Structure Discovery in Nonparametric Regression through Compositional Kernel Search
This addresses the challenge of kernel selection in machine learning for researchers and practitioners, offering a systematic approach to improve prediction accuracy and interpretability.
The paper tackles the problem of selecting kernel structures in nonparametric regression by proposing a compositional search method, which outperforms widely used kernels and combination methods on various prediction tasks, enabling interpretable decompositions and long-range extrapolation.
Despite its importance, choosing the structural form of the kernel in nonparametric regression remains a black art. We define a space of kernel structures which are built compositionally by adding and multiplying a small number of base kernels. We present a method for searching over this space of structures which mirrors the scientific discovery process. The learned structures can often decompose functions into interpretable components and enable long-range extrapolation on time-series datasets. Our structure search method outperforms many widely used kernels and kernel combination methods on a variety of prediction tasks.