EconDL is a comprehensive resource detailing applications of Deep Learning in Economics. This is a companion website to the paper Deep Learning for Economists and aims to be a go-to resource for economists and other social scientists for applying tools provided by deep learning in their research.
This website contains user-friendly software and dataset resources, and a knowledge base that goes into considerably more technical depth than is feasible in a review article. The demos implement various applications explored in the paper, largely using open-source packages designed with economists in mind. They require little background and will run in the cloud with minimal compute, allowing readers with no deep learning background to gain hands-on experience implementing the applications covered in the review.
Lecture 3: Covers language models before transformers: RNNs, Seq2Seq, LSTM models
Lecture 4: Covers transformer architecture in general and begins discussion on transformer language models
Lecture 7: Covers the basics of training and optimizing neural networks