Technical Program

Paper Detail

Paper Title Asymptotics of MAP Inference in Deep Networks
Paper IdentifierTU2.R1.2
Authors Parthe Pandit, Mojtaba Sahraee, University of California, Los Angeles, United States; Sundeep Rangan, New York University, United States; Alyson K. Fletcher, University of California, Los Angeles, United States
Session Jack Keil Wolf Award
Location Le Théatre (Parterre), Level -1
Session Time Tuesday, 09 July, 11:40 - 13:00
Presentation Time Tuesday, 09 July, 12:00 - 12:20
Manuscript  Click here to download the manuscript
Abstract Deep generative priors are a powerful tool for reconstruction problems with complex data such as images and text. Inverse problems using such models require solving an inference problem of estimating the input and hidden units of the multi-layer network from its output. Maximum a priori (MAP) estimation is a widely-used inference method as it is straightforward to implement, and has been successful in practice. However, rigorous analysis of MAP inference in multi-layer networks is difficult. This work considers a recently-developed method, multi-layer vector approximate message passing (ML-VAMP), to study MAP inference in deep networks. It is shown that the mean squared error of the ML-VAMP estimate can be exactly and rigorously characterized in a certain high-dimensional random limit. The proposed method thus provides a tractable method for MAP inference with exact performance guarantees.