Question

Which of the following statements is NOT true about Thompson Sampling or Posterior Sam- pling? After each sampling of an arm, the distributions for the arm becomes thinner(considering bell shaped reward distributions for arms) It is shown that Thompson sampling generally gives better regret bound than UCB In Thompson sampling, eliminating arms is necessary for getting better complexity All are true

Answer 1

Answer:

yes all are true..........

Answer 2

Answer: The statement that is NOT true about Thompson Sampling or Posterior Sampling is:

"In Thompson sampling, eliminating arms is necessary for getting better complexity."

Thompson Sampling does not require eliminating arms to improve its complexity. On the contrary, Thompson Sampling can handle a large number of arms efficiently because it only needs to maintain posterior distributions over the arms.

The sampling process can be performed independently for each arm, making it computationally scalable.

The other statements are true:

After each sampling of an arm, the distributions for the arm becomes thinner (considering bell-shaped reward distributions for arms).

It is shown that Thompson Sampling generally gives better regret bound than UCB.

In Thompson Sampling, eliminating arms is not necessary for getting better complexity.

Learn more about Sampling here

https://brainly.in/question/6485918

Learn more about complexity here

https://brainly.in/question/924407

#SPJ2