D
David Bau
Researcher at Massachusetts Institute of Technology
Publications - 84
Citations - 10608
David Bau is an academic researcher from Massachusetts Institute of Technology. The author has contributed to research in topics: Interpretability & Generative model. The author has an hindex of 35, co-authored 84 publications receiving 8193 citations. Previous affiliations of David Bau include Google & Business International Corporation.
Papers
More filters
Proceedings ArticleDOI
Network Dissection: Quantifying Interpretability of Deep Visual Representations
TL;DR: This work uses the proposed Network Dissection method to test the hypothesis that interpretability is an axis-independent property of the representation space, then applies the method to compare the latent representations of various networks when trained to solve different classification problems.
Proceedings ArticleDOI
Explaining Explanations: An Overview of Interpretability of Machine Learning
TL;DR: In an effort to create best practices and identify open challenges, the authors describe foundational concepts of explainability and show how they can be used to classify existing literature, and discuss why current approaches to explanatory methods especially for deep neural networks are insufficient.
Posted Content
Explaining Explanations: An Overview of Interpretability of Machine Learning
TL;DR: In an effort to create best practices and identify open challenges, the authors provide a definition of explainability and show how it can be used to classify existing literature, and discuss why current approaches to explanatory methods especially for deep neural networks are insufficient.
Journal ArticleDOI
Semantic photo manipulation with a generative image prior
David Bau,Hendrik Strobelt,William Peebles,Jonas Wulff,Bolei Zhou,Jun-Yan Zhu,Antonio Torralba +6 more
TL;DR: The authors adapts the image prior learned by GANs to image statistics of an individual image, which can accurately reconstruct the input image and synthesize new content consistent with the appearance of the original image.