← Search

Fernanda Viegas

2 accepted papers

2018

Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)

ICML 2018oral

The interpretation of deep learning models is a challenge due to their size, complexity, and often opaque internal state. In addition, many systems, such as image classifiers, operate on low-level features rather than high-level concepts. To address these challenges, we introduce Concept Activation…