We consider datasets consisting of arbitrarily structured en- tities (e.g., molecules, sequences, graphs, etc) whose sim- ilarity can be assessed with a reproducing kernel (or a family thereof). These entities are assumed to additionally have a set of named attributes (e.g.: number_of_atoms, stock_price, etc). These attributes can be used to classify the structured entities in discrete sets (e.g., ‘number_of_atoms < 3’, ‘stock_price ≤ 100’, etc) and can effectively serve as Boolean predicates. Our goal is to use this side-information to provide named kernel-based anomaly detection. To this end, we propose a method which is able to find among all possible entity subsets that can be described as a conjunction of the available predicates either a) the optimal cluster within the Reproducing Kernel Hilbert Space, or b) the most anomalous subset within the same space. Our method employs combinatorial optimisation of an adaptation of the Maximum-Mean-Discrepancy measure that captures the above intuition. Additionally, we propose a cri- terion to select the optimal one out of a family of kernels in a way that preserves the available side-information. Finally, we provide several real world datasets that demonstrate the usefulness of our proposed method.
AAAI Conference on Artificial Intelligence