Study

Review Final Exam

  •   0%
  •  0     0     0

      Back   Restart
  • If γ = 0, the agent cares about:
    Immediate reward only
  • MDP stands for:
    Markov Decision Process
  • What does transition function T(s,a,s') represent?
    Probability of moving to state s′ after action a in state s.
  • When do we stop Value Iteration?
    When values stop changing significantly.
  • How is a centroid updated in K-means?
    By taking the mean of all assigned points.
  • What does γ (gamma) represent?
    Discount factor
  • Why can K-means give different results on the same dataset?
    Random centroid initialization.
  • Policy iteration consists of:
    Policy evaluation + policy improvement
  • What is the main goal of PCA?
    PCA reduces dimensions while preserving maximum variance.
  • K-means is what type of learning?
    Unsupervised
  • What happens first in K-means?
    Assign points
  • If K increases, WCSS(cost function) usually:
    Decreases
  • In PCA, what does the first principal component maximize?
    Variance
  • When does Policy Iteration stop?
    When policy no longer changes.
  • If Q(s,a) values are: Left = 3 Right = 7 Up = 5 What action will policy choose?
    Right
  • What does K represent in K-means?
  • If living reward is negative, what behavior is encouraged?
    Faster finishing.
  • Which are components of an MDP?
    States, Actions, Rewards
  • What is a terminal state?
    A state where the episode ends.
    • Your experience on this site will be improved by allowing cookies.