How Machine Learning is Changing the Landscape for Investments

By Benjamin Richard Truitt, CFA, Spire Fund Advisory, LLC

 

What is Artificial Intelligence & Machine Learning?

Did you know the first mechanical calculator was created in the 17th century by Wilhelm Schickard and Blaise Pascal? This invention was the first concept that changed the human experience with relation to what is known as artificial intelligence. Artificial Intelligence (AI) is the simulation where computer systems exhibit the traits that would normally require the human mind such as learning and problem-solving. However, another part of artificial intelligence is machine learning, which is a combination of mathematical modeling and computer science.

Machine Learning (ML)is effectively changing the way we use financial technology. As technology continues to advance, large financial firms and asset managers are more aware of the opportunities that ML and AI innovations provide. ML is changing the investment industry and the way investments are done. We do not need human capital for as much of the work today, yet we need the workforce around ML to adopt how to use technology and interact with it. It has begun to drive success and will continue to do so going forward.

The ML Algorithm Shows Parallel to the Crypto Market

To some extent, there are parallels between advancement in the application of Machine Learning and what is happening with Bitcoin mining. In Bitcoin mining a very complex mathematical problem is solved by a computer. As more and more Bitcoins are mined, the math problem becomes increasingly complex requiring an increasingly sophisticated and expensive computer system. The more powerful the computer, the faster the Bitcoin mining algorithm and the more likely it is a miner will be successful in obtaining a Bitcoin. Likewise, as ML advances and the number of market participants utilizing ML increases, the competition for obtaining an edge requires increasingly sophisticated ML algorithms and computers. A key difference between the application of ML and mining for Bitcoins is that much of the ML algorithm design is still done my humans and its performance is significantly dependent on the skill or artistry of the designer, which will likely present an edge between one investment company over another for quite some time.

ML for Investing = No Humans? Or more opportunities?

Machine Learning for investing can look at a vast number of comparable data points and arrive at an analysis in a way that is impossible for a human to effectively replicate. The more analysis needed, the more refined the terms, the more accurate the predictions and outcome. While humans have and can perform accurate due diligence, there is a ceiling on production and output.

What Machine Learning can do is sift through large amounts of data in a reduced timeframe, something no human – even a Stephen Hawking – can replicate. As an example, AI can go through 14,000 companies over a few hours. It would take a great deal of manpower to pour through that level of data and identify correlations between signals in the data and performance. While this can be unsettling to analysts who understand the market nuances, the movements of industries they cover and are whizzes at Excel, they simply cannot compete with this level of automation. However, this automation only got to be this effective through human understanding and the ability to fine tune the ML.

There is a learning curve with technology that can pivot and grow by learning how to work with the technology. The portion of the inputs will, over time, need less human capital to function but more human capital to operate and interact with the systems. It is a natural evolution and not much different than the adoption of motorized vehicles. They may be smart in many ways, yet they require humans to design and build them and they still need – for the most part – humans to operate them.

The Art of the Data Processing Making Better Predictions & Detecting Fraud

A large part of ML is the art of data science. For example, a fund manager, starting up in 1997, would have never been able to keep staff lean and rely on technology for 99% of the analysis. However, today that has changed, and it is due to the evolution of technology in investments. Machine Learning can extract the benefits of a group of data and provide an output in a fraction of the time that a human can perform this function. It can also help firms benefit from data processing by detecting patterns, enabling firms to allocate more time and resources to make better stock predictions. The advancement in efficiency offers much in the form of cost savings, however, to really extract the power of decision from ML the algorithms themselves and the data that is utilized need to be engineered correctly. Many ML algorithms can effectively make predictions, but really honing the accuracy of these predictions is dependent on the skill of the practitioner. This is where the art of implementing ML comes to light.

Conclusion

Going with – and not against – the tide and leveraging new skills to move forward in a new direction is necessary. Technology has effectively changed the world around us. These types of evolutions happen every so often – just like the Industrial Revolution. We are amid a changing landscape, not unlike the time when that first mechanical calculator was introduced. It is a good time to embrace the technological advancements around us.