No abstract

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.

No abstract

No abstract

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.

The fact that deep neural networks are susceptible to crafted perturbations severely impacts the use of deep learning in certain domains of application. Among many developed defense models against such attacks, adversarial training emerges as the most successful method that consistently resists a wide range of attacks. In this work, based on an observation from a previous study that the representations of a clean data example and its adversarial examples become more divergent in higher layers of a deep neural net, we propose the Adversary Divergence Reduction Network which enforces local/global compactness and the clustering assumption over an intermediate layer of a deep neural network. We conduct comprehensive experiments to understand the isolating behavior of each component (i.e., local/global compactness and the clustering assumption) and compare our proposed model with state-of-the-art adversarial training methods. The experimental results demonstrate that augmenting adversarial training with our proposed components can further improve the robustness of the network, leading to higher unperturbed and adversarial predictive performances.

No abstract

No abstract