Forex Neural Network Tracker

Current State

Training Factory:

Testing Factory:

What this Means

Each network gets 5 days to trade in the market and the ones that do the best get to advance to the next generation. Fitness is a combination of money earned and other indicators of good performance, such as max drawdown, difference between balance and equity, and total loss.

The top 5 networks in each generation get sent to the testing factory. Here they are tested for a much longer period of time, 30 days. Each network gets tested 250 times and the testing factory uses these tests to determine what percent of the time the network is able to make money on average, along with other economic indicators of success, such as max drawdown, total profit/ loss, and profitability ratio

What is this?

This project uses Artificial Intelligence to learn to trade in the Forex (Foreign Exchange) market. It is currently training and you can see how it improves through training in real time on this website.

This project uses NeuroEvolution technology, specifically the Neuro Evolution through Augmenting Topologies (NEAT) algorithm. It is written in python and is hosted via [Host currently not decided]. It uses the Oanda API for historical market data, The Eurostat database for European economic data, and the Federal Reserve Economic Data (FRED) database for United States economic data.

The effectiveness of NeuroEvolution has not been well demonstrated on complex tasks but there is evidence that NeuroEvolution can provide similar results to Reinforcement Learning algorithms with shorter training times and less computing resources. I beleive that the technology has a lot of potential to solve the world's important problems and I hope to demonstrate its effectiveness on a complex problem through this project.

References and helpful links:
How Neural Networks work
NEAT algorithm paper  Paper summary
Open AI comparison of RL vs ES

About Me

My name is Cole Lashley. I am a 2nd year student at the University of Maryland, majoring in Computer Science. I enjoy singing and playing instruments, spending time with friends, playing video games, and programming. I am passionate about Computer Science because I believe that technology has the power to change the world in unprecedented ways.

Picture of Me

A picture of me