At Intelligent Trading Foundation, we are committed to building AI solutions that help humans solve hard problems.
Devising a good trading strategy based on a variety of technical analysis indicators definitely qualifies as a hard problem.
Since the inception of our company, our AI team has been diligently working on creating machine learning algorithms that help cryptocurrency traders profit from the market.
Artificial neural networks have been our primary focus. However, in the last few months we’ve started experimenting with genetic programming.
Genetic programming is a new branch of AI research that allows us to start from simple and probably not very useful trading strategies, and through simulating evolution including mutation, crossing, and survival of the fittest, arrive at complex trading strategies that work.
One great thing about genetic programming is that humans can look at these strategies, understand what they’re doing, and then decide whether to use them for future time periods, tweak and integrate them with trading systems, or just treat them as a learning tool.
This is a major advantage over artificial neural networks which behave like black boxes, with little chance of humans understanding how the network reached its decision.
Below is a video explaining the genetic programming work we are doing at ITF.