easchaw

Deep-Neural-Networks

Interpretable-Autoencoder-based-Neural-Networks

Deep Neural Networks achieve the state-of-the-art performance, this is not enough. They lack the ability to bring intuitiveness to the end user in terms of explaining the output of the model. The novel architectures proposed in this research will address vulnerabilities in the domain and hence will provide insights for the end users.

Intrusion-Detection-System-

Cyber-security is concerned with protecting information, a vital asset in today’s world. The volume of data that is generated and can be usefully analysed is such that cyber-security can only be effectively implemented with the aid of software support. Data must be analysed by software tools providing support for security analysts. Often event data generated by computer systems is sequential, that is, not only are the type of the events relevant, but the sequence in which events occur is also relevant. Examples of this include many log files and system call or software library call sequences. This research aims to provide the basis to build an Anomaly Detection based Host Intrusion Detection System (HIDS) that makes decisions based on sequential traces of operating system calls.

Multi-variate-Anomaly-Detection

To investigate the analysis of multivariant data in order to identify anomalous data instances. Root cause analysis benefits from this by filtering features whose values lead to such anomalies.

Publications

Springer/ IEEE ComSoc peer-reviewed list of publications

easchaw/Intrusion-Detection-System-

Cyber-security is concerned with protecting information, a vital asset in today’s world. The volume of data that is generated and can be usefully analysed is such that cyber-security can only be effectively implemented with the aid of software support. Data must be analysed by software tools providing support for security analysts. Often event data generated by computer systems is sequential, that is, not only are the type of the events relevant, but the sequence in which events occur is also relevant. Examples of this include many log files and system call or software library call sequences. This research aims to provide the basis to build an Anomaly Detection based Host Intrusion Detection System (HIDS) that makes decisions based on sequential traces of operating system calls.

easchaw/Deep-Neural-Networks

easchaw/Interpretable-Autoencoder-based-Neural-Networks

Deep Neural Networks achieve the state-of-the-art performance, this is not enough. They lack the ability to bring intuitiveness to the end user in terms of explaining the output of the model. The novel architectures proposed in this research will address vulnerabilities in the domain and hence will provide insights for the end users.

easchaw/Multi-variate-Anomaly-Detection

To investigate the analysis of multivariant data in order to identify anomalous data instances. Root cause analysis benefits from this by filtering features whose values lead to such anomalies.

easchaw/Publications

Springer/ IEEE ComSoc peer-reviewed list of publications