Investing with AI: Does It Actually Work?
Why financial machine learning hasn’t kept up
In recent times, thanks to developments in both research and computation power, we have seen vast advances in the field of machine learning with applications spanning fields such as e-commerce, manufacturing and transportation. So much so that in some narrow cases AI has been shown to exceed even human capabilities. Intuitively, it stands to reason that the investors big and small alike should also be beneficiary to this trend as the environment is rich with data and the mere mention of stocks is enough to conjure up mental images of price charts ripe with hidden patterns.
However this has not been the reality. There’ve been no breakthrough papers that have taken the community by storm, an increasing amount of investment managers are finding it more difficult to keep up with the index let alone exceed it and the predominant portfolio approach still remains the mean-variance paradigm imparted onto us by the financial academics of yesteryear. Thus it begs the question, why have we not seen widespread adoption of AI in investment?
