: Professor Beguš’s research focuses on developing deep learning models for speech data. Specifically, he trains models to learn representations of spoken words from raw audio inputs. 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. His work has been published on the sound systems of various language families, including Indo-European, Caucasian, and Austronesian languages.