Visible to the public Biblio

Filters: Keyword is complex problems  [Clear All Filters]
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z   [Show ALL]
A
Pyatnisky, I. A., Sokolov, A. N..  2020.  Assessment of the Applicability of Autoencoders in the Problem of Detecting Anomalies in the Work of Industrial Control Systems.. 2020 Global Smart Industry Conference (GloSIC). :234—239.

Deep learning methods are increasingly becoming solutions to complex problems, including the search for anomalies. While fully-connected and convolutional neural networks have already found their application in classification problems, their applicability to the problem of detecting anomalies is limited. In this regard, it is proposed to use autoencoders, previously used only in problems of reducing the dimension and removing noise, as a method for detecting anomalies in the industrial control system. A new method based on autoencoders is proposed for detecting anomalies in the operation of industrial control systems (ICS). Several neural networks based on auto-encoders with different architectures were trained, and the effectiveness of each of them in the problem of detecting anomalies in the work of process control systems was evaluated. Auto-encoders can detect the most complex and non-linear dependencies in the data, and as a result, can show the best quality for detecting anomalies. In some cases, auto-encoders require fewer machine resources.

B
Radhika, K. R., Nalini, M. K..  2017.  Biometric Image Encryption Using DNA Sequences and Chaotic Systems. 2017 International Conference on Recent Advances in Electronics and Communication Technology (ICRAECT). :164–168.

Emerging communication technologies in distributed network systems require transfer of biometric digital images with high security. Network security is identified by the changes in system behavior which is either Dynamic or Deterministic. Performance computation is complex in dynamic system where cryptographic techniques are not highly suitable. Chaotic theory solves complex problems of nonlinear deterministic system. Several chaotic methods are combined to get hyper chaotic system for more security. Chaotic theory along with DNA sequence enhances security of biometric image encryption. Implementation proves the encrypted image is highly chaotic and resistant to various attacks.