Deep Learning for Economists

Harvard FAS

Unstructured data – e.g., text, images, audio, and video – play a central role in economic analyses. Unstructured data are too high-dimensional and complex to use directly in empirical analyses, and instead economists extract low-dimensional, interpretable features. Deep neural networks are the state-of-the-art method for extracting this structured data from unstructured inputs. This course covers state-of-the-art deep neural network architectures and examines a range of applications in…