For more information please see our blog post, Investing in Exponential Times.
Although the average consumer may not be familiar with the concept of machine learning, it’s already a big part of our daily lives. Tom Schlebusch, co-founder and CEO of NMRQL Research, a local FinTech player offering South Africa’s first machine learning powered unit trust fund, says machine learning has a big role to play in how we invest our money. Amidst the explosion of data and the exponential advancement of technology, investors are beginning to realise the tremendous benefits of machine learning in investment management.
Schlebusch explains that when Amazon or Netflix makes recommendations which appeal to us, it’s because machine learning was used to analyse our behaviour. “All social media platforms use machine learning to provide personalised feedback based on behaviour patterns and relationships with other users. Machine learning enables Google maps to provide relevant traffic updates. These are just a few of the rapidly-growing number of areas where machine learning is adding value to our lives,” he says.
According to Schlebusch, machine learning is the study of algorithms that are able to learn how to solve a problem using data, usually lots of data. “One form of machine learning starts with inputs, desired outputs, and an algorithm whose behaviour can be tuned,” he says. “Over time, the algorithm basically ‘learns’ by tuning its parameters until it can produce the desired output. In essence, machine learning is when the machine (computer system) learns to make data-driven decisions rather than being explicitly programmed.”
Given the magnitude of data in investment decision-making, Schlebusch says much value can be derived from applying machine learning in the investment management process. “In this process, the inputs into machine learning models may include the news, fundamentals, technicals and economics and the desired outputs are future returns,” explains Schlebusch.
Unlike human-centric traditional investment decision-making which is underpinned by cognitive biases that negatively affect objectivity and critical reasoning, Schlebusch believes that machine learning algorithms drive unbiased, more efficient investment decisions. Furthermore, he says, in our rapidly-changing world and markets, machine learning algorithms are learning continuously from new data. “They evolve their ‘belief system’ as markets change, thereby enhancing the efficacy of investment decision-making.”
Machine learning algorithms can also process tremendous quantities of data in any format. This is particularly relevant given the exponential advancement of technology and the explosion of data. Schlebusch says, “By some estimates, 50% of all data ever created was created in the past 2 years. Furthermore, an incredible 70 – 80% of that data is unstructured, which means it exists as websites, social media, voice recording, images, video, or even sattelite photographs of oil carousels.”
According to Schlebusch, traditional investment techniques that rely on humans processing of the relevant data, inevitably fail to exploit unstructured data because of its sheer size and lack of inherent shape or form. However, according to Schlebusch, machine-learning algorithms can make sense of unstructured data and use it as inputs to optimise the investment decision-making process. He adds that machine learning is currently the only technology available that can process unstructured data on any significant scale.
While consumers have embraced the benefits of machine learning in certain areas of their lives, many remain cautious when it comes to investing. Schlebusch says, “People are generally resistant to change. It’s not just fear of change but our inbuilt cognitive bias which suggests that when you’ve been doing something a particular way for a long time, it must be a good way to do it. This mindset does not consider how the tremendous technology-driven change taking place around us is creating new and improved ways of doing things.”
While this general inertia may delay some investors from unlocking value from machine-based investing processes, Schlebusch believes innovators and early adopters will lead the way, embracing this new way of investing and influencing others to follow, as the benefits become evident.
“The investment management industry is on the cusp of a new era,” concludes Schlebusch, “as it moves towards unbiased, scalable, adaptive and testable algorithmic investment decision-making, which is also more cost-effective. As this shift gathers momentum, more and more investors will unlock the value of machine learning when investing.”