"Recent update: Struggling with reinforcement learning" (1/5) Investment AI development struggles

For now, here is a brief status update.

I have created a market-prediction AI for the next generation EA, and am currently building an AI that trades based on these predicted values.

Until now, the EA used deep learning to create the AI, and it traded when the predicted value exceeded the actual value.

However, the next target EA is to install an AI that makes trading decisions using the predictions from the deep learning AI.

This is the reinforcement learning AI.

I will post the performance of the reinforcement learning today as a memo.

Even with reinforcement learning over a one-month period, after about 400 episodes, the number of winning trades started to increase.

A few years of reinforcement learning seems challenging...