Questions tagged [thompson-sampling]

For questions about Thompson sampling, which is a technique for choosing actions (that addresses the exploration-exploitation dilemma) in the multi-armed bandit and reinforcement learning problems.

Filter by
Sorted by
Tagged with
0 votes
0 answers
17 views

Thompson sampling, is it accurate for smaller sample sizes?

For example are 500 samples enough? I tried this code ...
  • 1
2 votes
0 answers
13 views

Minimum sampling for maximising the prediction accuracy

Suppose that I'm training a machine learning model to predict people's age by a picture of their faces. Lets say that I have a dataset of people from 1 year olds to 100 year olds. But I want to choose ...
  • 121
1 vote
0 answers
32 views

Data Imbalance in Contextual Bandit with Thompson Sampling

I'm working with the Online Logistic Regression Algorithm (Algorithm 3) of Chapelle and Li in their paper, "An Empirical Evaluation of Thompson Sampling" (https://papers.nips.cc/paper/2011/...
  • 11
0 votes
0 answers
44 views

Is there a variant of Thompson Sampling that works with variable bandits?

Does there exist a variant of TS, such that, while computing the returns of multi-armed bandits, we have the possibility of introducing an extra bandit? For instance, while we are applying TS to 3 ...
4 votes
0 answers
98 views

What is the Thompson Sampling in simple terms?

I am looking at the different existing methods of action selection in reinforcement learning. I found several methods like epsilon-greedy, softmax, upper confidence bound and Thompson sampling. I ...
0 votes
1 answer
74 views

Why is Thompson Sampling considered a part of Reinforcement Learning?

I often see Thompson Sampling in RL literature, however, I am not able to relate it to any of the current RL techniques. How exactly does it fit with RL?
0 votes
0 answers
48 views

Thompson sampling with Bernoulli prior and non-binary reward update

I am solving a problem for which I have to select the best possible servers (level 1) to hit for a given data. These servers (level 1) in turn hit some other servers (level 2) to complete the request. ...
3 votes
3 answers
280 views

Why aren't exploration techniques, such as UCB or Thompson sampling, used in full RL problems?

Why aren't exploration techniques, such as UCB or Thompson sampling, typically used in bandit problems, used in full RL problems? Monte Carlo Tree Search may use the above-mentioned methods in its ...
  • 321
1 vote
0 answers
38 views

Multi-armed bandits: reducing stochastic multi-armed bandits to bernoulli bandits

Agrawal and Goyal (http://proceedings.mlr.press/v23/agrawal12/agrawal12.pdf page 3) discussed how we can extend Thompson sampling for bernoulli bandits to Thompson sampling for stochastic bandits in ...
  • 287
1 vote
1 answer
2k views

Why am I getting better performance with Thompson sampling than with UCB or $\epsilon$-greedy in a multi-armed bandit problem? [closed]

I ran a test using 3 strategies for multi-armed bandit: UCB, $\epsilon$-greedy, and Thompson sampling. The results for the rewards I got are as follows: Thompson sampling had the highest average ...
3 votes
2 answers
1k 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 ...
  • 165
7 votes
0 answers
94 views

Normalizing Normal Distributions in Thompson Sampling for online Reinforcement Learning

In my implementation of Thompson Sampling (TS) for online Reinforcement Learning, my distribution for selecting $a$ is $\mathcal{N}(Q(s, a), \frac{1}{C(s,a)+1})$, where $C(s,a)$ is the number of times ...
  • 71
4 votes
1 answer
544 views

How to compute the action probabilities with Thompson sampling in deep Q-learning?

In some implementations of off-policy Q-learning, we need to know the action probabilities given by the behavior policy $\mu(a)$ (e.g., if we want to use importance sampling). In my case, I am using ...
  • 43