Cisco Systems, Inc.
Automated identification of higher-order behaviors in a machine-learning network security system

Abstract:

In one example embodiment, a server compares derived values of a plurality of machine learning features with specified values of the plurality of machine learning features according to an ontology that defines a relationship between the plurality of machine learning features and a corresponding higher-order behavior for the specified values of the plurality of machine learning features. When the derived values of the plurality of machine learning features match the specified values of the plurality of machine learning features, the server aggregates the weights of the plurality of machine learning features to produce an aggregated weight. The server assigns the aggregated weight to the higher-order behavior so as to indicate a significance of the higher-order behavior in producing the machine learning result.