Timezone: »

 
Poster
Invert to Learn to Invert
Patrick Putzky · Max Welling

Thu Dec 12 10:45 AM -- 12:45 PM (PST) @ East Exhibition Hall B + C #161
in Deep Learning -- Efficient Training Methods »

Iterative learning to infer approaches have become popular solvers for inverse problems. However, their memory requirements during training grow linearly with model depth, limiting in practice model expressiveness. In this work, we propose an iterative inverse model with constant memory that relies on invertible networks to avoid storing intermediate activations. As a result, the proposed approach allows us to train models with 400 layers on 3D volumes in an MRI image reconstruction task. In experiments on a public data set, we demonstrate that these deeper, and thus more expressive, networks perform state-of-the-art image reconstruction.

Keywords: Efficient Training Methods · Deep Learning · Algorithms -> Large Scale Learning; Algorithms -> Sparsity and Compressed Sensing; Algorithms · Structured Prediction; Applica
[ Paper]  [ Slides] 

Author Information

Patrick Putzky (University of Amsterdam)
Max Welling (University of Amsterdam / Qualcomm AI Research)

More from the Same Authors