No abstract

Signed networks are mathematical structures that encode positive and negative relations between entities such as friend/foe or trust/distrust. Recently, several papers studied the construction of useful low-dimensional representations (embeddings) of these networks for the prediction of missing relations or signs. Existing network embedding methods for sign prediction, however, generally enforce different notions of status or balance theories in their optimization function. These theories, are often inaccurate or incomplete which negatively impacts method performance. In this context, we introduce conditional signed network embedding (CSNE). Our novel probabilistic approach models structural information about the signs in the network separately from finegrained detail. Structural information is represented in the form of a prior, while the embedding itself is used for capturing fine-grained information. These components are then integrated in a rigorous manner. CSNE's accuracy depends on the existence of sufficiently powerful structural priors for modelling signed networks, currently unavailable in the literature. Thus, as a second main contribution, which we find to be highly valuable in its own right, we also introduce a novel approach to construct priors based on the Maximum Entropy (MaxEnt) principle. These priors can model the polarity of nodes (the degree to which their links are positive) as well as signed triangle counts (a measure for the degree structural balance holds to in a network). Experiments on a variety of real-world networks confirm that CSNE outperforms the state-of-the-art on the task of sign prediction. Moreover, the MaxEnt priors on their own, while less accurate than full CSNE, achieve accuracies competitive with the state-of-the-art at very limited computational cost, thus providing an excellent runtime-accuracy trade-off in resource-constrained situations. CCS CONCEPTS • Computing methodologies → Machine learning; Knowledge representation and reasoning.

Botnets have been recently recognized as one of the most formidable threats on the Internet. Different approaches have been designed to detect these types of attacks. However, as botnets evolve their behavior to mislead the signaturebased detection systems, learning-based methods may be deployed to provide a generalization capacity in identifying unknown botnets. Developing an adaptable botnet detection system, which incrementally evolves with the incoming flow stream, remains as a challenge. In this paper, a self-learning botnet detection system is proposed, which uses an adaptable classification model. The system uses an ensemble classifier and, in order to enhance its generalization capacity, updates its model continuously on receiving new unlabeled traffic flows. The system is evaluated with a comprehensive data set, which contains a wide variety of botnets. The experiments demonstrate that the proposed system can successfully adapt in a dynamic environment where new botnet types are observed during the system operation. We also compare the system performance with other methods.

No abstract

No abstract