Gitta Kutyniok: "An Information Theoretic Approach to Validate Deep Learning-Based Algorithms" скачать в хорошем качестве

Gitta Kutyniok: "An Information Theoretic Approach to Validate Deep Learning-Based Algorithms"

Machine Learning for Physics and the Physics of Learning 2019 Workshop III: Validation and Guarantees in Learning Physical Models: from Patterns to Governing Equations to Laws of Nature "An Information Theoretic Approach to Validate Deep Learning-Based Algorithms" Gitta Kutyniok - Technische Universität Berlin, Program in Applied and Computational Mathematics Abstract: In this talk, we provide a theoretical framework for interpreting neural network decisions by formalizing the problem in a rate-distortion framework. The solver of the associated optimization, which we coin Rate-Distortion Explanation (RDE), is then accessible to a mathematical analysis. We will discuss theoretical results as well as present numerical experiments showing that our algorithmic approach outperforms established methods, in particular, for sparse explanations of neural network decisions. Institute for Pure and Applied Mathematics, UCLA October 30, 2019 For more information: http://www.ipam.ucla.edu/mlpws3