Charles Tripp is a machine learning researcher who has spent over a decade on professional research and development projects. After graduating from Stanford with his PhD, he founded Terrain Data, Inc., a Silicon Valley data science startup. After guiding that business to a successful acquisition, he returned to research to pursue a career focused on the aspects of his work he enjoys most.
At NREL, Charles applies machine learning techniques to renewable energy and energy efficiency. He works to bridge the gap between emerging artificial intelligence and machine learning technologies and real-world practical applications He is particularly interested in applications of, and inquiry into, the nature of reinforcement learning and derivative-free optimization algorithms. As a byproduct, he is additionally interested in the fusion of classical techniques such as analytical models, linear and model predictive control, dynamic programming, with modern machine learning methods. Charles is a Bayesian and Decision Analyst at heart, and is always looking to apply solid decision science to machine learning problems.
Research Interests
Reinforcement Learning
Derivative-Free Optimization
Probabilistic Modeling
Stochastic Simulation
Bayesian Methods
Non-Convex Optimization
Non-Linear Control
Kalman Filters
Particle Filters
Education
Ph.D., Electrical Engineering, Stanford University
M.S., Electrical Engineering, Stanford University
B.S., Electrical Engineering, Rice University
Associations and Memberships
Member, ACM
Member, INFORMS
Member, SIGPLAN
Featured Work
FRC: A high-performance concurrentparallel deferred reference counter for C++, ACM SIGPLAN International Symposium on Memory Management (2018)
Approximate Kalman filter Q-learning for continuous state-space MDPs, Proceedings of the Twenty-Ninth Conference on Uncertainty in Artificial Intelligence (2013)
Heuristics for Large Scale Dynamic Programming, Thesis (Ph.D.) Stanford University Department of Electrical Engineering (2013)
Backtracking for More Efficient Large Scale Dynamic Programming, 11th International Conference on Machine Learning and Applications (2012)
Patents
Finding Similar Images Based on Extracting Keys from Images, US Patent 9384519B (2016)
ACID database: Adaptive Buffered B-Tree, Graphical Query Management System, US Patent Pending 62181622 (2015)
Share