Timezone: »

 
Poster
Flowification: Everything is a normalizing flow
Bálint Máté · Samuel Klein · Tobias Golling · François Fleuret

Wed Nov 30 09:00 AM -- 11:00 AM (PST) @ Hall J #917

We develop a method that can be used to calculate the likelihood contribution of linear and convolutional layers allowing multi-layer perceptrons and convolutional networks to be converted into normalizing flows. We term this process flowification. In some cases flowification requires the addition of uncorrelated noise to the model but in the simplest case no additional parameters. The technique we develop can be applied to a broad range of architectures, allowing them to be used for a wide range of tasks. Our models also allow existing density estimation techniques to be combined with high performance feature extractors. In contrast to standard density estimation techniques that require specific architectures and specialized knowledge, our approach can leverage design knowledge from different domains and is a step closer to the realization of general purpose architectures. We investigate the efficacy of linear and convolutional layers for the task of density estimation on standard datasets.

Author Information

Bálint Máté (University of Geneva)
Samuel Klein (University of Geneva, Switzerland)
Tobias Golling (University of Geneva)
François Fleuret (University of Geneva)

François Fleuret got a PhD in Mathematics from INRIA and the University of Paris VI in 2000, and an Habilitation degree in Mathematics from the University of Paris XIII in 2006. He is Full Professor in the department of Computer Science at the University of Geneva, and Adjunct Professor in the School of Engineering of the École Polytechnique Fédérale de Lausanne. He has published more than 80 papers in peer-reviewed international conferences and journals. He is Associate Editor of the IEEE Transactions on Pattern Analysis and Machine Intelligence, serves as Area Chair for NeurIPS, AAAI, and ICCV, and in the program committee of many top-tier international conferences in machine learning and computer vision. He was or is expert for multiple funding agencies. He is the inventor of several patents in the field of machine learning, and co-founder of Neural Concept SA, a company specializing in the development and commercialization of deep learning solutions for engineering design. His main research interest is machine learning, with a particular focus on computational aspects and sample efficiency.

More from the Same Authors