Source (reference) of the implementation

[akmandor]

Is your implementation based on this paper?

[peter-ch]

No. This code attempts to construct neural networks by selecting and adjusting the NN weights as if it was a computer game, also adding/removing nodes. It uses RL to learn how to do that, but it doesn't do hyperparameter optimization.

[NimmiW]

@akmandor https://github.com/hadijomaa/HypRL