New answers tagged policy-gradients
1
vote
In policy gradient methods why do we compute the gradient of the objective function through a one-trajectory estimate?
Sutton & Barto's book of Reinforcement Learning mentions below:
When the state-value function is used to assess actions in this way it is called a critic, and the overall policy-gradient method ...
1
vote
Accepted
In policy gradient methods why do we compute the gradient of the objective function through a one-trajectory estimate?
The one-trajectory sample is the Monte-Carlo way to estimate the gradient, which, turns out, to be an unbiased estimator although with high-variance due to relying only on one sample: I think this is ...
Top 50 recent answers are included
Related Tags
policy-gradients × 194reinforcement-learning × 173
deep-rl × 31
reinforce × 31
actor-critic-methods × 21
proximal-policy-optimization × 20
neural-networks × 14
ddpg × 11
deep-learning × 10
machine-learning × 8
proofs × 8
sutton-barto × 8
comparison × 7
python × 7
math × 7
objective-functions × 7
pytorch × 7
rewards × 7
backpropagation × 6
markov-decision-process × 6
off-policy-methods × 6
continuous-action-spaces × 6
trust-region-policy-optimization × 6
policy-gradient-theorem × 6
keras × 5