The programme works alongside dedicated Industry Partners towards training the next generation of students to understand, utilise and contribute to the forefront of ML and AI research. A brief summary could be found in this pdf, while a longer thought piece follows below.
Teaching for the Future
It is widely recognised that advances at the forefront of Artificial Intelligence (AI) and Machine Learning (ML) have the potential to disrupt industries. In light of this, numerous governments have hedged their bets and produced a wide variety of national AI strategies. Here in sub-Saharan Africa, Gartner forecasts a thirty-fold growth in the business value of AI over the next seven years, to $46.6 billion. With these estimates of growth and potential, how can we best position ourselves to benefit from and, more importantly, contribute to the fast-changing world of AI?
Amongst the many avenues of growth, this article considers one of the most fundamental roads: education.
Tech startups and companies leading the AI revolution are built on the foundation of a strong talent pool. However, the size of the current talent pool will not match the projected growth of AI in business, and the sector will face a growing skills shortage. A recent Tencent Research Institute report suggests that there are around 300,000 AI researchers and practitioners in the world, with a market demand of millions. Element AI, a leading Canadian lab, estimates that there are only 10,000 individuals worldwide with the right skillset to initiate serious AI projects. This shortfall is as prevalent in Africa as it is worldwide.
In Africa, there are many more grassroots initiatives, which make the present a very exciting time for AI in Africa. Across the continent, growing and flourishing grassroots movements can be seen, which are building and equipping communities of AI aficionados. The Deep Learning Indaba movement has inspired and stirred up the AI community across Africa; the main Indaba in Stellenbosch brought 550 researchers and practitioners from 35 African countries together in Stellenbosch last year. Its offshoots, the IndabaX events, are hosted in 27 African countries concurrently in 2019! Additionally, Data Science Africa plays a leading role in equipping people through workshops. Finally, the proliferation of cheap or free Massive Open Online Courses (MOOCs) provides accessible introductory entry points into AI and ML. Despite these exciting developments, we are not blind to the challenges of strengthening our human capital of AI researchers and practitioners. To have a clear voice at the forefront of AI research and development, we should pool our collective resources together even more.
In the university ecosystem, a pattern of specialised teaching is emerging in Europe, North America and China in the form of dedicated one-year taught Masters degrees in AI and ML. One such programme is being launched in the Tübingen AI cluster, which is a focal point of the German national AI strategy. The UK’s Cambridge-Oxford-London AI triangle has taught MSc degrees in AI at each of its corners. Due to the rapid and pervasive growth of AI as a field, leading universities such as Carnegie Mellon University in the US are already setting up undergraduate programmes that specialise in AI.
To our knowledge, the only dedicated AI Masters programme on the African continent is a very innovative African Masters of Machine Intelligence (AMMI) recently established at AIMS (the African Institute for Mathematical Sciences) in Kigali. AMMI is generously funded by Facebook and Google, and has an intake of around 30 students a year, with a new hub in Ghana in the pipeline. We need more formal programmes like these in Africa, to purposefully spearhead growth of the African AI talent pool!
What is next for southern Africa, given these investments internationally and further north? At Stellenbosch University, a coursework-based MSc in Machine Learning and Artificial Intelligence is actively being developed, with plans to have its first intake of students in January 2020. The blueprint of the degree is inspired by similar one-year Masters programmes like University College London’s Machine Learning MSc, the University of Cambridge’s MPhil in Machine Learning, Speech and Language Technology, and the University of Edinburgh’s Artificial Intelligence MSc.
AI today is fast-paced and diverse, and to ensure diversity and relevance as a flagship degree, the Stellenbosch MSc programme makes provision for researchers and academics from beyond our shores to teach certain advanced modules, or parts of modules. Moreover, starting a new programme is akin to founding and growing a startup, with the same estimation and belief of what the future might need. In the blueprint for the Stellenbosch MSc, we have chosen to include two modules that are somewhat novel to AI curriculums, “Applied Machine Learning at Scale” and “Artificial Intelligence and the Brain”. ML is finding its way into the terabyte and petabyte world of banking, online search and auctions, retail and media. In light of this, the “Applied Machine Learning at Scale” module will help students wear the “computational thinking” hat of scalability and distributed systems when designing ML solutions. The future research agenda for further advances in AI is increasingly looking to the neuroscience community for templates like short and long term memory formation and retrieval, one-shot and continual learning, attention mechanisms and many others. The “Artificial Intelligence and the Brain” module will lay the groundwork for students to contribute novel ideas at this exciting intersection of silicon and wet science.
Call to action, call for support
The Stellenbosch MSc programme is keen to work alongside dedicated industry partners towards training the next generation of students to understand, utilise and contribute to the forefront of AI research. Our industry partners supporting our mission will have a guiding voice on the programme’s advisory board. Partners could offer internships, suggest topics for students’ research projects, and even co-supervise these research projects if the interest is mutual.
We are making every effort to ensure that talented and hard-working students from across Africa will be able to be in the class of 2020. Yet, we cannot do so on our own. Scholarships are part of the investment to build tomorrow’s AI talent pool in Africa, and, in this regard, we will heavily rely on industrial partners to be the wind beneath the programme’s wings.
We invite you, dear reader, to join us in this exciting project for teaching ML and AI in the future. Together, we can do so much.
More information about industry partnership can be obtained from Dr Willie Brink at wbrink [at] sun [.] ac [.] za