It has been a while since I wrote my last article. There has been hectic time with: work, Xmas break and developing AI based strategies applied to commodities and volatility products.
As I am getting better and better in understanding advanced forms of AI and their potential, I told to myself why not to share with the community what it is and how I am applying it to our preferred strategy: NEXT-alpha.
What is it?
In computer science, artificial intelligence (AI), sometimes called machine intelligence, is intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans. Leading AI textbooks define it as any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals. Colloquially, the term "artificial intelligence" is often used to describe machines (or computers) that mimic "cognitive" functions that humans associate with the human mind, such as "learning" and "problem solving".
A typical AI analyzes its environment and takes actions that maximize its chance of success. An AI's intended goal can be as simple as: 1 if the AI wins a game, 0 otherwise, or complex such as repeat actions that have been successful in the past. AI often revolves around the use of algorithms. An algorithm is a set of unambiguous instructions that a mechanical computer can execute. Many AI algorithms are capable of learning from data; they can enhance themselves by learning new strategies, or "rules of thumb", that have worked well in the past. Alternatively in the most complex cases: they can themselves write other algorithms.
Intelligent but not as human mind can do
The cognitive capabilities of current architectures are very limited, using only a simplified version of what intelligence is really capable of. For instance, the human mind has come up with ways to reason beyond measure and logical explanations to different occurrences in life. In simpler words: humans can innovate and move beyond the status quo, machines can only move on a straight line and hardly make evolutionary innovations that goes beyond the current status of the world we live in.
Does AI learn?
Yes, it does. It does it continually and thus constantly improving its ability to take decisions and successfully execute a task. The computer algorithms automatically constantly improve based on what they learn on daily basis!
How to apply this to finance?
The equity market has existed for centuries, the future market for millennial (Yes! From the time of ancient Greece). Like Ray Dalio says understanding the mechanisms behind long and short term cyclicality of the market in its broader sense (stocks, bonds, metals,…) can help to forecast the most likely event to take place in the future.
Now, once a mechanism is well understood, because it is a set of consecutive logical instructions, it can then be transformed in an algorithm, and once we have it in our hands we can then train it using a multitude of historical data that in some cases are tracing back to the beginning of 1900.
With NEXT-alpha this is exactly what we are doing. We have used historical data to train our AI model to understand the most likely asset that is able to generate money in the next 1 to 3 months. Algorithms keep on evolving over time and so it was for NEXT-alpha. It went through some adjustments since it is inception back in 2017. The core principles are still there, the percentage of positive months and magnitude of returns keeps getting better and better.
note: part of the article has been a re-edit from Wikipedia