Deep Unfolding: Model-Based Inspiration of Novel Deep Architectures

TL;DR

This paper introduces a framework called deep unfolding that transforms model-based inference algorithms into trainable deep neural network architectures, combining domain knowledge with the flexibility of deep learning.

Contribution

It proposes a novel method to unfold inference algorithms into deep networks, enabling the integration of domain knowledge into neural architectures and demonstrating its effectiveness in speech enhancement.

Findings

Deep unfolding can interpret conventional networks as mean-field inference in Markov random fields.

New architectures can be created using belief propagation instead of mean-field inference.

The approach achieves competitive speech enhancement results with fewer parameters.

Abstract

Model-based methods and deep neural networks have both been tremendously successful paradigms in machine learning. In model-based methods, problem domain knowledge can be built into the constraints of the model, typically at the expense of difficulties during inference. In contrast, deterministic deep neural networks are constructed in such a way that inference is straightforward, but their architectures are generic and it is unclear how to incorporate knowledge. This work aims to obtain the advantages of both approaches. To do so, we start with a model-based approach and an associated inference algorithm, and \emph{unfold} the inference iterations as layers in a deep network. Rather than optimizing the original model, we \emph{untie} the model parameters across layers, in order to create a more powerful network. The resulting architecture can be trained discriminatively to perform accurate inference within a fixed network size. We show how this framework allows us to interpret conventional networks as mean-field inference in Markov random fields, and to obtain new architectures by instead using belief propagation as the inference algorithm. We then show its application to a non-negative matrix factorization model that incorporates the problem-domain knowledge that sound sources are additive. Deep unfolding of this model yields a new kind of non-negative deep neural network, that can be trained using a multiplicative backpropagation-style update algorithm. We present speech enhancement experiments showing that our approach is competitive with conventional neural networks despite using far fewer parameters.