Quantum Computing and Machine Learning

QSPectral is interested in the potential of leveraging Quantum properties to provide tractable cyber solutions. The advent of a family of new algorithms  based on solving a system of linear equations have shown significant speedups over classical algorithms for eminently-practical problems in Machine Learning such as clustering, classification, and finding patterns in huge amounts of data.

At the interface between quantum computing and machine learning, the field of quantum machine learning aims to improve classical machine learning algorithms with the help of quantum computers. It has been shown that data pertaining to some real-world problems can be represented by Graph Convolutional Neural Networks. This opens up, for example, the potential of quantum kernel methods in the convolutional layer for use as a pre-processing layer in the final neural network.

The aims of this research are as follows:

Investigate the application of Quantum Hopfield Nets and Quantum Kernels for in realistic problem domains.

Develop a Quantum Deep Neural Network model based on investigations  to enable accurate computationally competitive processing of networks.

Data Analytics, Modeling and  Simulation for Tele-health Solutions 

We are also researching and developing advanced data analytics on the MonitorMyHealth platform to process the disparate data from the devices to provide insights, outcomes and alerts. The algorithm for detecting a person’s activity based on just smartphone sensors is available to monitor activities undertaken in the home and is more accurate than anything that has been reported. The system can be employed as is, however, benefits come from blending with other data sources (i.e. from other sensors) to create a complete systems picture of a patient’s health.

We are investigating the application of streaming and learning algorithms to discover patterns in sensor data and provide predictive power and timely alerts for patients and healthcare providers.

Dynamic Wireless Networks
In remote monitoring, emergency and defence operations, real time communication leading to accurate positioning, tracking and measurement are essential, especially when these tasks support life threatening situations, command and control, intelligence, and sensor-linked systems.

This research program will investigate ad-hoc Software Defined Networking (SDN) for:

• providing a scheme for the use of communication networks in challenging situations such as bush fire monitoring. The optimal placement of limited resources to create reliable and robust networks would be critical to accurate diagnosis and saving life.
• addressing a significant gap in existing defence, emergency and automated mining capabilities by analysing requirements and identifying optimal technology outcomes through modelling and simulation that will enable provision of appropriate responses in a timely and cost-efficient manner.