Quantum AI Research Group
Embedded Files
Research
Investigation on a systematic approach to determine the optimal structure of a quantum neural network that is applied in the forecasting of financial time series data.
The Travelling Salesman problem is NP-Hard and has important industry applications. We find good solutions for networks with up to 12 locations with Qiskit simulations of quantum circuit model devices using a penalty-free formulation.
We utilise a real-world cryptocurrency dataset to train a quantum generative adversarial networks (qGAN), and implement the real-life NVIDIA stock price data on quantum LSTM (QLSTM), comparing its performance with classical Long Short Term Memory (LSTM) model.
Benchmark evaluations of QIGA on six well-known multi-modal optimisation functions. These tests validate QIGA’s robustness and generalizability in exploring complex and deceptive search spaces to consistently locate global optima.
Awarded/Flagged Projects
Eureka, 2026-2028: "Vulnerability-based Smart Prevention, Defense and Mitigation using Generative AI for Cyber Security", Total Budget: €5.3M, KU: €182K, Consortium: 12 partners from Portugal, UK, Spain, Turkiye, UK Academic Lead/PI at KU: Xing Liang, CoI at KU: JC Nebel.
Ongoing Projects
Seedcorn Funding Grant, 2025: AI-Infused Cloud Security: Enabling Sustainable and Zero Trust Paradigms through Automation, £10K, PI: Xing Liang
Royal Academy of Engineering, 2024-2027: Harnessing AI Power for Stronger and Resilient Cyber Defences, £30K, PI: Xing Liang
Completed Projects
KU First Grant, 2024: Delivering Zero Trust with AI-driven Security Automation, £10K, PI: Xing Liang
KU First Grant, 2024: Improved British Sign Language (BSL) Recognition Using Deep Learning, £10K, CoI: Xing Liang
KU WISH, 2024: Dataset Development for Data-Driven Predictive Modelling of Innovative Building Façades, £10K, CoI: Xing Liang
Acknowledgements
Google Sites
Report abuse