Apr 14, 2021
We present eigenvalue estimates of integral operators associated with compositional dot-product kernels. The estimates improve on previous ones established for power series kernels on spheres. Our estimates allow us to obtain the volumes of balls in the corresponding reproducing kernel Hilbert spaces. We then discuss the consequences in statistical estimation with compositional dot-product kernels and highlight interesting trade-offs between the approximation error and the statistical error depending on the length of the compositions and the smoothness of kernels.
The 24th International Conference on Artificial Intelligence and Statistics was held virtually from Tuesday, 13 April 2021 to Thursday, 15 April 2021.
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Presentations on similar topic, category or speaker