Gasper Begus

I am an Assistant Professor at the Department of Linguistics at UC Berkeley, where I’m also affiliated with the Institute of Cognitive and Brain Sciences. Previously, I was an Assistant Professor at the University of Washington. Before that, I graduated with a Ph.D. from Harvard.

My research focuses on developing deep learning models for speech data and using well-understood dependencies in speech to interpret internal representations in deep neural networks. More specifically, I build models that learn representations of spoken words from raw audio inputs. I combine 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. I have 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), I propose 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.

I direct the Berkeley Speech and Computation Lab. Feel free to contact me if you’re interested in getting involved with the lab.

You can follow me on Twitter for latest updates.

Department of Linguistics

UC Berkeley

1203 Dwinelle Hall #2650

Berkeley, CA 94720-2650


Nov 25, 2020 New preprint out on CNNs, iterative learning & language change.
Nov 25, 2020 In Fall 2021, I plan to co-teach at Berkeley CS: EECS225d: Spoken Language Processing: Human and Computational Models (with Gopala Anumachipalli).
Nov 25, 2020 A new paper on estimating historical probabilities in sounds systems of world's languages to appear in Phonology
Nov 24, 2020 I gave three invited talks recently: at UCSC Phlunch, at UC Davis Phon Lab and at ASEF.
Oct 26, 2020 How to combine human behavioral experiments and CNNs? A new preprint is on arXiv
Sep 23, 2020 Can deep convolutional network learn identity-based patterns? A new preprint is on arXiv
Sep 15, 2020 I'll give a talk at MIT BCS Comp Lang
Jul 25, 2020 A deep neural network learns to produce new words
Jul 22, 2020 A new paper on lexical learning in GANs & latent space interpretability is out on arXiv.
Jul 8, 2020 A new paper is out in Frontiers in Artificial Intelligence.
Jul 1, 2020 I started a new job at UC Berkeley.