In Feb 2015, Ray Dalio, founder and co-CIO of Bridgewater Associates, one of the biggest hedge funds all over the world, starts the artificial intelligence (AI) team. According to Bloomberg.
I wonder why he does so. He already maintains fantastic track records for the long term. In order to understand what is in his mind, I read the articles about Ray Dalio and watch the videos including famous "HOW THE ECONOMIC MACHINE WORKS". He repeated many times "We should know how the economic machine works", "We always should learn what we do not know". These philosophies are always consistent in his statements. Then I found that these are similar to logic of artificial intelligence or machine learning. When we use artificial intelligence or machine learning, first we construct the mathematical models, which are sometimes called "hypothesis". In my view, Ray's "economic machine"can be represented as models in computers. Then these models can learn new things when new data are provided. In other words, models may adjust to environmental changes by themselves. So I imagine that Ray likes these characteristics about artificial intelligence or machine learning. This is my view on why he starts AI team in the hedge fund.
Then I wonder what is good if AI or machine learning is used for investments. According to the article, the team reports to David Ferrucci. He used to be the researcher who led the development of Watson before joining Bridgewater Associates. Watson is a famous "cognitive computing" machine. I read the article about Watson which applies to cancer treatments in US and imagine how it may be used to make investment decisions and advantages of usage of computers. It gives us some clues about what investment models will be developed in the hedge fund.
1. Computers can operate 24hours 365 days at full speed
Computers need no rest time. Whatever data should be analyzed, computers can do that all day. Global macro investments like Bridgewater Associates need a lot of information all over the world to make the diversified asset allocation. The more data we have, the more advantages the computers have.
2. Computers can read natural language, such as English.
This technology is sometimes called "Natural language processing". Watson became famous when it defeated the human champions on Jeopardy!, American game show in 2011. Therefore, we analyze not only numerical data, but also natural language at tremendous speed. So economic research papers can be used for analysis. This is also an advantage for using computers. With advantage 1 above, it makes computers more powerful for investment.
3. Computers can learn from data by themselves
This is the biggest advantage to use computers for investment. Although computers might make bad decisions initially, they can be smarter and better by learning. This is just like process where children learn languages. The more they learn, the smarter they become. The economic environment is always changing so it is critically important to adjust the changes.
No one knows what Ray thinks in his mind. So there is no clear answer why he starts Artificial-Intelligence Team. Therefore, you can have your own answer to that. This is a good starting point to think how AI or machine learning will impact financial industry in the future. We have some clues according to the comments from Bridgewater Associates. What do you think?