We have updated our software for improved interpretability since this post was written. For the latest on this topic, read our new series on model-agnostic interpretability.  While some machine learning models – like decision trees – are transparent, the majority of models used today – like deep neural networks, random forests, gradient boosting