Despite the successful application of machine learning (ML) in a wide range of domains, adaptabilitythe very property that makes machine learning desirable-can be exploited by adversaries to contaminate training and evade classification. In this paper, we investigate the feasibility of applying a specific class of machine learning algorithms, namely, reinforcement learning (RL) algorithms, for autonomous cyber defence in software-defined networking (SDN). In particular, we focus on how an RL agent reacts towards different forms of causative attacks that poison its training process, including indiscriminate and targeted, white-box and black-box attacks. In addition, we also study the impact of the attack timing, and explore potential countermeasures such as adversarial training.
Spatial Information Systems and their recent temporal extensions typically store large volumes of geo-referenced information. Having such size, it becomes increasingly difficult to explore their contents with current querying techniques. In this paper, we examine how data mining methods can help users in die analysis of the contents of Spatial and Spatio-Temporal Information Systems. We review existing spatial applications and investigate how they can be extended to deal with time. We also look at new, alternative methods that utilise the inherent structure of spatio-temporal information as well as its rich semantics to derive rules about changes and movement dictated by the querying requirements of the users of Spatio-Temporal Information Systems. The identification of these requirements thus remains important in helping to accommodate data mining techniques within STIS. Our current research concentrates on the development of generic knowledge discovery tools to be used on spatiotemporal data. By incorporating these tools into various types of STIS, we are aiming to provide users with enhanced analytical and data management capabilities. To this effect, we shall define a framework for mining evolution and meta-rules, and test their use in a prototype spatio-temporal knowledge discovery system.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.