Data scientist Vukosi Marivate has helped to build scientific communities and networks for African researchers in machine learning and artificial intelligence (AI).Credit: Mariki Uitenweerde, EYEscape Corporate Photography
During his PhD in the United States, computer scientist Vukosi Marivate discovered both the power of inclusive spaces and the sense of isolation created by non-diverse gatherings. Since then, Marivate, now a data scientist at the University of Pretoria in South Africa, has been building a critical mass of African data scientists — so that their voices can be part of the artificial intelligence (AI) revolution.
In 2016, Marivate and his fellow researchers established the Deep Learning Indaba conference (in Bantu languages, indaba means a consultation or an important discussion) to help to build a local community around machine learning and AI. The 2023 conference, held in Accra, Ghana, attracted 800 researchers from more than 62 countries, half of which were in Africa.
Trekking across Nepal to bridge its gender gap in the sciences
Marivate also co-founded Masakhane, a grass-roots community of more than 1,000 researchers, software developers and language specialists from 30 African countries, which aims to galvanize research in natural language processing in African languages, for and by Africans. Natural language processing uses machine-learning techniques to process text and speech, for purposes such as translation and transcription. Historically, Africa’s 2,000 languages have been neglected in digital products. One of Masakhane’s projects created bespoke scientific terms for African languages, and the community continues to generate African-led AI research groups and companies on the continent.
When did you realize that you wanted to tackle the lack of diversity in science?
During my PhD at Rutgers University in New Brunswick, New Jersey, I remember going to the International Conference on Machine Learning, one of the world’s top meetings on the subject. I walked into a room filled with people and on the other side of the giant hall was the only other Black person there. It messes with your mind and you start asking, “What’s happening here? Why are there only two of us?” You feel as if you do not belong.
That’s why I need to ensure that I open doors to others. It’s about access. How do we make sure that the people who are applying for international funding or fellowships better represent the African continent, and help them to realize that they can be excellent, like anybody else?
I also learnt that it doesn’t necessarily need to be me who helps them: it’s about using the network, asking who else is working in these subject areas — in the same country or on the continent — and then asking how I can connect students who are in those spaces. The Deep Learning Indaba and Masakhane are networks that are much, much bigger than any individual.
Why is diversity, equity and inclusion (DEI) work important?
‘Anyone engaging in scientific practice shouldn’t be excluded’: the United Kingdom’s only Black chemistry professor, on diversity
Once you start engaging with science in a societal way, you see that the people who are really pushing for innovation to come into their communities tend to be very different from those who are chasing the highest levels of prestige of science.
I think the best of us are still to come. Young people on the African continent are going to do so much more than us, older scientists. They have better networks; they have more global recognition of the work that they do. We’re going to have amazing work come from these young people and others who are not typically represented in these spaces.
What obstacles do Africans face in global AI and computing spaces?
The AI researcher archetype tends to be male, white and from the global north. We’re trying to say that your race, disability, gender and even your geographical area shouldn’t matter. We want people to feel comfortable, so that they can be themselves and bring their full selves into AI and computing spaces, without changing themselves to fit into the system.
The other barriers include funding and the pressures on students. For example, many South African students are likely to be breadwinners at home and are often pushed towards non-research industry jobs. How do we make sure that the ones that have that spark find places in the discipline and have options?
Who has been your biggest influence or mentor?
One is Tshilidzi Marwala, a South African AI engineer and rector of the United Nations University in Tokyo. Back in 2006, I was studying engineering at the University of the Witwatersrand (Wits) in Johannesburg, and wanted to work for a big engineering company and make money. Marwala said, “There are more interesting things out there in the world that you might want to look at.” His laboratory was at Wits at the time, and there I saw people who looked like me, who had backgrounds similar to mine. That experience changes you. He was also so generous with everyone, asking, “How do we open opportunities for you?” That was a major turning point in my scientific life.
