Ian Char

About Me

I am a machine learning scientist at Lila Sciences. My research interests include reinforcement learning, Bayesian optimization, and uncertainty quantification. I am particularly interested in how these topics can accelerate scientific discovery.

Previously, I was a PhD student in the Machine Learning Department at Carnegie Mellon University advised by Jeff Schneider. There, my thesis revolved around model-based reinforcement learning applied to tokamak control for nuclear fusion.

Previously, I earned a MS in Applied Mathematics from University of Colorado Boulder and was advised by Manuel Lladser. I also went to University of Colorado Boulder for my undergraduate and earned a BS in both applied math and computer science.

Selected Publications and Pre-Prints

• PID-Inspired Inductive Biases for Deep Reinforcement Learning in Partially Observable Control Tasks
  Ian Char, Jeff Schneider
  Advances in Neural Information Processing Systems. NeurIPS 2023.
  [PDF] [Code]

• Full Shot Predictions for the DIII-D Tokamak via Deep Recurrent Networks
  Ian Char, Youngseog Chung, Joseph Abbate, Egemen Kolemen, Jeff Schneider
  Pre-Print.
  [PDF]

• Offline Model-Based Reinforcement Learning for Tokamak Control
  Ian Char, Joseph Abbate, Laszlo Bardoczi, Mark D. Boyer, Youngseog Chung, Rory Conlin, Keith Erickson, Viraj Metha, Nathan Richner, Egemen Kolemen, Jeff Schneider
  Learning for Discovery and Control 2023
  [PDF]

• Near-optimal Policy Identification in Active Reinforcement Learning
  Xiang Li, Viraj Mehta, Johannes Kirschner, Ian Char, Willie Neiswanger, Jeff Schneider, Andreas Krause, Ilija Bogunovic
  International Conference on Learning Representations 2023 Notable-Top-5% (Oral)
  [PDF]

• Exploration via Planning for Information about the Optimal Trajectory
  Viraj Mehta, Ian Char, Joseph Abbate, Rory Conlin, Mark D Boyer, Stefano Ermon, Jeff Schneider, Willie Neiswanger
  Advances in Neural Information Processing Systems. NeurIPS 2022.
  [PDF] [Code]

• Beyond Pinball Loss: Quantile Methods for Calibrated Uncertainty Quantification
  Youngseog Chung, Willie Neiswanger, Ian Char, Jeff Schneider
  Advances in Neural Information Processing Systems. NeurIPS 2021.
  [PDF] [Code]

• BATS: Best Action Trajectory Stitching
  Ian Char^*, Viraj Mehta^*, Adam Villaflor, John Dolan, Jeff Schneider
  NeurIPS 2021 Offline Reinforcement Learning Workshop
  [PDF]

• Offline Contextual Bayesian Optimization
  Ian Char, Youngseog Chung, Willie Neiswanger, Kirthevasan Kandasamy, Andrew Oakleigh Nelson, Mark D Boyer, Egemen Kolemen, Jeff Schneider
  Advances in Neural Information Processing Systems. NeurIPS 2019.
  [PDF] [Code]

Software

• Uncertainty Toolbox
  A python toolbox for predictive uncertainty quantification, calibration, metrics, and visualizations. Also: a glossary of useful terms and a collection of relevant papers and references.
  [Code] [Website]

Awards

• Teaching Assistant of the Year
  Machine Learning Dperatment, Carnegie Mellon Univeristy (2021-2022)

• NSF GRFP (2018)
  NSF Graduate Research Fellowship Program, 2018

• Outstanding Graduate for Academic Achievement
  University of Colorado at Boulder, May 2018
  Awarded to students graduating with a 4.0.

Teaching

• (10-606/607) Mathematical/Computational Foundations for Machine Learning
  Teaching Assistant
  Carnegie Mellon University, Fall 2021

• (10-716) Advanced Machine Learning: Theory and Methods
  Teaching Assistant
  Carnegie Mellon University, Spring 2020

• (APPM 4350) Fourier Series and Boundary Value Problems
  Teaching Assistant
  University of Colorado Boulder, Fall 2016

Service

• Machine Learning Department Social Committee
  2019-2023

• Reviewer for NeurIPS 2023
  NeurIPS 2023

• Reviewer for Real World Experiment Design and Activate Learning Workshop
  ICML 2022

Powered by Jekyll and Minimal Light theme.