Credit: Google News
In this profile series, we interview AI innovators on the front-lines – those who have dedicated their life’s work to improving the human condition through technology advancements.
Meet Muthoni Wanyoike and Jade Abbott, two women at the forefront of Africa’s growing AI ecosystem. Abbott is a senior software developer and machine learning engineer at South African software company Retro Rabbit, and Wanyoike leads a team at InstaDeep — an AI startup with offices in Tunisia, Nigeria, Kenya, England and France.
Both Abbott and Wanyoike are also involved in efforts to build a community of developers across the continent.
Over the past few years, several conferences focused on AI and emerging technology have sprung up across Africa. Wanyoike is an organizing committee member for the largest of these events, Deep Learning Indaba, and the co-organizer of the Nairobi Women in Machine Learning and Data Science community. Abbott, a co-organizer of Women in Tech ZA, had a paper on neural machine translation for African languages accepted in 2018 into a workshop at the global AI conference NeurIPS.
In this interview, Wanyoike and Abbott share their perspectives on the AI community gaining traction in Africa, and discuss their vision for developing diverse tech talent in the region.
Why did you get involved in fostering the machine learning community in Africa?
Abbott: In Africa, even coming from a very privileged background, I still don’t feel like I’m on the global playing field. So imagine people who aren’t from that privileged background, which is the majority of the continent. What I care about is bringing people together who are interested in machine learning — inspiring people, showing them that they can do it, and they’re not excluded, and they can be just as good as any of the international players. They can contribute, not just within their own communities, but also to the global scientific community.
Wanyoike: My focus is to provide learning opportunities for women who are interested in getting into AI. For me, my vision and why I continue to work in AI is because I see AI as an enabler, and as an opportunity for us to solve African problems, to solve our day-to-day challenges, be it health care, congestion, or planning.
What’s your vision for AI across Africa going forward?
Wanyoike: Currently in Kenya, it’s a really young space. My country’s just 50 years old, so we have the opportunity to start from a very interesting place, because we’re starting from a point of information, a point where we make sense of the data that we collect. I think this is such a great opportunity for us. It’s an opportunity for us to define what Africa looks like in the next 100 years. So my vision is that we will use data that we are collecting to make sense of our continent, our life, and our spaces, and also build products people care about and solutions that can actually tackle the challenges that we face.
Abbott: And even beyond that, the way I see it is Africa becomes a hub for innovation, not only within our problems, but also on a fundamental, theoretical level. I don’t see why foundational research from a very low-level, technical, machine learning theory environment can’t come out of Africa. We can apply our unique perspective of looking at problems, whether they be ones that are directly applicable to our environment, or ones that might be applicable to the world.
Africa, in being younger, is more agile and will be able to innovate very fast compared to some of the international players. For example, the major banks in Kenya are way younger than banks in America and Europe. And with that, they don’t carry a legacy load that they have to take with them, which does allow them to move significantly faster. Everywhere else caught on to mobile money five or ten years after M-Pesa already had permeated the entirety of Kenya and East Africa.
What are some applications of AI in Africa you are most proud of?
Abbott: There are honestly so many, it’s hard to pick. If you asked me two years ago what applications I was most proud of, I would not have known who was applying AI in Africa. That’s the beauty of events like the Deep Learning Indaba, which bring together people doing such amazing work. The visibility of AI in Africa across the continent has been very siloed up until recently.
Some favorites are Dr. Justine Nasejje’s work in survival analysis, which seeks to uplift our continent by addressing its pressing public health challenges. Raesetje Sefala’s work in using satellite images and computer vision to study the evolution and effects of spatial apartheid in South Africa. Or more in the governmental space, the City of Cape Town, using data science and machine learning for forecasting to aid problems in their water crisis. And Praekelt Foundation’s MomConnect, who’ve improved their ability to connect and advise pregnant women across Africa by building chatbots that can work with multiple African languages.
How have you seen the AI ecosystem in Africa change over time?
Wanyoike: I think in Kenya, a lot has changed in the last one year. So for example, with Women in Machine Learning and Data Science, we moved from maybe 500 members to now almost 2,000 — in just one year. We also have many more new AI communities. There’s AI Kenya and Nairobi.AI. We have groups that are doing amazing, incredible work, in terms of sharing knowledge, of inviting new sources or speakers, and just making information available to their networks. So we’re seeing a lot of change, much of which can be attributed to the community initiatives around the continent.
What impact do these conferences and community events held on the continent have for women interested in AI?
Wanyoike: There are several challenges facing women in machine learning and data science. First is their education. There are a lot of field gaps. And aside from field gaps, there are a lot of gaps in terms of information and networking. Who do you connect with if you’re trying to solve a language processing problem, for example? The community provides a space for people to not only learn, but to also collaborate and work together. We’ve seen a lot of projects springing up from just people meeting from the meetups, and access to job opportunities that they wouldn’t have had before.
Abbott: It’s very multi-layered. But the first barrier, which is a very hard one to change at this stage until we reach scale, is there’s a very much a traditional perception. Let’s say if you come from a rural background or traditional background, which in South Africa is definitely the case for a majority of the population, if you are female and you are going to go into business, your parents are going to pay for you to become a dentist, an accountant, a doctor or a lawyer. You aren’t actually too exposed to technology at a very young age.
What we’ve found is that you can challenge that at multiple levels. Running programming days or STEM days at schools works really well. When you go in there with women representation, the students see that, ‘Hey look, she does that, I could do that, too.’ I often say the reason that I ended up in technology was because my mother was a programmer. It was always something I could see women doing. So I think they’re grassroots issues.
It’s empowering to have mentors and people you can talk to who are similar to you, or look like you. It seems like we shouldn’t have to have a role model who is also female, but it is that way. We do. It shows us that we can do it, too.
Credit: Google News