Timezone: »
Poster
Matrix Inference and Estimation in Multi-Layer Models
Parthe Pandit · Mojtaba Sahraee Ardakan · Sundeep Rangan · Philip Schniter · Alyson Fletcher
We consider the problem of estimating the input and hidden variables of a stochastic multi-layer neural network from an observation of the output. The hidden variables in each layer are represented as matrices with statistical interactions along both rows as well as columns. This problem applies to matrix imputation, signal recovery via deep generative prior models, multi-task and mixed regression, and learning certain classes of two-layer neural networks. We extend a recently-developed algorithm -- Multi-Layer Vector Approximate Message Passing (ML-VAMP), for this matrix-valued inference problem. It is shown that the performance of the proposed Multi-Layer Matrix VAMP (ML-Mat-VAMP) algorithm can be exactly predicted in a certain random large-system limit, where the dimensions $N\times d$ of the unknown quantities grow as $N\rightarrow\infty$ with $d$ fixed. In the two-layer neural-network learning problem, this scaling corresponds to the case where the number of input features, as well as training samples, grow to infinity but the number of hidden nodes stays fixed. The analysis enables a precise prediction of the parameter and test error of the learning.
Author Information
Parthe Pandit (University of California, Los Angeles)
Parthe is a Postdoc at the Halicioglu Data Science Institute at UCSD. He is interested in high-dimensional optimization problems in machine learning and signal processing.
Mojtaba Sahraee Ardakan (UCLA)
Sundeep Rangan (NYU)
Philip Schniter (The Ohio State University)
Alyson Fletcher (UCLA)
More from the Same Authors
-
2023 Poster: A Regularized Conditional GAN for Posterior Sampling in Image Recovery Problems »
Matthew Bendel · Rizwan Ahmad · Philip Schniter -
2022 Poster: Instability and Local Minima in GAN Training with Kernel Discriminators »
Evan Becker · Parthe Pandit · Sundeep Rangan · Alyson Fletcher -
2019 Poster: Input-Output Equivalence of Unitary and Contractive RNNs »
Melikasadat Emami · Mojtaba Sahraee Ardakan · Sundeep Rangan · Alyson Fletcher -
2018 Poster: Plug-in Estimation in High-Dimensional Linear Inverse Problems: A Rigorous Analysis »
Alyson Fletcher · Parthe Pandit · Sundeep Rangan · Subrata Sarkar · Philip Schniter -
2017 Poster: Rigorous Dynamics and Consistent Estimation in Arbitrarily Conditioned Linear Systems »
Alyson Fletcher · Mojtaba Sahraee-Ardakan · Sundeep Rangan · Philip Schniter -
2014 Poster: Scalable Inference for Neuronal Connectivity from Calcium Imaging »
Alyson Fletcher · Sundeep Rangan -
2014 Spotlight: Scalable Inference for Neuronal Connectivity from Calcium Imaging »
Alyson Fletcher · Sundeep Rangan -
2012 Poster: Approximate Message Passing with Consistent Parameter Estimation and Applications to Sparse Learning »
Ulugbek S Kamilov · Sundeep Rangan · Alyson Fletcher · MIchael Unser -
2011 Poster: Neural Reconstruction with Approximate Message Passing (NeuRAMP) »
Alyson Fletcher · Sundeep Rangan · Lav R Varshney · Aniruddha Bhargava