Regular biography
Jonathan Frankle is a Fellow at Harvard University's John A. Paulson School of Engineering and Applied Sciences, affiliated with the Computer Science department. His research interests include Applied Mathematics, Machine Learning, Artificial Intelligence, and related fields such as Computation and Society, Computational and Data Science, Science, Technology, Innovation, and Public Policy, and Science and Technology Policy. He is involved in teaching within the Computer Science department and can be contacted via email at jfrankle@seas.harvard.edu.
Scholar-generated biography
Jonathan Frankle is a researcher specializing in deep learning, with a focus on neural network pruning and optimization. His work explores techniques such as the Lottery Ticket Hypothesis, which identifies sparse, trainable subnetworks within larger neural networks. Frankle's research also addresses the challenges of training and fine-tuning models, including the role of BatchNorm and the impact of initialization on pruning effectiveness. Additionally, he investigates the broader implications of AI systems, such as facial recognition bias and the ethical considerations of secret processes. His contributions span both theoretical and applied aspects of deep learning, with an emphasis on improving model efficiency and accountability.