‘I’ve been inspired by the generation of mathematicians and theoretical computer scientists that laid the mathematical foundations of modern cryptography. They contributed an incredible amount of deep conceptual insights – such as the many different ways to mathematically formalise the notion of a proof, or what it means to gain knowledge or to learn. This conceptual innovation is something that inspires me, and to which I aspire in my own work,’ states Dr Ryan Sweke, the Alexander von Humboldt German Research Chair of Mathematics, and its applications at the African Institute for the Mathematical Sciences (AIMS). He leads the Quantum at AIMS group, which conducts research under the broad umbrella of quantum computing and quantum information, and covers a wide variety of topics. He is also a senior lecturer at Stellenbosch University.
International Year of Quantum Science and Technology
‘I firmly believe that our current understanding of quantum mechanics, and the window into the fundamental nature of our world that it provides, is an incredible collective achievement that belongs to everyone, and that everyone should be able to gain joy, excitement, inspiration, understanding and wonder from. Therefore, I’m very happy about the year 2025 being proclaimed the International Year of Quantum Science and Technology (IYQ),’ Ryan says.
‘The IYQ provides a co-ordinated opportunity to convey this excitement and wonder to the broader public, and in particular to those who either are excluded, or feel excluded, from the scientific community. In addition, this opportunity is particularly timely, given the unprecedented opportunities that quantum computers provide both for conveying the ideas and applications of quantum mechanics, and for participation in its development! In this regard, organisations like NITheCS have a massive role to play as facilitators of scientific communication initiatives, and of concrete opportunities for students and researchers who are inspired and want to develop their careers.’
He continues: ‘Quantum computing sits at the intersection of theoretical computer science, physics, mathematics and engineering, and certainly couldn’t have been developed to where it is today without significant interdisciplinary collaboration between all those fields. To some extent, one can think of quantum computing as a testament to the power of interdisciplinary work.’
Ryan believes inter-cultural collaboration to be equally important. ‘For the sustainability of our planet, and for developing just societies, it’s crucial that there is no dominant culture of science in which some are able to feel comfortable and invited, and others are excluded. I believe that research done together by diverse groups of individuals, with diverse perspectives, backgrounds and ways of thinking, is almost certain to be more successful, more relevant to the problems we face as a society, and more enjoyable! This is also a reason why I have always been particularly drawn to AIMS, since it provides an incredibly diverse pan-African environment for research.’
Background
Ryan Sweke completed his PhD at the University of KwaZulu-Natal in 2017, with a focus on quantum algorithms for the simulation of open quantum systems. He then spent almost five years (2018-2022) as a post-doctoral researcher at the Freie Universität Berlin, first as an Alexander von Humboldt postdoctoral fellow (2018-2019) and then as a senior postdoc of the PlanQK project, the flagship quantum machine learning initiative of the German government.
‘At the time I was lucky enough to be part of a really large research group led by Prof. Jens Eisert at the Freie Universität Berlin, with an incredibly supportive and collaborative research culture. There were no silos, nor was there internal competition; everyone was welcome to collaborate with anyone else and to pursue the directions that were interesting to them. This culture really allowed me to grow as a scientist, to gain a much broader picture and understanding of the interesting questions in and around quantum computing, and to work with a wide range of collaborators. There is joy in consistent collaboration over many years, because eventually one gets to do science with one’s friends! I feel very lucky to have been able to experience that.’ After Berlin, Ryan spent just over two years (2022-2024) as research scientist at IBM Quantum in San Jose, California, before joining AIMS in January 2025.
Quantum focus areas
His current main interests relate to today’s ‘hot topics’ in science, technology and general discussions. ‘Firstly, one could ask simply “Can quantum computers be useful for machine learning (ML)?” Over the last decade, advances in ML have had a huge impact across all sectors of society – including the natural sciences – and given the fact that quantum computers offer a different way of computing that we know sometimes has advantages, it’s natural to ask whether or not they can offer any advantages over standard computers for ML problems.’
He continues: ‘There are many ways in which one could imagine answering this question. Ideally, one would want to try benchmark quantum ML proposals against state-of-the-art classical algorithms, such as large language models. Unfortunately, current generation quantum computers don’t allow for such experiments: they are too small and too noisy. So, my work is focused on trying to use theoretical analyses of proposed quantum ML algorithms to gain evidence for or against their usefulness for different ML tasks. Unfortunately, most of the evidence accumulated so far points to quantum computers not being advantageous for the standard ML problems that we are used to – like natural language processing, image generation etc. However, the jury is still out. Ultimately, we will probably have to wait until we have large-scale quantum computers before we can get any definitive answers.’
Ryan also cites ‘many genuinely quantum problems, for which the data that you have, or want to generate, are quantum states. For these problems, “standard” ML algorithms can’t be directly applied, and so quantum ML algorithms need to be developed.’
In terms of security and trustworthy machine learning, Ryan says there are many places in the modern ML pipeline where different parties have to trust each other. ‘For example, when a client pays a cloud computing provider to train a ML model, they have to trust that the provider will actually use the resources they say they will. Another example: sometimes a data provider might want to keep their data secret from someone who wants to use it to train a ML algorithm. Without a specific solution they will simply have to trust that the other party will not steal their data.’
He names another example: ‘When you ask a human to do a task for them, you have to trust that the human has not simply used a ML algorithm to do the task. There are many more such examples! Luckily, however, a very well-developed branch of theoretical computer science deals with how to develop systems in a way that removes the need for different parties to trust each other – cryptography! I’m currently very interested in the application of cryptographic techniques and ideas in the development of secure and trustworthy ML pipelines. This is particularly interesting, because we already know that quantum cryptography offers advantages over standard cryptography, and this may also turn out to be the case for applications within ML settings.’
Ryan is also interested in whether state-of-the-art ML tools can be used to solve difficult problems in the development of quantum computers themselves. ‘Examples include the use of ML for the development of quantum compilers, quantum error correction codes, etc.’
Turning to the burning questions in the field, Ryan says ‘the most consistent question is “When will quantum computers provide meaningful advantages for problems that are actually relevant in scientific or industrial contexts?”. Unfortunately, given the hype around (and investment into) quantum computing, it’s very hard to extract an answer to this question from all of the public-facing information about quantum computing. My own personal opinion on this is that we simply don’t know, and that it depends strongly on the area of application. For some problems, such as the simulation of quantum mechanical systems – which is relevant for material or drug design – we know of good quantum algorithms that could outperform the best known classical algorithms, and it’s “just” a matter of waiting for large-scale quantum computers that could be made robust enough to withstand the noise. For other problems, such as ML, we do not even really have many established candidate quantum algorithms.’
He adds: ‘There are certainly many significant and technical challenges to be overcome if we are to build large scale quantum computers. We have made incredible progress over the last decade – definitely more progress than I think anyone would have predicted – and I believe that there are no physical obstacles preventing the construction of large-scale quantum computers eventually. However, the timeline of when we will have large-scale fault-tolerant quantum computers is still unclear to me.’
The scientific process matters
‘I feel extremely privileged and lucky to be able to spend my time solving problems together with collaborators. Essentially, almost all of the time of a scientist is spent being confused to some extent, slowly, slowly, developing the insight and intuition – with the help of collaborators and students and mentors – that sometimes, finally, provides a moment of understanding. This sense of mystery and uncertainty, and the process of slowly stripping this away in a collective effort, is something that I find exciting and deeply satisfying,’ Ryan says.
‘I truly love being at the blackboard with collaborators, working together to understand something that we didn’t understand when we started. Of course, it’s essential to pick problems of which the solutions will be meaningful, but at least for me personally doing science would be totally unsustainable without really enjoying the process of science – the ups and the downs! At least, given the challenges and setbacks involved in research, I don’t think I would have the perseverance if my only motivation was the solution, as opposed to the process itself.’
Finally, Ryan says it is extremely important that young researchers actively seek out scientific environments that are healthy, diverse, supportive and enjoyable. ‘At the end of the day, while science is certainly challenging, it should feel like joy. Unfortunately, some research cultures are competitive, intimidating or isolating – and often very successful researchers are too busy to commit sufficient time to mentoring, collaboration or supervision.’
He advises that young researchers should be up-front and open about asking potential colleagues, collaborators and mentors about mutual expectations and support. ‘Concrete questions like “How many times can we expect to meet in person?”, “How long will it take for you to reply to my emails?”, “How much of my time do you expect me to commit to this project – and how much of your time do you expect to commit to it”, “What are your expectations of me on a weekly, monthly, yearly basis?”, “Do you expect me to be in the lab every day?” do matter – and I think it’s important for researchers to ensure that expectations are matched before taking up any potential opportunity. Notably, it is critical for junior or young researchers to feel supported in practical ways.’