Modeling Individuals Without Data via Secondary Task Transfer Learning

Explore a novel transfer learning method for modeling individual human behavior in this 34-minute conference talk from the Toronto Machine Learning Series. Discover how to produce high-quality deep neural networks (DNNs) for unseen target tasks using limited data from a secondary task. Learn about the specialized transfer learning representation and Monte Carlo Tree Search (MCTS) approach that outperforms standard methods in human modeling domains such as financial health and video game design. Gain insights from Matthew Guzdial, an award-winning Assistant Professor and CIFAR AI Chair at the University of Alberta, as he presents innovative solutions for modeling individuals without extensive datasets.