The past decade has witnessed a proliferation of data, where the notion of data science plays a central role in the envisioned many technological advances. A number of applications (like the online recommendation systems, the automatic driving systems, the financial Robo-Advisor systems, etc.) involving big data have already brought huge benefits to many facets of our daily lives. Advances in data analytics motivate a systematic way to uncover the hidden insights, mining the useful information, and carrying forward inferences from these massive datasets. The overarching objective of research in our group is to wed state-of-the-art high-performance methods with the emerging big data problems, in a way that they can inspire and reinforce the development of each other, with the ultimate goal of benefiting our societies. In pursuit of such goals, the research in our group can be summarized into the following topics.

Optimization, Learning, and Inference in Big Data Analytics


This topic aims at developing high-performance methods at the intersection of optimization, signal processing, statistics, and machine learning to address emerging challenges in large-scale data science applications. Especially, this topic will unveil the sparsity, and more generally, low-dimensional structures, in big data analytics. Structured signal processing has attracted much attention in many data-driven problems like sparse representation and inverse problems built upon low-dimensional modeling, including compressed sensing, matrix completion, robust principal component analysis, dictionary learning, super resolution, phase retrieval, neural networks, etc. This topic is aimed to focus on designing optimization-based algorithms that are effective in both theory and practice.

Artificial Intelligence for Financial Technology and Business Analytics


The application of research ideas from theoretical physics, mathematics, and control theory to the financial markets has been a common industrial practice for now almost three decades. Engineering has also witnessed a steady flow of contributions to the financial world from fields like computer science and optimization theory. Signal processing (SP), without exception, has benefited financial markets substantially through well-known and widely applied techniques as well, to name a few, the Fourier transform, the Kalman filter, and shrinkage methods. Recently years, artificial intelligence (AI) has been largely believed to be a promising technology in the renovation of financial engineering (FinEng), financial technology (FinTech), and financial informatics (FinInfo), where the booming robo-advisor system is a good example. In view of this, this topic is to explore the worlds of finance from perspectives of optimization, signal processing, and AI.

AI-Enabled Accelerated and Decentralized Analysis for Financial Data


Arguably, it’s never been more important for financial institutions to react quicker and more effectively – faced with the task of juggling important tasks such as detailed risk analysis, modelling, and simulation, there is no doubt that in modern era speed is king. Traditional financial workloads have huge data sets that need to be processed and pored through. An example workload is back-testing, which is a simulated trade on reliable historical data. It is a computationally intensive task because of the sheer volume of data. With standard data centre or cloud service provider infrastructure struggling to keep up with ever-increasing workloads, it’s clear that CPUs need to be supplemented with some serious silicon support. For these high-frequency trading and risk management applications, the software demands are time and resource intensive. As a result, the hardware must keep up with the computing demands and constantly changing industry parameters. Hardware options include CPUs, GPUs, ASICs, and FPGAs.