How Artificial Intelligence Is Transforming The Future Of Investing In Africa

By Tom Jackson AFKI Original Published: October 23, 2017, 1:13 pm
Artificial Intelligence Is Transforming The Future Of Investing In Africa. Photo - CNBCArtificial Intelligence Is Transforming The Future Of Investing In Africa. Photo - CNBC

Making investments, in any asset class, is risky business, and requires heaps of research and bravery. This is especially so in the unpredictable African context, where artificial intelligence could have the solution to issues such as these.

In South Africa, a startup has launched what it calls the country’s first “machine learning-powered unit trust”, which uses artificial intelligence to improve stock selection.

NMRQL Research – which is co-founded by former First National Bank CEO Michael Jordaan – plans to invest in a diversified portfolio of domestic and international assets, using machine-learning algorithms to make investment decisions. The process allows the company to discover hidden patterns in big data, which can then be exploited to forecast returns.

Stuart Reid, chief engineer and partner at NMQRL, said the use of machine learning at NMRQL Research is innovative in any context, not simply an African one.

“Progress rarely happens linearly, it comes in sprints. Historically there have been three major machine learning “booms”. The first boom introduced the ideas of search, symbolic logic, and natural language processing and occurred from 1956 to 1974. The second boom introduced the ideas of expert systems, fuzzy logic, and support vector machines, and occurred from 1980 to 1987. We are currently living in the third boom which has introduced hundreds of new ideas,” he said.

“There are three major differences between the previous two booms and this one, namely, the availability of open source software, the availability of large publicly curated datasets, and the availability of cloud computing services like Amazon Web Services and Google Compute Engine.”

Open source software, open source datasets, and on-demand cloud computing services, coupled with an “open and honest” research community accessible via Twitter, Reddit, and Medium, is according to Reid a recipe which results in state-of-the-art machine learning algorithms being available to anybody, anywhere, at very low cost.

“This is particularly good news for the African continent, which has possibly not been in as good of a position to benefit from previous booms as this one,” he said.

The use of artificial intelligence in making investment decisions, he said, will change the industry in two major ways.

“Firstly, machine learning algorithms are able to process massive quantities of unstructured data intelligently. Unstructured data sources include sound, image, video, and more. These forms of data are not being traded on at the moment and therefore have a lot of potential to produce abnormal returns. There is no way that any fund manager using traditional investment models or even human analysts could process the same amount of data as machine learning algorithms can,” said Reid.

“Secondly, machine learning algorithms are free of the various cognitive biases which plague human investors and cause them to make bad decisions, which are bad for your wealth. These cognitive biases have been studied for decades by academics in the field of behavioral finance.”

Machine learning offers massive potential to the African continent in fields other than investments too, given the fact it is now possible for anybody with an internet connection to get access to the best machine learning frameworks, open source datasets, and data centres in the world.

“What that means is that even individuals with rudimentary, self-taught programming skills could design, develop, and deploy machine learning algorithms to solve theirs and other people’s problems,” said Reid.

“That having been said, in order for machine learning to make a much bigger impact on the continent we believe that society needs to place more of an emphasis on data collection. Data collected from individuals on a day to day basis, schools, shops, banks and clinics could be used by machine learning algorithms to solve complex problems relating to the optimal allocation of resources such as textbooks, medicine and water, as well as predicting their supply and demand.”

Tom Jackson is the co-founder of tech news and research platform Disrupt Africa and a journalist covering innovation on the continent from the Cape to Cairo.

Sign up for the AFKInsider newsletter — the most compelling business news you need to know from Africa and the African diaspora, delivered straight to your inbox.

Tags: , , , ,