Interpolation and Learning with Scale Dependent Kernels
This work addresses theoretical understanding of interpolation in machine learning, providing insights into error regimes for researchers, but it is incremental as it builds on existing kernel methods.
The paper investigates the learning properties of nonparametric ridge-less least squares with scale-dependent kernels, showing that the scale controls stability and that learning error decreases when sample size is less than exponential in data dimension, while variance due to noise remains bounded at larger sample sizes.
We study the learning properties of nonparametric ridge-less least squares. In particular, we consider the common case of estimators defined by scale dependent kernels, and focus on the role of the scale. These estimators interpolate the data and the scale can be shown to control their stability through the condition number. Our analysis shows that are different regimes depending on the interplay between the sample size, its dimensions, and the smoothness of the problem. Indeed, when the sample size is less than exponential in the data dimension, then the scale can be chosen so that the learning error decreases. As the sample size becomes larger, the overall error stop decreasing but interestingly the scale can be chosen in such a way that the variance due to noise remains bounded. Our analysis combines, probabilistic results with a number of analytic techniques from interpolation theory.