Build an AI to beat the game Breakout part-2
Now after resoliving the environment problem we started implementing the brain of the project which is the A3C algorithm, so now we have to create a convolutional layer ->pooling layer-> flattering layer
And in the convolutional neural network we have to implement the A3C algoritm with 3 Agent ,we can have more than three Actor as of now we are just creating the 3 Agent to working simultaneously by using the SHARED MODEL
so this is how how the model is training faster than the other model but when we include this algorithm alone the result wasn’t that great to say because it took a longer time to train this is not the optimal part of it so when we came to see the backend of how it works in order to solve this problem .
we came to know that this feature extraction works on the principle of taking pixels from each movement for your better understanding i have enclosed the image below,
as we see in the (fig 1) we can see that the next move will be left downwards but the AI dont know what to do, it will only come to know that it is moving left downwards after many softamx functions.
so this is were we found the problem that it is taking much time to think(train) and make the correct decision
so Here comes the idea to the problem ,what if we use the LSTM layer🤔
so finally the LSTM layer has solved the problem because, the process of LSTM layer was to save the provious move of the input in the hidden layers itself, now there wont be a problem of making the prediction of the ball where is going , all these thing are done in the hidden layer itself, this makes the algorithm works way better that the others
so finally after this, the problem we faced in improving the model was solved and gave a very good result of traing the model in just ~9hours with just 8GB RAM i7 configuration desktop/laptop