CompaniesPREMIUM

On the road to AI-led decision-making in SA

Technology set to be the recipient of greater investment

Picture: 123RF/KITTIPONG JIRASUKHANONT
Picture: 123RF/KITTIPONG JIRASUKHANONT

After helping to identify the Omicron variant of the deadly coronavirus — infamously first reported in SA — the country is set to see greater investment and use of artificial intelligence (AI) in enterprise decision-making. 

There is little doubt that the current crop of Covid-19 vaccines —  from Pfizer, Johnson & Johnson or Moderna — are some of the fastest developed in human history. A process that would typically take years to trial, let alone to complete, took a little more than a year, thanks in part to the use of AI tools and technology. 

The biggest advantage that AI systems have is their ability to work through mountains of data that would ordinarily take many man hours to complete, within a relatively short time. 

One company, InstaDeep, is looking to focus much of its recently raised funding to deepen its research & development activities in SA.

In January InstaDeep raised $100m (R1.457bn), closing a Series B funding round led by Luxembourg’s Alpha Intelligence Capital together with Hong Kong’s CDIB Capital. 

“InstaDeep will use the funding to advance its high-performance computing infrastructure optimised for decision-making AI, continue to hire elite talent — particularly in SA — and accelerate the launch of disruptive AI products across multiple industries, including biotech, logistics, transportation and electronics manufacturing,” the company says.

The firm is also expanding its global presence into the US.

InstaDeep, founded in 2014, specialises in using artificial intelligence to help companies and organisations with decision making on projects. Its AI is mainly used across the biology, logistics, electrical design and energy industries. The company is based in the UK with offices in France, Nigeria, Tunisia, Dubai and SA. 

“Through participation in events in the machine learning community, in particular the Deep Learning Indaba, we had the opportunity to actually come to SA — in Cape Town and Stellenbosch. This was in 2019 and we got the chance to meet very talented engineers and researchers in AI,” Karim Beguir, InstaDeep’s co-founder and CEO, told Business Day.

“And so it was a sort of natural thing for a company like ours which is really trying to showcase the talent that we have in Africa to have a presence in SA.”

The latest funding round also saw investment from German biotechnology firm BioNTech, Abu Dhabi investment firm Chimera Capital, railway operator Deutsche Bahn’s DB Digital Ventures, Google, investment firms G42 and Synergie.

InstaDeep has been working with one of these backers — BioNTech — for a number of years. The German firm is best known for its partnership with Pfizer on one of the world’s main Covid-19 vaccines. 

The Tunisia-founded start-up worked with BioNTech to develop an AI tool which detects high-risk Covid-19 variants based on their genetic code. The resultant algorithm identified more than 90% of variants of concern, including the Omicron strain, in advance.

While the world is still a long way from having drugs or treatments designed purely by AI, the technology has helped to optimise them. For example, Moderna’s vaccine platform learnt from producing 20,000 mRNA sequences, which helped it create the first batch of its Covid-19 vaccine for testing in 42 days.

Early on in the pandemic, AI was also used to help to narrow down the best available drugs to treat Covid-19. 

Experts say repurposing drugs is simply a “hack” and that the long-term hope for AI is that it can transform humanity’s understanding of disease. In such a world, AI systems will identify new diseases and find existing medication to treat them, all while working to develop new treatments or drugs specifically for those conditions. 

At the same time, the technology still has some way to go before it can be considered perfect, or at least mainstream. First, AI systems — though referred to as “intelligence” — still have a way to go before they can be considered to be doing actual “thinking”. For now, most are confined to a process more accurately referred to as “machine learning”, which, in practice, is a form of pattern recognition. 

Beguir believes decision-making AI to follow as the “second generation” of the technology.

“Second generation is about helping industrial players make complex decisions by saving time through the use of computers. For example: if you’re a railway company that has to schedule 10,000 rides a day over 30,000km of railway, that’s a complex decision.”

Outside biotech, InstaDeep’s team has also developed multi-agent reinforcement learning (MARL), an approach for solving large-scale problems by getting many AI devices to work together. An example is eliminating traffic congestion with self-driving cars. The approach has already had an impact on SA’s AI sector, with Cape Town “fast becoming an academic centre for MARL research”, the company says.

The company has also worked with Google to predict desert locust breeding grounds in advance to alert local farmers across Africa of potential threats from devastating outbreaks. 

gavazam@businesslive.co.za

Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.

Comment icon