Building Smart Agents with Reinforcement Learning Meeting 8

Machine Learning Reinforcement: Strategies and Challenges - Quiz M8 Dashboard / My courses Started on State Completed on Time taken Question 1 Monday, April 26, 2021, 12:07 PM Finished Monday, April 26, 2021, 12:50 PM 43 mins 42 secs Correct The effect of an action chosen from a state that depends only on that state and not on previous states is known as a property: Marked out of 5.00 Select one: a. Stochastic b. Markov c. Bellman d. Transition Question 2 In the Markov Decision Process, the selection of road segments in the shortest path problem is part of: Correct Marked out of 5.00 Select one: a. State b. Value c. Action d. Reward Question 3 Reinforcement learning is appropriate for solving the main problems in: Correct Marked out of 5.00 Select one: a. Chess game b. Object recognition c. Customer segmentation d. Weather forecasting Question 4 The following is a false statement regarding the Markov Decision Process: Correct Marked out of 5.00 Select one: a. The reward function can be expressed as an expectation function b. The total policy in a state for all rewards is 1 c. The state-action value function can be expressed in the Bellman expectation equation d. Elements of the transition probability matrix are between 0 and 1 Markov Decision Process can be defined in terms of n-tuples, which have values: Select one: a. 6 Question 5 b. 5 Correct c. 4 Marked out of 5.00 d. 3 Question 6 Reinforcement learning is a part of: Correct Marked out of 5.00 Select one: a. All right b. Markov Decision Process c. Dynamic Programming d. Machine learning Question 7 The solution of the Markov Decision Process is referred to as: Correct Marked out of 5.00 Select one: a. Reward b. Policy c. Model d. Action Question 8 The reward function in Markov Decision Process can be any number: Correct Marked out of 5.00 Select one: a. Original b. Round c. All right d. Real Question 9 Reinforcement learning can be illustrated as an interaction between: Correct Marked out of 5.00 Select one: a. Reward with next state b. All right c. Agent with its environment d. State with selected action Question 10 Markov properties in Markov Decision Process are required directly in determining: Correct Marked out of 5.00 Select one: a. State-action value function b. Selected action c. Reward function d. Transition probability matrix Reinforcement learning can be applied to problems: Select one: a. All right Question 11 b. Shortest path determination Correct c. Robot navigation Marked out of 5.00 d. Supervised learning Question 12 A set of tokens representing each possible state of the agent is referred to as: Correct Marked out of 5.00 Select one: a. Reward b. Action c. All wrong d. State Question 13 In choosing an action for a state, the following points need to be considered: Correct Marked out of 5.00 Select one: a. Selects the action that gives the state the highest value b. All right c. The state-action value at a time applies to any policy d. Maximizing future rewards Question 14 The following is an example of a problem that can be solved with reinforcement learning: Correct Marked out of 5.00 Select one: a. Self-driving car b. Conversation system c. Playing Games d. All right Question 15 Reward in the shortest path problem modeled with Markov Decision Process is characterized by: Correct Marked out of 5.00 Select one: a. Weight value at node b. Total weight value of nodes and internodes c. Weight value on the segment d. All wrong Question 16 The following are some reinforcement learning algorithms, except: Correct Marked out of 5.00 Select one: a. Hill-Climbing Bagged Ensemble Selection (HCES-Bag) b. State-Action-Reward-State-Action (SARSA) c. Q-learning d. Deep Deterministic Policy Gradient (DDPG The following is an incorrect statement regarding reinforcement learning (RL): Select one: a. RL can be used in creating a customized training system for students b. RL requires an environment to interact with repetitively Question 17 Correct c. RL can be used in robotics for industrial automation d. RL solved with Markov Decision Process requires training data to determine the transition probabilities. Marked out of 5.00 Question 18 The selection of an action at a particular state in the Markov Decision Process is determined by: Correct Marked out of 5.00 Select one: a. Reward b. Policy c. Model d. Value Question 19 Correct statement regarding reinforcement learning: Correct Marked out of 5.00 Select one: a. The actions that can be selected in each state are not unique b. For actions to be stochastic, they must fulfill Markov properties c. All right d. Input received by the agent in the form of state, action and reward Question 20 The effect of an action choice on a state is determined by: Correct Marked out of 5.00 Select one: a. Transition model b. Reward function c. Policy d. All wrong - Jump to...