Questions tagged [policies]
For questions related to policies (as defined in reinforcement learning or other AI sub-fields).
76
questions
16
votes
3
answers
4k
views
Is the optimal policy always stochastic if the environment is also stochastic?
Is the optimal policy always stochastic (that is, a map from states to a probability distribution over actions) if the environment is also stochastic?
Intuitively, if the environment is ...
- 39.7k
15
votes
4
answers
7k
views
What does "stationary" mean in the context of reinforcement learning?
I think I've seen the expressions "stationary data", "stationary dynamics" and "stationary policy", among others, in the context of reinforcement learning. What does it mean? I think stationary policy ...
- 428
8
votes
3
answers
16k
views
What is the difference between a stochastic and a deterministic policy?
In reinforcement learning, there are the concepts of stochastic (or probabilistic) and deterministic policies. What is the difference between them?
- 39.7k
8
votes
1
answer
7k
views
What is the difference between a stationary and a non-stationary policy?
In reinforcement learning, there are deterministic and non-deterministic (or stochastic) policies, but there are also stationary and non-stationary policies.
What is the difference between a ...
- 39.7k
6
votes
1
answer
549
views
What is the relation between a policy which is the solution to a MDP and a policy like $\epsilon$-greedy?
In the context of reinforcement learning, a policy, $\pi$, is often defined as a function from the space of states, $\mathcal{S}$, to the space of actions, $\mathcal{A}$, that is, $\pi : \mathcal{S} \...
- 39.7k
6
votes
0
answers
99
views
Proof that there always exists a dominating policy in an MDP
I think that it is common knowledge that for any infinite horizon discounted MDP $(S, A, P, r, \gamma)$, there always exists a dominating policy $\pi$, i.e. a policy $\pi$ such that for all policies $\...
- 185
5
votes
2
answers
618
views
Why is the derivative of this objective function 0 if the policy is deterministic?
In the Berkeley RL class CS294-112 Fa18 9/5/18, they mention the following gradient would be 0 if the policy is deterministic.
$$
\nabla_{\theta} J(\theta)=E_{\tau \sim \pi_{\theta}(\tau)}\left[\left(\...
- 153
5
votes
2
answers
1k
views
Given two optimal policies, is an affine combination of them also optimal?
If there are two different optimal policies $\pi_1, \pi_2$ in a reinforcement learning task, will the linear combination (or affine combination) of the two policies $\alpha \pi_1 + \beta \pi_2, \alpha ...
- 53
5
votes
1
answer
136
views
How do I compute the variance of the return of an evaluation policy using two behaviour policies?
Suppose there is an evaluation policy called $\pi_{e}$ and there are two behavior policies $\pi_{b1}$ and $\pi_{b2}$. I know that it is possible to estimate the return of policy $\pi_{e}$ through ...
- 481
4
votes
1
answer
145
views
Why doesn't value iteration use $\pi(a \mid s)$ while policy evaluation does?
I was looking at the Bellman equation, and I noticed a difference between the equations used in policy evaluation and value iteration.
In policy evaluation, there was the presence of $\pi(a \mid s)$, ...
- 125
4
votes
1
answer
510
views
A few questions regarding the difference between policy iteration and value iteration [closed]
The question already has some answer. But I am still finding it quite unclear (also does $\pi(s)$ here mean $q(s,a)$ ?):
The few things I do not understand are:
Why the difference between 2 ...
user9947
4
votes
1
answer
449
views
Why do we have two similar action selection strategies for UCB1?
In the literature, there are at least two action selection strategies associated with the UCB1's action selection strategy/policy. For example, in the paper Algorithms for the multi-armed bandit ...
- 39.7k
4
votes
2
answers
440
views
Why is having low variance important in offline policy evaluation of reinforcement learning?
Intuitively, I understand that having an unbiased estimate of a policy is important because being biased just means that our estimate is distant from the truth value.
However, I don't understand ...
- 227
4
votes
1
answer
621
views
Can someone please help me validate my MDP?
Problem Statement :
I have a system with four states - S1 through S4 where S1 is the beginning state and S4 is the end/terminal state. The next state is always better than the previous state i.e if ...
- 83
4
votes
1
answer
330
views
An example of a unique value function which is associated with multiple optimal policies
In the 4th paragraph of
http://www.incompleteideas.net/book/ebook/node37.html
it is mentioned:
Whereas the optimal value functions for states and state-action pairs are unique for a given MDP, ...
- 143
4
votes
1
answer
741
views
Why does having a fixed policy change a Markov Decision Process to a Markov Reward Process?
If a policy is fixed, it is said that a Markov Decision Process (MDP) becomes a Markov Reward Process (MRP).
Why is this so? Aren't the transitions and rewards still parameterized by the action and ...
- 43
3
votes
1
answer
555
views
Are there some notions of distance between two policies?
I want to determine some distance between two policies $\pi_1 (a \mid s)$ and $\pi_2 (a \mid s)$, i.e. something like $\vert \vert \pi_1 (a \mid s) - \pi_2(a \mid s) \vert \vert$, where $\pi_i (a\mid ...
- 287
3
votes
1
answer
149
views
In reinforcement learning, why are policies defined as functions of states and not observations?
I am new to RL and I am following Sutton & Barto's book.
My doubt is, when we talk about the policy of our agent, we say it is the probability of taking some action $a$ given the state $s$. ...
- 153
3
votes
1
answer
303
views
Why must the value of a state under an optimal policy equal the expected return for the best action from that state?
The Sutton and Barto reinforcement learning textbook states that
the value of a state under an optimal policy must equal the expected return for the best action from that state.
That is,
$$v_*(s) = \...
3
votes
1
answer
1k
views
What is a "learned policy" in Q-learning?
I am completing an assignment at the moment. One of the assignment questions asks how you identified the learned policy and how you obtained it. The question is a reinforcement learning question, and ...
user43972
3
votes
1
answer
327
views
How can we find the value function by solving a system of linear equations without knowing the policy?
An MDP is a Markov Reward Process with decisions, it’s an environment in which all states are Markov. This is what we want to solve. An MDP is a tuple $(S, A, P, R, \gamma)$, where $S$ is our state ...
- 45
3
votes
1
answer
153
views
What is meant by "generate the data" in describing the difference between on-policy and off-policy?
From the book:
Sutton, Richard S.,Barto, Andrew G.. Reinforcement Learning (Adaptive Computation and Machine Learning series) (p. 100). The MIT Press. Kindle Edition. "
following is stated:
"...
- 335
3
votes
2
answers
408
views
What is the difference between a non-stationary policy and a state that stores time?
This question is related to What does "stationary" mean in the context of reinforcement learning?, but I have a more specific question to clarify the difference between a non-stationary ...
- 428
3
votes
1
answer
1k
views
Why do RL implementations converge on one action?
I have seen this happening in implementations of state-of-the-art RL algorithms where the model converges to a single action over time after multiple training iterations. Are there some general ...
3
votes
2
answers
2k
views
Why do value iteration and policy iteration obtain similar policies even though they have different value functions?
I am trying to implement value and policy iteration algorithms. My value function from policy iteration looks vastly different from the values from value iteration, but the policy obtained from both ...
- 325
3
votes
3
answers
943
views
Is the policy really invariant under affine transformations of the reward function?
In the context of a Markov decision process, this paper says
it is well-known that the optimal policy is invariant to positive affine transformation of the reward function
On the other hand, ...
- 171
3
votes
1
answer
339
views
Can an optimal policy have a value function that has a smaller value for a state than a non-optimal policy?
I'm starting to learn about the Bellman Equation and a question came to my mind.
A policy $\pi$ is optimal if the value $v_\pi(s)$ is greater or equal than the value $v_{\pi'}(s)$ for all states $s \...
- 133
3
votes
2
answers
214
views
How does the AlphaGo Zero policy decide what move to execute?
I was going through the AlphaGo Zero paper and I was trying to understand everything, but I just can't figure out this one formula:
$$
\pi(a \mid s_0) = \frac{N(s_0, a)^{\frac{1}{\tau}}}{\sum_b N(s_0,...
- 33
3
votes
0
answers
110
views
Representation of state space, action space and reward system for Reinforcement Learning problem
I am trying to solve the problem of an agent dynamically discovering(start with no information about the environment) the environment and to explore as much of the environment as possible without ...
3
votes
0
answers
70
views
Can the importance sampling estimator have a non-stationary behaviour policy even if the target policy is stationary?
The inverse propensity score (IPS) estimator, which is used for off-policy evaluation in a contextual bandit problem, is well explained in the paper Doubly Robust Policy Evaluation and Optimization.
...
- 227
2
votes
3
answers
329
views
Why is the policy not a part of the MDP definition?
I'm reading an article on reinforcement learning, and I don't understand why the agent's policy $\pi$ is not part of definition of Markov Decision process(MDP):
Bu, Lucian, Robert Babu, and Bart De ...
- 187
2
votes
1
answer
738
views
Is Q-learning only capable of learning a deterministic policy?
I was following a reinforcement learning course on coursera and in this video at 2:57 the instructor says
Expected SARSA and SARSA both allow us to learn an optimal
$\epsilon$-soft policy, but, Q-...
2
votes
2
answers
58
views
What happens when the probability of either one of the policies is 0 in Importance Sampling?
I have a general question about the methods that use importance sampling in RL. What happens when the probability of either one of the policies is 0?
- 21
2
votes
1
answer
156
views
Why is there an inconsistency between my calculations of Policy Iteration and this Sutton & Barto's diagram?
In equation 4.9 of Sutton and Barto's book on page 79, we have (for the policy iteration algorithm):
$$\pi'(s) = arg \max_{a}\sum_{s',r}p(s',r|s,a)[r+\gamma v_{\pi}(s')]$$
where $\pi$ is the previous ...
- 147
2
votes
2
answers
862
views
Is my understanding of the value function, Q function, policy, reward and return correct?
I'm a beginner in the RL field, and I would like to check that my understanding of certain RL concepts.
Value function: How good it is to be in a state S following policy π.
...
- 123
2
votes
1
answer
61
views
If deep Q learning involves adjusting the value function for a specific policy, then how do I choose the right policy?
I wrote a simple implementation of Flappy Bird in Python, and now I'm trying to train an agent to play it at a reasonable skill level using TFLearn.
I feed the network an input vector of size 4:
...
- 53
2
votes
1
answer
93
views
Is it common to have extreme policy's probabilities?
I have implemented several policy gradient algorithms (REINFORCE, A2C, and PPO) and am finding that the resultant policy's action probability distributions can be rather extreme. As a note, I have ...
- 33
2
votes
1
answer
194
views
Why do I need an initial arbitrary policy to implement value iteration algorithm
I've been recently given an assignment based on Reinforcement Learning and I'm supposed to implement the value iteration algorithm in a grid environment.
The assignment:
My doubt is why do I even ...
- 121
2
votes
1
answer
101
views
Possible inconsistency in the Policy Improvement equation
I came across this formula in Sutton And Barto: RL an Intro (2nd Edition) equation number 4.7 (page number 78).
If $\pi$ and $\pi'$ are deterministic policies and $q_\pi(s, \pi'(s)) \geq v_\pi(s)$ ...
user9947
2
votes
1
answer
322
views
Is the initialisation of $V(s)$ and $\pi(s)$ really arbitrary in policy iteration?
In Sutton and Barto's book (Reinforcement learning: An introduction. MIT press, 2018), the algorithm "Policy Iteration" is:
Here, $V(s)$ is initialized arbitrarily, meaning that I can ...
- 121
2
votes
1
answer
106
views
Why do I get the best policy before Q values converge using DQN?
I have implemented DQN algorithm and wonder why during testing, the best performance is achieved by a policy from about 300 episode, when mean Q values converge at about 800 episode?
Mean Q-values ...
user43110
2
votes
1
answer
104
views
Is a learned policy, for a deterministic problem, trained in a supervised process, a stochastic policy?
If I trained a neural network with 4 outputs (one for each action: move down, up, left, and right) to move an agent through a grid (deterministic problem). The output of the neural network is a ...
- 21
2
votes
3
answers
175
views
How to estimate a behavior policy for off-policy learning based on data?
I have a dataset which includes states, actions, and reward. The dataset includes information on the transition, i.e., $p(r,s' \mid s,a)$.
Is there a way to estimate a behavior policy from this ...
- 341
2
votes
0
answers
54
views
Does there necessarily exist "dominated actions" in a MDP?
In a Markov Decision Process, is it possible that there exists no "dominated action"?
I define a dominated action the following way:
we say that $(s,a)$ is a dominated action, if $\forall \...
- 31
2
votes
0
answers
103
views
What kind of reinforcement learning method does AlphaGo Deepmind use to beat the best human Go player?
In reinforcement learning, there are model-based versus model-free methods. Within model-based ones, there are policy-based and value-based methods.
AlphaGo Deepmind RL model has beaten the best Go ...
- 201
2
votes
0
answers
52
views
What are some other real-life examples of simple policies but complex value functions?
Hado van Hasselt, a researcher at DeepMind, mentioned in one of his videos (from 7:20 to 8:20) on Youtube (about policy gradient methods) that there are cases when the policy is very simple compared ...
- 1,131
2
votes
0
answers
28
views
Do we assume the policy to be deterministic when proving the optimality?
In reinforcement learning, when we talk about the principle of optimality, do we assume the policy to be deterministic?
- 153
2
votes
1
answer
75
views
Is a policy in reinforcement learning analogous to a field such as APF?
If a policy maps states to actions in reinforcement learning, then for a path planning with obstacles, can't we simply use Artificial Potential Field fields for path planning and model policy ...
- 171
2
votes
0
answers
115
views
Does the off-policy evaluation work for non-stationary policies?
As the title says, in reinforcement learning, does the off-policy evaluation work for non-stationary policies?
For example, IS (importance sampling)-based estimators, such as weighted IS or doubly ...
- 227
2
votes
0
answers
282
views
AlphaGo Zero: Does the policy head give a probability for every possible move?
If I understood correctly, the AlphaGo Zero network returns two values: a vector of logit probabilities p and a value v.
My question is: in this vector that it is outputted, do we have a probability ...
- 265