Ian Goodfellow

Ian Goodfellow is a computer scientist best known for his research in the field of deep learning, including the invention of generative adversarial networks (GANs) in 2014. Goodfellow developed the first defenses against adversarial examples, was among the first to study the security and privacy of neural networks, and helped popularize the field of machine learning security and privacy. He is the lead author of the MIT Press textbook Deep Learning alongside Yoshua Bengio and Aaron Courville and wrote the deep learning chapter in the textbook Artificial Intelligence: A Modern Approach. In 2017, he was listed among MIT Technology Review's 35 Innovators under 35, and in 2019, he was included on Foreign Policy's list of 100 Global Thinkers.

Goodfellow attended Stanford University, completing a Bachelor and Master of Science in computer science between 2004 and 2009. While at Stanford, he studied with Andrew Ng and Gary Bradski. In 2010, he began a PhD in machine learning at Université de Montréal. He submitted his thesis, titled "Deep learning of representations and its application to computer vision," in April 2014. His thesis advisor and co-advisor were Yoshua Bengio and Aaron Courville, respectively. During his PhD studies, Goodfellow invented maxout networks, generative adversarial networks, multi-prediction deep-Boltzmann machines, and a new fast inference algorithm for spike-and-slab sparse coding, and led the development and popularization of Pylearn2.

Since leaving the Université de Montréal, Goodfellow has worked at Google twice (July 2014 - March 2016 & March 2017 - March 2019), OpenAI (March 2016 - March 2017), and Apple (March 2019 - May 2022) as the director of machine learning in the special projects group. In June 2022, Goodfellow joined DeepMind, working as a research scientist in Oriol Vinyals' deep learning team.