Held at the the 14th Session of the LSSTC Data Science Fellowship Program.
- Lecture 1: An introduction to Variational Autoencoders
- Lecture 2: Introduction to Probabilsitic Machine Learning and Considerations for Real Data
Each lecture is accompanied by coding exercises. The jupyer notebooks can be run on Google Colab. The underlying machine learning framework is pytorch.
- Notebook 1: From Autoencoders to Variational Autoencoders (VAEs) + Training a VAE on galaxy spectra
- Notebook 2: Improving VAEs with Normalizing Flows + denoising and inpainting masked and noisy Data