Deep Probabilistic Programming Languages: A Qualitative Study
This addresses usability issues for researchers and practitioners in machine learning and programming languages, but is incremental as it focuses on qualitative analysis rather than new methods.
The paper tackles the challenge of deep probabilistic programming languages being difficult to use and understand by explaining them and characterizing their current strengths and weaknesses, but does not report concrete numerical results.
Deep probabilistic programming languages try to combine the advantages of deep learning with those of probabilistic programming languages. If successful, this would be a big step forward in machine learning and programming languages. Unfortunately, as of now, this new crop of languages is hard to use and understand. This paper addresses this problem directly by explaining deep probabilistic programming languages and indirectly by characterizing their current strengths and weaknesses.