Efficient Online Relative Comparison Kernel Learning
This addresses a scalability issue for applications like collaborative filtering and search, but it is incremental as it builds on existing kernel learning approaches.
The paper tackles the scalability problem of learning kernel matrices from relative comparison feedback by proposing ERKLE, an efficient online method that uses stochastic gradient descent and passive-aggressive updates, achieving improved speed and comparable or better accuracy compared to existing methods.
Learning a kernel matrix from relative comparison human feedback is an important problem with applications in collaborative filtering, object retrieval, and search. For learning a kernel over a large number of objects, existing methods face significant scalability issues inhibiting the application of these methods to settings where a kernel is learned in an online and timely fashion. In this paper we propose a novel framework called Efficient online Relative comparison Kernel LEarning (ERKLE), for efficiently learning the similarity of a large set of objects in an online manner. We learn a kernel from relative comparisons via stochastic gradient descent, one query response at a time, by taking advantage of the sparse and low-rank properties of the gradient to efficiently restrict the kernel to lie in the space of positive semidefinite matrices. In addition, we derive a passive-aggressive online update for minimally satisfying new relative comparisons as to not disrupt the influence of previously obtained comparisons. Experimentally, we demonstrate a considerable improvement in speed while obtaining improved or comparable accuracy compared to current methods in the online learning setting.