Towards Deterministic Diverse Subset Sampling
We discuss a greedy deterministic adaptation of a k-DPP. Deterministic algorithms are interesting for many applications, as they provide interpretability to the user by having no failure probability and always returning the same results. First, the ability of the method to yield low-rank approximations of kernel matrices is evaluated by comparing the accuracy of the Nyström approximation on multiple datasets. Afterwards, we demonstrate the method on an image search task.
Joachim Schreurs was born in Hasselt Belgium, October 13, 1994. In 2017 he received a Master’s degree in Mathematical Engineering at the KU Leuven. He is currently a doctoral student in machine learning, at the STADIUS research division of the Department of Electrical Engineering (ESAT) at KU Leuven, under the supervision of prof. Johan A. K. Suykens.