Skip to content

Latest commit

 

History

History
160 lines (119 loc) · 1.73 KB

File metadata and controls

160 lines (119 loc) · 1.73 KB

PyData 2016 - London

Presentations

Probablistic Programming Data Science with PyMC3 (Thomas Wiecki)
  https://www.youtube.com/watch?v=LlzVlqVzeD8
  https://docs.google.com/presentation/d/1QNxSjDHJbFL7vFwQHHheeGmBHEJAo39j28xdObFY6Eo/edit
  https://gist.github.com/anonymous/9287a213fe188a79d7d7774eef79ad4d

  • prob_diff = np.mean(trace['mean_in_sample'] < trace['mean_out_of_sample])
  • ppc = pm.sample_ppc(trace, model=model_name, samples=500); pred = ppc['out'].mean(axis=0)
  • Example of Bayesian Neural Network in the above gist
  • Example of mapping out the entire classifier space, including the uncertainty.
  • Example of using ADVI, and mini-batch ADVI!
  • ADVI must be continuous.
  • "Puppy Book" code ported to PyMC3
  • https://github.com/aloctavodia/Doing_bayesian_data_analysis

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

Tutorials

Deep learning tutorial - advanced techniques (Geoffrey French)
https://www.youtube.com/watch?v=DlNR1MrK4qE

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)

**** (__)