Abstract

 In this paper, we propose a general smoothing framework for graph kernels by taking structural similarity into account, and apply it to derive smoothed variants of popular graph kernels. Our framework is inspired by state-of-the-art smoothing techniques used in natural language processing (NLP). However, unlike NLP applications that primarily deal with strings, we show how one can apply smoothing to a richer class of inter-dependent sub-structures that naturally arise in graphs. Moreover, we discuss extensions of the Pitman-Yor process that can be adapted to smooth structured objects, thereby leading to novel graph kernels. Our kernels are able to tackle the diagonal dominance problem while respecting the structural similarity between features. Experimental evaluation shows that not only our kernels achieve statistically significant improvements over the unsmoothed variants, but also outperform several other graph kernels in the literature. Our kernels are competitive in terms of runtime, and offer a viable option for practitioners.