Tutorial 1 Recap Quiz

JavaScript isn't enabled in your browser, so this file can't be opened. Enable and reload.

How frequently does the agent receive rewards (including when the reward is 0) in MDPs? *

1 point

Never

Once ever

Alternate timesteps

Every timestep

What are states and actions in Reinforcement Learning? *

1 point

States are chosen by the agent to interact with the environment. Actions are given to the agent at each timestep, and can be thought of as a feedback signal, telling it which states are good and bad.

States fully define the state of the environment. Actions are chosen by the agent to interact with the environment.

States are given as feedback by the environment each timestep, and can be thought of as a feedback signal. Actions are taken by the agent to interact with the environment.

States are chosen by the agent to interact with the environment. Actions are what you train in reinforcement learning, trying to learn the optimal action.

What is the name of the type of problems that Reinforcement Learning solves? *

1 point

Markov Reward Processes

Markov Directed Problems

Markov Decision Processes

Markov Directed Processes

What is a policy in reinforcement learning? *

1 point

The way that rewards are given out by the environment.

The function that determines the successor state given the current state and action taken. It can also be called the transition function.

A function that chooses an action given the state as input

A set of rules followed by airline company staff which make no logical sense.

What's the goal of reinforcement learning algorithms? *

1 point

Finding the optimal path through the MDP

Finding the maximum reward received at any timestep

Finding the optimal policy - the policy which maximises the cumulative future reward

Finding the best MDP (state transitions & rewards to give) to represent the task to solve

Submit

Clear form

This form was created inside of Learney. Report Abuse

Forms