• Comment classification for an online news domain 

      Brand, Dirk; Van der Merwe, Brink (2014-12)
      ENGLISH ABSTRACT: In online discussion forums, comment moderation systems are often faced with the problem of establishing the value of an unseen online comment. By knowing the value of comments, the system is empowered ...
    • Firegaze : processing and visualizing firewall logs in the cloud 

      Van Tonder, R.; Visser, W. (SATNAC, 2013)
      This project aims to visualise packet counts filtered by iptables at the network layer, and allows for performing network forensics in a distributed environment. For example, anomalies such as bandwidth spikes and port ...
    • Flow allocation in wireless networks with selfish nodes 

      Krzesinski, A. E. (SATNAC, 2013)
      Consider an ad hoc network where packet transmissions occur between the nodes. Optimal flow allocation in such systems can be modelled as a constrained nonlinear optimisation problem. This problem can be solved either ...
    • Learning dynamics of linear denoising autoencoders 

      Pretorius, Arnu; Kroon, Steve; Kamper, Herman (PMLR, 2018)
      Denoising autoencoders (DAEs) have proven useful for unsupervised representation learning, but a thorough theoretical understanding is still lacking of how the input noise influences learning. Here we develop theory ...
    • N-gram representations for comment filtering 

      Brand, Dirk; Kroon, Steve; Van der Merwe, Brink; Cleophas, Loek (ACM, Inc., 2015-09)
      Accurate classifiers for short texts are valuable assets in many applications. Especially in online communities, where users contribute to content in the form of posts and comments, an effective way of automatically ...
    • Unsupervised pre-training for fully convolutional neural networks 

      Wiehman, Stiaan; Kroon, Steve; De Villiers, Hendrik (Institute of Electrical and Electronics Engineers, 2016)
      Unsupervised pre-training of neural networks has been shown to act as a regularization technique, improving performance and reducing model variance. Recently, fully con-volutional networks (FCNs) have shown state-of-the-art ...