Questions tagged [exploration-exploitation-tradeoff]
For questions related to the exploration-exploitation trade-off (or dilemma) in the context of reinforcement learning.
34
questions
0
votes
0
answers
21
views
thompson, UCB, e-greedy vs information state space algorithms for bandits
I am trying to understand why UCB, thompson sampling etc are inferior to information state space bandits in certain cases.
Consider page 25 topic "Value of information" and below in the ...
1
vote
1
answer
23
views
Practical differences between the different RL architectures
I have tried many different RL architectures: DQN, PPO, Policy optimization, and for my specific problem they all failed in their basic setup. Eventually I discovered that my problem had too sparse/...
0
votes
0
answers
94
views
Methods for sequential decision optimization problem with nonlinear bayesian reward function
I am attempting to grasp if there are any other methods out there that i am not aware of that can be beneficial given my problem context.
Being inspired from optimal experimental design communities ...
2
votes
1
answer
167
views
In which community does using a Bayesian regression model as a reward function with exploration vs. exploitation challenges fall under?
I am trying to find research papers addressing a problem that, in my opinion, deserves significant attention. However, I am having difficulty locating relevant information.
To illustrate the problem ...
0
votes
1
answer
45
views
In DQN, how to increase epsilon but not too much?
I am using DQN algorithm in a non-stationary problem in a continuous learning. My environment gives me some new states each T steps. For example after 10 000 steps, I get some new states and I need to ...
1
vote
1
answer
56
views
Metrics to compare the exploraation of RL Algorithms
I am looking for metrics to compare the exploration under different RL Algos/reward functions. I want to somehow quantify how big of a region of the policy space is explored. What are common measures ...
2
votes
0
answers
218
views
Choosing and Designing Decay Types for Epsilon-Greedy Exploration in Reinforcement Learning
I am working on a reinforcement learning project that involves epsilon-greedy exploration. I have two questions regarding the choice between linear and exponential decay for epsilon, and the ...
1
vote
2
answers
57
views
Are RL algorithms suppose to keep learning?
I don't understand if the purposes of RL agents is simply optimizing a model with a reward instead of using labeled data (i.e. in a supervision fashion), or they have also the purpose of keep training ...
4
votes
1
answer
954
views
Mathematically, what is happening differently in the neural net during exploration vs. exploitation?
I want to understand roughly what is happening in the neural network of an RL agent when it is exploring vs. exploiting. For example, are the network weights not being updated when the agent is ...
5
votes
6
answers
3k
views
Why is exploitation necessary during training?
I have read many blog articles making all kinds of broad analogies to explain the exploration/exploitation trade-off. However, I still can't fully grasp it. On an extremely abstract level, I ...
0
votes
1
answer
241
views
How to reduce the number of episodes before the agent learns in this game?
The initial environment state is 0.25. Each time step the agent performs a discrete action of 0 or ...
1
vote
1
answer
123
views
Is there a notion of exploration-exploitation tradeoff in dynamic programming (or model-based RL)?
Is there a notion of exploration-exploitation tradeoff in dynamic programming (or model-based RL)?
1
vote
1
answer
70
views
Which policy has to be followed by a player while construction of its own Q-table?
Consider the scenario, where there are two players. One of the players perform the action randomly, whereas I want second player as a Q-player. I mean, the player selects a best action from the Q-...
5
votes
3
answers
651
views
In Q-learning, wouldn't it be better to simply iterate through all possible states?
In Q-learning, all resources I've found seem to say that the algorithm to update the Q-table should start at some initial state, and pick actions (which are sometimes random) to explore the state ...
2
votes
0
answers
362
views
Why is the ideal exploration parameter in the UCT algorithm $\sqrt{2}$?
From Wikipedia, in the Monte-Carlo Tree Search algorithm, you should choose the node that maximizes the value:
$${\displaystyle {\frac {w_{i}}{n_{i}}}+c{\sqrt {\frac {\ln N_{i}}{n_{i}}}}},$$
where
${...
1
vote
0
answers
268
views
What is the difference between exploitation and exploration in the context of optimization?
In the paper Moth-flame optimization algorithm: A novel nature-inspired heuristic paradigm (2015, published in Knowledge-Based Systems)
The test functions are divided to three groups: unimodal, multi-...
1
vote
0
answers
314
views
Should the exploration rate be updated at the end of the episode or at every step?
My agent uses an $\epsilon$-greedy strategy to learn. The exploration rate (i.e. $\epsilon$) decays throughout the training. I've seen examples where people update $\epsilon$ every time an action is ...
3
votes
1
answer
212
views
Can we stop training as soon as epsilon is small?
I'm new to reinforcement learning.
As it is common in RL, $\epsilon$-greedy search for the behavior/exploration is used. So, at the beginning of the training, $\epsilon$ is high, and therefore a lot ...
4
votes
2
answers
996
views
What is the meaning of "exploration" in reinforcement and supervised learning?
While exploration is an integral part of reinforcement learning (RL), it does not pertain to supervised learning (SL) since the latter is already provided with the data set from the start.
That said, ...
2
votes
1
answer
173
views
What is the optimal exploration-exploitation trade-off in Q*bert?
I am training an RL agent with Deep Q-learning + Experience Replay on the Q*bert Atari environment. After 400,000 frames, my agent appears to have learned strategic information about the game, but ...
0
votes
0
answers
81
views
How to modify the Actor-Critic policy gradient algorithm to perform Safe exploration in Reinforcement Learning
I am trying to implement safe exploration technique in [Ref.1]. I am using Soft Actor-Critic algorithm to teach an agent to introduce a bias between 0 and 1 to a specific state of interest in my ...
2
votes
1
answer
167
views
Why do some DQN implementations not require random exploration but instead emulate all actions?
I've found online some DQN algorithms that (in a problem with a continuous state space and few actions, let's say 2 or 3), at each time step, compute and store (in the memory used for updating) all ...
2
votes
1
answer
498
views
Why is 100% exploration bad during the learning stage in reinforcement learning?
Why can't we during the first 1000 episodes allow our agent to perform only exploration?
This will give a better chance of covering the entire space state. Then, after the number of episodes, we can ...
3
votes
2
answers
209
views
Why is it not advisable to have a 100 percent exploration rate? [duplicate]
During the learning phase, why don't we have a 100% exploration rate, to allow our agent to fully explore our environment and update the Q values, then during testing we bring in exploitation? Does ...
1
vote
1
answer
71
views
How to deal with the addition of a new state to the environment during training?
Let's say we have a dynamic environment: a new state gets added after 2000 episodes have been done. So, we leave room for exploration, so that it can discover the new state.
When it gets to that new ...
1
vote
2
answers
62
views
Why do we explore after we have an accurate estimate of the value function?
Suppose we have a small space state and that, after about 2000 episodes, we've accurately explored the environment and known the accurate $Q$ values. In that case, why do we still leave a small ...
1
vote
2
answers
145
views
Why can't we fully exploit the environment after the first episode in Q-learning?
During the first episode, it's 100% exploration, because all our Q values are 0. Suppose we have 1000 time steps, and it's terminated by meeting a reward. So, after the first episode, why can't we ...
4
votes
2
answers
2k
views
Should I use exploration strategy in Policy Gradient algorithms?
In policy gradient algorithms the output is a stochastic policy - a probability for each action.
I believe that if I follow the policy (sample an action from the policy) I make use of exploration ...
1
vote
1
answer
488
views
Can tabular Q-learning converge even if it doesn't explore all state-action pairs?
My understanding of tabular Q-learning is that it essentially builds a dictionary of state-action pairs, so as to maximize the Markovian (i.e., step-wise, history-agnostic?) reward. This incremental ...
1
vote
0
answers
270
views
How can I increase the exploration in the Proximal Policy Optimation algorithm?
How can I increase the exploration in the Proximal Policy Optimation reinforcement learning algorithm? Is there a variable assigned for this purpose? I'm using the stable-baseline implementation: ...
1
vote
1
answer
71
views
Should I just use exploitation after I have trained the Q agent?
When using a trained Q-learning algorithm in an actual game, would I just use exploitation and no longer use exploration? Should I use exploration only during the training phase?
1
vote
0
answers
588
views
Why is the $\epsilon$ hyper-parameter (in the $\epsilon$-greedy policy) annealed smoothly?
As far as I understand, RL is a process that can be divided into 2 stages:
Exploring a wide range of paths (acting randomly)
Refining the current optimal paths (revolving around actions with a so-...
8
votes
1
answer
680
views
What is the purpose of the actor in actor-critic algorithms?
For discrete action spaces, what is the purpose of the actor in actor-critic algorithms?
My current understanding is that the critic estimates the future reward given an action, so why not just take ...
2
votes
1
answer
315
views
Does eligibility traces and epsilon-greedy do the same task in different ways?
I understand that, in Reinforcement Learning algorithms, such as Q-learning, to prevent selecting the actions with greatest q-values too fast and allow for exploration, we use eligibility traces.
Here ...