ProbAI 2019

Recorded lectures and tutorials, slides and program code from the very first Nordic Probabilistic AI School (ProbAI 2019).

Lectures and Tutorials

• Day 1 (June 3):
  • [Lecture, Materials] Introduction to Probabilistic Modelling
  • [Tutorial, Materials] Probabilistic Programming
• Day 2 (June 4):
  • [Lecture, Materials] Variational Inference and Optimization 1
  • [Tutorial, Materials] Variational Inference and Probabilistic Programming 1
• Day 3 (June 5):
  • [Lecture, Materials] Variational Inference and Optimization 2
  • [Tutorial, Materials] Variational Inference and Probabilistic Programming 2
• Day 4 (June 6):
  • [Lecture, Materials] Introduction to Deep Learning & Deep Generative Models
  • [Tutorial, Materials] Deep Latent Variable Models for Imputation of Incomplete Data Sets (Imputations with MIWAE)
• Day 5 (June 7):
  • [Lecture, Materials] Deep Generative Models
  • [Tutorial, Materials] Bayesian Sparsification of Neural Networks

Google Colab

If your local computer doesn't have all the software packages and you are not able to finish the installation, you can try Google Colab.

Start the Google Colab notebook with the following line to install the necessary packages !pip install -q --upgrade pyro-ppl torch.

• Day 1 - students_PPLs_Intro.ipynb
• Day 1 - students_Bayesian_regression.ipynb
• Day 2 - students_simple_model.ipynb
• Day 2 - students_lin_reg.ipynb
• Day 3 - student_simple_model.ipynb
• Day 3 - student_BBVI.ipynb
• Day 3 - Bayesian_linear_regression.ipynb
• Day 3 - VAE.ipynb
• Day 3 - FA.ipynb

Talks

• Day 2:
  • Evolving Deep Neural Networks by Keith L. Downing
• Day 3:
  • Value of Information by Jo Eidsvik
  • Bayesian Methods for Rank and Preference Data: From Recommender Systems to Cancer Genomics by Valeria Vitelli
• Day 4:
  • Combining Model and Parameter Uncertainty in BNNs by Aliaksandr Hubin