Hello

[yingshaoxo]

I really appreciate what you have done there.

But according to what I have seen from your gif, it seems work badly at even this simple game(Mario).

---

Will it be better to use traditional deep learning tech? for example, supervised learning.

First, we train a model(neural network) by supervised learning(capture screen image as an input, capture keyboard as an output), then we use this model as the base of deep Q network.

---

I don't know if it's work.

I only knew some basics about keras. But I knew Python very well. I have one month to do this, will you give me some advice when I'm getting stuck later?

[Naereen]

Hi !
Yes, you guessed correctly from the gif, the learning approach I tried her did not work at all. I tried to play and experiment with the network types, sizes etc, and I got literally nothing but discouraging results.
Your idea sounds interesting and could work.

However, I am sorry but I have no time to help you, not even to give advice or hints, sorry.
I'm not sufficiently skilled on supervised learning and what you propose to try, so I have no expertise to help you.

My code is under an open source license, so you are of course free to fork it and try as much as you want!
But no, sorry, I cannot help. Good luck and enjoy 😄!