google brain interpretability

Today, we’re excited to publish “The Building Blocks of Interpretability,” a new Distill article exploring how feature visualization can combine together with other interpretability techniques to understand aspects of how networks make decisions. A Benchmark for Interpretability Methods in Deep Neural Networks Sara Hooker, Dumitru Erhan, Pieter-Jan Kindermans, Been Kim Google Brain shooker,dumitru,pikinder,beenkim@google.com Abstract We propose an empirical measure of the approximate accuracy of feature impor-tance estimates in deep neural networks. Been Kim, Google Brain, USA. To address these challenges, we introduce Concept Activation Vectors (CAVs), which provide an interpretation of a neural net's internal state in terms of human-friendly concepts.

I am a researcher at Google Brain working on training models that go beyond test-set accuracy to fulfill multiple desired criteria -- interpretable, compact, fair and robust. This researcher with Google Brain has mitigated this problem through translation and interpretability research to bridge the gap between AI and humans. Been Kim, Elena Glassman, Brittney Johnson and Julie Shah Pieter-Jan Kindermans, Sara Hooker, Julius Adebayo, Maximilian Alber, Kristof T. Schütt, Sven Dähne, Dumitru Erhan, Been Kim I am interested in designing high-performance machine learning methods that make sense The recent advancements in computation models … The Google Brain team conducted several experiments to evaluate the efficiency of TCAV compared to other interpretability methods. The ground truth of the experiment was that the image concept was more relevant than the caption concept.

The recent advancements in computation models and deep learning research have enabled the creation of highly sophisticated models that can include thousands of hidden layers and tens of millions of neurons. We show how to use CAVs as part of a technique, Testing with CAVs (TCAV), that uses directional derivatives to quantify the degree to which a user-defined concept is important to a classification result--for example, how sensitive a prediction of “zebra” is to the presence of stripes. ~ Kim B., Google Brain, Interpretable Machine Learning (ICML 2017) ... 2018, An Introduction to Machine Learning Interpretability Zhao Q., Hastie T., 2017, Causal Interpretations of Black-Box Models Kim B., Doshi-Velez F., 2017, Interpretable Machine Learning: The fuss, the concrete and the questions Google Brain is a machine intelligence research team, with a long research horizon and a primary focus on deep neural networks. Join to Connect. The key idea is to view the high-dimensional internal state of a neural net as an aid, not an obstacle. Many deep learning techniques are complex in nature and, although they result very accurate in many scenarios, they can become incredibly difficult to interpret. In contrast, TCAV results correctly show that the image concept was more important.TCAV is one of the most innovative approaches to neural network interpretability of the last few years. CAVs are learned by training a linear classifier to distinguish between the activations produced by a concept’s examples and examples in any layer.The second step is to generate a TCAV score that quantifies the sensitivity of the predictions to a specific concept. We show that these combinations can allow us to sort of “stand in the middle of a neural network” and see some of the decisions being made at that …

Prior to Google, Sara taught an open source machine learning for social good course in Nairobi, Kenya.