Africa’s contribution to the world of artificial intelligence is in the spotlight this week with the announcement by born-in-Africa AI disruptor InstaDeep that it has raised $100 million. It also welcomed high-profile backers including BioNTech, Deutsche Bahn and others – to help scale its decision-making AI products that solve real-world problems.
The startup made international headlines this month after it co-developed an AI-powered early warning system for high-risk Covid variants with BioNTech. It was founded in 2014 in North Africa. Today, it is a leader in decision-making AI products, with a headquarters in London and AI research and engineering teams in Tunis, Paris, Lagos, Dubai, and Cape Town.
“When we started InstaDeep, forecasts for AI’s contribution to global economic growth neglected to even mention Africa,” says Karim Beguir, InstaDeep’s co-founder and CEO. “For me, this was a call to arms. We needed to change the status quo. It could not be that Africa was excluded when it came to shaping the future of AI.”
The status quo has changed. InstaDeep announced 25 January that it has raised $100 million and closed a Series B round led by Alpha Intelligence Capital together with CDIB. Investors in the round included BioNTech, Chimera Abu Dhabi, Deutsche Bahn’s DB Digital Ventures, Google, G42, and Synergie.
The company will use the funding to advance its high-performance computing infrastructure optimised for decision-making AI, continue to hire elite talent – particularly in South Africa. This will accelerate the launch of disruptive AI products across multiple industries, including biotech, logistics, transportation, and electronics manufacturing. The firm is also expanding its global presence into the United States.
Detecting new high-risk Covid variants
The global interest in the BioNTech-InstaDeep Covid Early Warning System showed the appetite for AI thinking from beyond Silicon Valley. It worked with BioNTech, the German biotech company behind the Pfizer Covid-19 vaccine, to develop an AI tool which detects high-risk SARS-CoV-2 variants based on their genetic code.
The AI system identified more than 90% of variants of concern, on average two months before their designation by the World Health Organisation. It detected the highly transmissible Omicron on the day its sequence became available among more than 70,000 novel variants discovered in October and November 2021.
“More than 10,000 novel variant sequences are discovered every week and human experts simply cannot cope with complex data at this scale,” Beguir says. “For the first time, high-risk variants could be detected on the spot, potentially saving months of precious time.”
African AI expertise to address African problems
The company’s South African team is busy working on new ways to advance AI so it can help solve challenges in areas such as health, transportation, and resource management.
A particular interest for the Cape Town-based team is “Multi-Agent Reinforcement Learning” (MARL), which is an approach for solving large-scale problems by getting many AI devices to work together, for example, eliminating traffic congestion with self-driving cars. Its focus on the subject had an impact on South Africa’s AI ecosystem with Cape Town fast becoming an academic centre for MARL research.
Developing and working with local AI communities is also central to the company’s ethos. To empower other AI researchers, its South African team also created an open-source MARL framework, called Mava – named after the Xhosa word for ‘experience’. It has the aim of accelerating solutions to problems across the continent.
“This is a powerful technology given the number of important and highly impactful problems that it could help us solve,” says Arnu Pretorius, the AI research scientist who leads InstaDeep’s South Africa team. “Some of our work in this direction looked at managing scarce resources using MARL to potentially overcome issues such as the economic problem of balancing community and self-interest when managing common-pool resources such as water, arable land or the atmosphere.
“This work was presented at the most prestigious AI conference in the world, the conference on Neural Information Processing Systems, or NeurIPS – only the second time that research from Africa was presented there.”
InstaDeep’s pan-African team also collaborated with Google to predict desert locust breeding grounds in advance to alert local farmers across Africa of potential threats from devastating locust outbreaks.
Empowering South Africa’s AI community
Pretorius says: “We have research engineers not only focusing on research but also working on applied projects in collaboration with other offices, clients and the research teams. These projects range from biotechnology to hardware design using AI.
“In South Africa’s AI landscape, InstaDeep occupies a very niche space, blending engineering and research, while having a positive impact on the world. We also see ourselves as playing an important role in the larger AI community when it comes to strengthening connections, upskilling and providing opportunities for learning.
“Examples of these efforts include our close collaboration with the Deep Learning Indaba, an organization dedicated to building Africa’s AI community, where we sit on the steering committee as well as participate as sponsors, volunteers and mentors. We have a passion for being of service.”