As an active member and participant of the machine learning team at the University, David's expertise lies in applications, particularly in translating content from one language to another. “My work is focused on identifying and addressing current solutions and gaps that could be filled to mitigate challenges faced by millions of users while solving real-life problems,” explain David.
“On a daily basis we look at forecasting for the operations department and by analysing data and trends, assist in identifying customer volumes and channel preferences, enabling the team to optimise their focus and resources effectively. This practical experience has allowed me to experience the impact of machine learning in driving operational efficiency.”
Working with machine learning systems is not without its challenges. “Incorporating legacy organisations and businesses like Multichoice into a machine learning solution means working with stakeholders to provide new ways of doing things. There are obstacles to overcome in convincing stakeholders who are still opposed to the automated processes,” says David.
Working with various data is also a challenge, according to David, and learning different aspects that one might not have heard of in university which are needed in building a model that may require more interaction with current data to comprehend underlying concepts and implications. He also emphasises the relevance of ethical considerations and notes the need to train unbiased machine learning models while acknowledging the impact that bias in data can have on these models, and it is the responsibility of the person who built the model to ensure fairness and inclusivity.
David is interested in exploring other areas where machine learning may be used, such as generating new music and videos. He further explains that machine learning can generate visual content from scripts that are provided.
“Although the current models may not be perfect, I think that they will improve with time.”
He believes that anyone interested in machine learning should begin with their domain, and that having a comprehensive understanding of that industry can lead to more meaningful and relevant solutions. His love and passion for computers and artificial intelligence contribute to shaping a future where machine learning enhances various domains.
To learn more about the Multichoice Chair in Machine Learning, click here.
