Just like learning to ride a bike, humans have the remarkable ability to learn complex functional relationships between variables from only a small amount of data. But how exactly do we do it? Well, it turns out that we have an internal toolbox filled with prior expectations about the forms these relationships may take. In three fascinating experiments, scientists have discovered that these expectations are not set in stone, but rather, they can be adjusted and refined through experience. It’s like having a flexible bike that can adapt to different terrains! This groundbreaking research builds upon previous studies that used fancy statistical modeling techniques (called Gaussian processes) to explain how humans learn functions. By applying a hierarchical Bayesian inference approach to the Gaussian process, researchers were able to capture the dynamic nature of learning to learn functions. The findings suggest that our brains have a sophisticated system for understanding and predicting complex relationships between variables. To dive deeper into the details of this captivating research, check out the full article!

Abstract
Humans can learn complex functional relationships between variables from small amounts of data. In doing so, they draw on prior expectations about the form of these relationships. In three experiments, we show that people learn to adjust these expectations through experience, learning about the likely forms of the functions they will encounter. Previous work has used Gaussian processes—a statistical framework that extends Bayesian nonparametric approaches to regression—to model human function learning. We build on this work, modeling the process of learning to learn functions as a form of hierarchical Bayesian inference about the Gaussian process hyperparameters.

Read Full Article (External Site)