Sequential decision making using online variational Bayes

Title: Sequential decision making using online variational Bayes

Abstract: After reviewing the basics of of approximate sequential Bayesian inference for state space models, I will present several algorithms we have developed [1-4] to make this process more efficient and robust, by combining various tricks (eg. linearization, low rank matrix updates, natural gradients, generalized Bayes). I then show how this can be applied to the problem of learning neural networks from streaming, non-stationary data, which is needed when tackling various kinds of sequential decision making problems, such as bandits, Bayesian optimization, and RL.

[1] https://arxiv.org/abs/2405.19681
[2] https://arxiv.org/abs/2405.05646
[3] https://arxiv.org/abs/2305.19535
[4] https://arxiv.org/abs/2112.00195

Bio: Kevin was born in Ireland, but grew up in England. He got his BA from U. Cambridge, his MEng from U. Pennsylvania, and his PhD from UC Berkeley. He then did a postdoc at MIT, and was an associate professor of computer science and statistics at the University of British Columbia in Vancouver, Canada, from 2004 to 2012. After getting tenure, he went to Google in California on his sabbatical and then ended up staying. Kevin is now a Principal Research Scientist / Director, and manages a team of 28 researchers and engineers at Google Deepmind; the team works on generative models (including diffusion and LLMs), reinforcement learning, robotics, Bayesian inference,  and various other topics. Kevin has published over 140 papers in refereed conferences and journals, as well 3 textbooks on machine learning published in 2012, 2022 and 2023 by MIT Press. (The 2012 book was awarded the DeGroot Prize for best book in the field of Statistical Science.) Kevin was the (co) Editor-in-Chief of JMLR 2014-2017, and is currently a senior  editor at ACM/IMS Journal of Data Science and at Foundations and Trends in Machine Learning.