Deep equilibrium models as estimators for continuous latent variables

TL;DR

This paper demonstrates that deep equilibrium models can serve as estimators for continuous latent variables, establishing a connection between neural network architectures and statistical models, and providing a foundation for unsupervised learning.

Contribution

It introduces a systematic framework linking deep equilibrium models to MAP estimation of latent variables, bridging neural networks with statistical modeling principles.

Findings

DEQ models solve MAP estimates for latents and parameters.

Activation functions and dropout relate to statistical assumptions.

DEQ feature maps are end-to-end differentiable for downstream tasks.

Abstract

Principal Component Analysis (PCA) and its exponential family extensions have three components: observations, latents and parameters of a linear transformation. We consider a generalised setting where the canonical parameters of the exponential family are a nonlinear transformation of the latents. We show explicit relationships between particular neural network architectures and the corresponding statistical models. We find that deep equilibrium models -- a recently introduced class of implicit neural networks -- solve maximum a-posteriori (MAP) estimates for the latents and parameters of the transformation. Our analysis provides a systematic way to relate activation functions, dropout, and layer structure, to statistical assumptions about the observations, thus providing foundational principles for unsupervised DEQs. For hierarchical latents, individual neurons can be interpreted as nodes in a deep graphical model. Our DEQ feature maps are end-to-end differentiable, enabling fine-tuning for downstream tasks.