Data science and cyber security systems: A statistical perspective

Data science is empowering many contemporary innovations, including the development of robust cybersecurity systems. It is becoming a significant research area that embedded the combination of modern computing tools and statistical inferences, with the demand of adequately maintaining data privacy within the systems. For example, in e-healthcare systems, IoT devices monitor the patient’s health and upload classified data as Electronic Medical Records (EMRs) to the cloud. Sharing EMRs data in hyper-connected systems with a limited storage capacity of IoT devices in a third party based service provider is more vulnerable to attack by sophisticated cybercriminals. Existing privacy protection schemes, including cryptography, cannot ensure the privacy demands in a cloud environment. Medical data sharing need multiple data accessing paradigms with different levels of privacy according to the need for consultation by a specialist, e.g., surgeons, general practitioners, allied health staff. Besides, none of these techniques is robust in privacy preservation because of not determining privacy weight (i.e. an orthogonal design matrix-based weight generation strategy for determining data privacy) for the e-Health industry using a cloud environment or automated by adaptation of AI. These effectively limit the full potential of the dynamic cloud-based e-healthcare systems and business environment and the wider uptake and comprehensive benefits of data science more generally.

In this inaugural talk, I will address some significant computing and inferential challenges in data science with a particular focus on cyber design space, security of data, computation, and inferences. It will not only highlight computing statistics for the decision-making process or running intelligent algorithms, but also there must be a focus on reliability measures on any applied outputs. This keynote presentation will also illustrate some real-world case studies from data science, including a quick overview of one of our project aims to develop a robust solution for privacy-preserving ehealth data sharing in a cloud environment with an AI-based automated technique to ensure data privacy and measuring reliability.