Artificial intelligence is actively being used by a growing set of asset managers. For managers that already a quantitative shop, this should not be a large stretch. Artificial intelligence is not just advancement in data modeling. It is a natural progression for those managers searching for an edge and have exhausted older more well-known data analysis techniques.
A recent white paper from Meketa Investment Group called "Artificial Intelligence" walks through some of the basics of artificial intelligence and discuss how these techniques are being used in asset management. As well as providing some basic information and definitions, this paper provides context for how artificial intelligence fits within overall data analysis. AI techniques are clear extensions of simpler techniques. The concepts of deep learning and machine learning are just subsets within the greater field of artificial intelligence.
Innovation is about pushing the boundaries for what is possible. Once the existing techniques have been exhausted for analysis new ideas have to be tried. In quant investing, the normal process of hypotheses and testing has given way to unsupervised learning and more complex ways to examine data. If you cannot think of anymore ideas, let the data speak for themselves. Artificial intelligence arrives when existing learning is at an impasse.
The firms that are actively using artificial intelligence techniques have in common a history of using quantitative analysis. In the case of the trend-followers, there existence is based on searching for data patterns. With the market environment become more competitive and recent periods of poorer performance, artificial intelligence and machine learning are being used to search for new return opportunities. Competition and failure coupled with a desire to succeed provides the impetus to try new techniques. From failure and new questions comes the desire for new learning and the use of new techniques for analysis. Of course, cheap computing power helps.
