Hello everyone) I’m learning to work with Learning Agents in Unreal Engine 5.4. I got used to the machines and sort of understood how they work. I wanted to create a simple neural network to test my knowledge. And I had a suspicion that the plugin does not interact with the input values in any way.
I have 2 Float variables in my test project. In theory, the neural network should receive 1 value A as input, and should produce value B as output.
To simplify, the neural network should arrive at the formula A=B.
The problem is that even after 5000 iterations, the neural network still produces “random” answers. I got the feeling that I did something wrong with the input signals. Please help me figure it out.
As a reward, I compare the difference between A and B. Also, if they match completely, the agent will receive a super reward.
I am attaching screenshots. If a similar issue has already been resolved, please provide a link.
