1
$\begingroup$

I see that my policy entropy is decreasing very fast and converges to zero in no time, causing the policy to sample the same action again and again (which results in a sub-optimal behavior). I think a reasonable approach here will be to add an entropy term to the A3C loss function. In this case, how do you choose a value for the entropy coefficient in the loss function of an A3C (actor-critic) algorithm?

$\endgroup$