Quantum systems can generate patterns in data that are hard for classical systems to generate. This talk shows that quantum information processors can also recognize and classify patterns that are hard for classical learning systems to find. The ability of quantum computers to perform linear algebra operations such as matrix inversion exponentially faster than classical computers translates into the ability to perform multiple data analysis tasks such as principal component analysis more efficiently. The talk discusses how to load classical data into quantum learners, and explores prospects for deep quantum learning.