Bayes in the age of intelligent machines

TL;DR

This paper discusses how Bayesian models and neural networks are complementary approaches for understanding human cognition and intelligent machines, emphasizing their different levels of analysis and mutual benefits.

Contribution

It clarifies that Bayesian models and neural networks operate at different analysis levels and can be combined to better understand cognition and AI systems.

Findings

Bayesian models and neural networks are complementary.

Bayesian approaches help interpret large neural networks.

Different analysis levels enhance understanding of cognition and AI.

Abstract

The success of methods based on artificial neural networks in creating intelligent machines seems like it might pose a challenge to explanations of human cognition in terms of Bayesian inference. We argue that this is not the case, and that in fact these systems offer new opportunities for Bayesian modeling. Specifically, we argue that Bayesian models of cognition and artificial neural networks lie at different levels of analysis and are complementary modeling approaches, together offering a way to understand human cognition that spans these levels. We also argue that the same perspective can be applied to intelligent machines, where a Bayesian approach may be uniquely valuable in understanding the behavior of large, opaque artificial neural networks that are trained on proprietary data.