Skip to main content
Artificial Intelligence

Return to Question

edited tags; edited title
Source Link
nbro
nbro
  • 41.4k
  • 12
  • 115
  • 205

Reinforcement learning: Output layer Regarding the output layer's activation function for continuous actionsaction space problems

I'm interested in building a (deep) RL agent for solving a continuous problem (which splits something into portions).

In all examples I've seen so far, e.g., solving the continuous lunar lander, always a tanh$\tanh$ output layer activation was used, which produces values between -1$-1$ and +1$+1$.

Is this just because it fits the use case or is this a general rule for RL agents with continuous action spaces?

What if I just want values between 0$0$ and 1$1$? Could I simply use a softmax$\operatorname{softmax}$ activation for my output layer?

Reinforcement learning: Output layer activation function for continuous actions

I'm interested in building a (deep) RL agent for solving a continuous problem (which splits something into portions).

In all examples I've seen so far, e.g., solving the continuous lunar lander, always a tanh output layer activation was used, which produces values between -1 and +1.

Is this just because it fits the use case or is this a general rule for RL agents with continuous action spaces?

What if I just want values between 0 and 1? Could I simply use a softmax activation for my output layer?

Regarding the output layer's activation function for continuous action space problems

I'm interested in building a (deep) RL agent for solving a continuous problem (which splits something into portions).

In all examples I've seen so far, e.g., solving the continuous lunar lander, always a $\tanh$ output layer activation was used, which produces values between $-1$ and $+1$.

Is this just because it fits the use case or is this a general rule for RL agents with continuous action spaces?

What if I just want values between $0$ and $1$? Could I simply use a $\operatorname{softmax}$ activation for my output layer?

comma
Source Link
stefanbschneider
stefanbschneider
  • 183
  • 2
  • 11

I'm interested in building a (deep) RL agent for solving a continuous problem (in which portions to splitwhich splits something into portions).

In all examples, I've seen so far, e.g., solving the continuous lunar lander, always a tanh output layer activation was used, which produces values between -1 and +1.

Is this just because it fits the use case or is this a general rule for RL agents with continuous action spaces?

What if I just want values between 0 and 1? Could I simply use a softmax activation for my output layer?

I'm interested in building a (deep) RL agent for solving a continuous problem (in which portions to split something).

In all examples, I've seen so far, e.g., solving the continuous lunar lander, always a tanh output layer activation was used, which produces values between -1 and +1.

Is this just because it fits the use case or is this a general rule for RL agents with continuous action spaces?

What if I just want values between 0 and 1? Could I simply use a softmax activation for my output layer?

I'm interested in building a (deep) RL agent for solving a continuous problem (which splits something into portions).

In all examples I've seen so far, e.g., solving the continuous lunar lander, always a tanh output layer activation was used, which produces values between -1 and +1.

Is this just because it fits the use case or is this a general rule for RL agents with continuous action spaces?

What if I just want values between 0 and 1? Could I simply use a softmax activation for my output layer?

Source Link
stefanbschneider
stefanbschneider
  • 183
  • 2
  • 11

Reinforcement learning: Output layer activation function for continuous actions

I'm interested in building a (deep) RL agent for solving a continuous problem (in which portions to split something).

In all examples, I've seen so far, e.g., solving the continuous lunar lander, always a tanh output layer activation was used, which produces values between -1 and +1.

Is this just because it fits the use case or is this a general rule for RL agents with continuous action spaces?

What if I just want values between 0 and 1? Could I simply use a softmax activation for my output layer?