Using Logic to Understand Learning
TimeTuesday, July 12th4pm - 4:30pm PDT
Location3003, Level 3
Special Session (Research)
DescriptionThe second talk will show how ideas from EDA and hardware design can be used to gain insights into fundamental questions in deep learning such as why deep neural networks are able to generalize well on unseen data when they have sufficient capacity to simply memorize their training sets. These insights can lead to more efficient algorithms for deep learning in resource constrained environments.