Regularization for Multiple Kernel Learning via Sum-Product Networks
This work addresses the problem of improving regularization in multiple kernel learning for machine learning practitioners, but it appears incremental as it builds on existing SPN and MKL frameworks without indicating a paradigm shift.
The paper tackles the problem of constructing graph-based regularizers for multiple kernel learning by representing kernel combination structures with sum-product networks, proposing a convex regularization method that encodes the SPN structure via a path-dependent kernel weighting function. The result includes an analysis of convexity and classifier complexity, along with an efficient wrapper algorithm for optimization, though no concrete numbers are provided in the abstract.
In this paper, we are interested in constructing general graph-based regularizers for multiple kernel learning (MKL) given a structure which is used to describe the way of combining basis kernels. Such structures are represented by sum-product networks (SPNs) in our method. Accordingly we propose a new convex regularization method for MLK based on a path-dependent kernel weighting function which encodes the entire SPN structure in our method. Under certain conditions and from the view of probability, this function can be considered to follow multinomial distributions over the weights associated with product nodes in SPNs. We also analyze the convexity of our regularizer and the complexity of our induced classifiers, and further propose an efficient wrapper algorithm to optimize our formulation. In our experiments, we apply our method to ......