Yoel Zeldes • 7/31/2018

Using Uncertainty to Interpret your Model

This technical article discusses the importance of model interpretability in complex deep neural networks. It focuses on using uncertainty estimation methods to debug models, handle out-of-distribution examples, and improve decision-making in high-risk applications like healthcare and autonomous vehicles. The post covers different types of uncertainty (epistemic) and their practical uses for practitioners.

0 comments

#Machine Learning #Deep Neural Networks #Model Interpretability