Machine-learning research earns UCT academic NRF P-rating

Dr Jan Buys, a senior lecturer at the University of Cape Town (UCT), specialising in natural language processing (NLP) and machine-learning (ML), has been awarded a prestigious P-rating by the National Research Foundation (NRF). This recognition is a testament to his outstanding contributions to the field of computer science and the global recognition of his research.

The NRF rating system is a cornerstone of the foundation’s mission to establish a globally competitive science system in South Africa. It serves as a vital tool for evaluating the quality of researchers in comparison to their international counterparts. NRF ratings are bestowed upon researchers based on the excellence and impact of their recent research outputs. This system encourages researchers to consistently produce high-quality work, publish in influential journals and attend top international conferences.

The NRF P-rating is reserved for individuals who have attained the highest level of recognition in their respective fields. It signifies that a researcher is not only a leader but also a pioneer in their area of expertise. Moreover, this rating is a strong indicator of a researcher’s potential to shape the future of their field and mentor the next generation of scholars.

“UCT is exceptionally proud of Dr Buys’ contributions to the exciting fields of natural language processing and machine-learning.”

UCT’s acting deputy vice-chancellor for Research and Internationalisation, Professor Jeff Murugan, lauded the achievement. “UCT is exceptionally proud of Dr Buys’s contributions to the exciting fields of natural language processing and machine-learning. His achievement highlights the university’s commitment to excellence and innovation and further underscores UCT’s reputation as a hub of cutting-edge, world-class research.”

Trailblazer in natural language processing

Researchers, normally younger than 35, who have held a doctorate or equivalent qualification for less than five years at the time of application are eligible for consideration. These individuals must demonstrate exceptional potential through their published doctoral work and research outputs in their early post-doctoral careers. They are recognised as having the potential to become future international leaders in their field based on exceptional research performance and output from their doctoral and early post-doctoral research careers.

Dr Buys, with his extensive research background and expertise in NLP and ML, meets all these requirements. His journey in academia has been marked by dedication, innovation and a commitment to advancing the frontiers of knowledge.

Buys’ research primarily focuses on text generation, linguistic structure prediction and low-resource language processing. The most prominent examples of applications in this research field are large language models (LLMs) such as ChatGPT, Microsoft’s Bing Copilot and Google’s Bard, which have now been widely adopted due to their ability to generate fluent text in response to user prompts.

“This opens many opportunities for impactful applications that were hard to imagine until recently.”

His research outputs have not only been well received within the academic community but also represent foundational research related to the technology that underlies many of these systems.

His pioneering work in NLP has opened new possibilities for text generation and linguistic analysis. “The speed of advances in the abilities of text generation systems such as ChatGPT in the past few years has surprised many experts in the field. This opens many opportunities for impactful applications that were hard to imagine until recently,” Buys said. “This rating is a validation of progress in this direction. But at the same time, many challenges and open research questions remain.”

A journey of excellence and innovation

His journey towards becoming a prominent researcher in the field of NLP is both inspiring and illustrative of the potential young researchers hold in shaping the future of science. Buys’ academic path has been marked by milestones that reflect his exceptional potential and dedication to his craft.

He completed his undergraduate and master’s degrees in computer science at Stellenbosch University. Motivated to further his learning, he pursued a doctoral degree at the University of Oxford, where he was under the expert guidance of Professor Phil Blunsom, an early pioneer in research on language models. This was the foundation for his future contributions to the field.

Following the completion of his doctorate, he served as a postdoctoral researcher at the University of Washington, collaborating with renowned researcher Professor Yejin Choi in one of the world’s leading NLP research groups.

“This experience expanded my horizons and allowed me to delve deeper into the intricacies of approaches to text generation from language models.”

“This experience expanded my horizons and allowed me to delve deeper into the intricacies of approaches to text generation from language models. The research pushed the boundaries of what was possible at a time of rapid advances in the field, just before interest in LLMs exploded with the release of models such as ChatGPT,” he said.

After joining UCT at the end of 2019, Buys shifted his research focus from earlier foundational work to the next challenge – developing models for languages not yet benefiting from these advances.

“One of the biggest current challenges is that LLMs work well for English and other high-resource languages for which large amounts of text is already available on the web, but not as well for low-resource languages, which includes most African and South African languages,” he said.

Buys’ research on advancing NLP for South African languages is a representation of the importance that researchers based in Africa should be active participants in these developments. “I am encouraged by the increase in local and grassroots-driven research into NLP for African languages, including an increase in papers from researchers based in Africa published at top international NLP conferences,” he said.

Catalyst for future breakthroughs

In many ways, Buys represents the future of NLP. His pioneering spirit sets a precedent for the next generation of researchers in the field. His NRF P-rating not only honours his past achievements but also serves as a catalyst for future breakthroughs.

He reflected, “Receiving the NRF P-rating is a humbling recognition of the path I’ve chosen, and it motivates me to continue pushing the boundaries of research that can expand the positive impacts of artificial intelligence systems that can understand and generate human language. I see it as a call to mentor and inspire the next generation of researchers who will shape the future of this field.”