The Human Kernel
This work addresses the challenge of making machine learning models more human-like in their reasoning for cognitive science and AI applications, though it is incremental in improving kernel design.
The authors tackled the problem of automating human expertise in function extrapolation by reverse engineering human inductive biases through a kernel learning framework, resulting in kernels that enable human-like extrapolation beyond traditional methods.
Bayesian nonparametric models, such as Gaussian processes, provide a compelling framework for automatic statistical modelling: these models have a high degree of flexibility, and automatically calibrated complexity. However, automating human expertise remains elusive; for example, Gaussian processes with standard kernels struggle on function extrapolation problems that are trivial for human learners. In this paper, we create function extrapolation problems and acquire human responses, and then design a kernel learning framework to reverse engineer the inductive biases of human learners across a set of behavioral experiments. We use the learned kernels to gain psychological insights and to extrapolate in human-like ways that go beyond traditional stationary and polynomial kernels. Finally, we investigate Occam's razor in human and Gaussian process based function learning.