Timezone: »

 
Poster
Simple and Effective VAE Training with Calibrated Decoders
Oleh Rybkin · Kostas Daniilidis · Sergey Levine

Thu Jul 22 09:00 PM -- 11:00 PM (PDT) @
in Poster Session 6 »

Variational autoencoders (VAEs) provide an effective and simple method for modeling complex distributions. However, training VAEs often requires considerable hyperparameter tuning to determine the optimal amount of information retained by the latent variable. We study the impact of calibrated decoders, which learn the uncertainty of the decoding distribution and can determine this amount of information automatically, on the VAE performance. While many methods for learning calibrated decoders have been proposed, many of the recent papers that employ VAEs rely on heuristic hyperparameters and ad-hoc modifications instead. We perform the first comprehensive comparative analysis of calibrated decoder and provide recommendations for simple and effective VAE training. Our analysis covers a range of datasets and several single-image and sequential VAE models. We further propose a simple but novel modification to the commonly used Gaussian decoder, which computes the prediction variance analytically. We observe empirically that using heuristic modifications is not necessary with our method.

Keywords: Bayesian Nonparametrics · Algorithms -> Missing Data; Algorithms -> Uncertainty Estimation; Probabilistic Methods · Causal Inference; Probabilistic Meth · Embedding and Representation learning

Author Information

Oleh Rybkin (University of Pennsylvania)

Oleg is a Ph.D. student in the GRASP laboratory at the University of Pennsylvania advised by Kostas Daniilidis. He received his Bachelor's degree from Czech Technical University in Prague. He is interested in deep learning and computer vision, and, more specifically, on using deep predictive models to discover semantic structure in video as well as applications of these models for planning. Prior to his Ph.D. studies, he worked on camera geometry as an undergraduate researcher advised by Tomas Pajdla. He was a visiting student researcher at INRIA advised by Josef Sivic, Tokyo Institute of Technology advised by Akihiko Torii, and UC Berkeley advised by Sergey Levine.

Kostas Daniilidis (University of Pennsylvania)
Sergey Levine (UC Berkeley)
Sergey Levine

Sergey Levine received a BS and MS in Computer Science from Stanford University in 2009, and a Ph.D. in Computer Science from Stanford University in 2014. He joined the faculty of the Department of Electrical Engineering and Computer Sciences at UC Berkeley in fall 2016. His work focuses on machine learning for decision making and control, with an emphasis on deep learning and reinforcement learning algorithms. Applications of his work include autonomous robots and vehicles, as well as computer vision and graphics. His research includes developing algorithms for end-to-end training of deep neural network policies that combine perception and control, scalable algorithms for inverse reinforcement learning, deep reinforcement learning algorithms, and more.

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors