Machine Learning Reinforcement Learning 2

Question 1

What is Q-learning in reinforcement learning?

Accepted Answer

A model-free RL algorithm that learns action-value (Q) functions

Answer

A clustering algorithm adapted for sequential environments

Answer

A method for dimensionality reduction in state spaces

Answer

A supervised learning technique applied to sequential prediction

Question 2

What does the Bellman equation capture in reinforcement learning?

Accepted Answer

The recursive relationship between the value of a state and its successor states

Answer

The relationship between learning rate and convergence speed

Answer

The optimal policy expressed as state transition probabilities

Answer

The formula for calculating policy entropy

Question 3

What is temporal difference (TD) learning?

Accepted Answer

Learning from differences between consecutive time steps using bootstrapping before episode completion

Answer

A method for predicting future timestamps in time-series forecasting

Answer

Training recurrent networks across multiple time periods simultaneously

Answer

Measuring the delay between taking an action and receiving a reward

Question 4

What is the key distinction between model-based and model-free reinforcement learning?

Accepted Answer

Model-based learns an environment model to plan ahead; model-free learns directly from interaction

Answer

Model-based uses neural networks while model-free uses lookup tables

Answer

Model-based requires more data while model-free requires less data to converge

Answer

Model-based applies to discrete spaces while model-free applies to continuous spaces

Question 5

In the SARSA algorithm, what does each letter represent?

Accepted Answer

State, Action, Reward, State, Action

Answer

State, Action, Reward, Space, Algorithm

Answer

Sample, Average, Reward, Step, Agent

Answer

State, Agent, Return, Space, Action

Question 6

What is the purpose of the discount factor (γ) in reinforcement learning?

Accepted Answer

To determine how much future rewards are valued relative to immediate rewards

Answer

To control the learning rate during gradient updates

Answer

To reduce the exploration rate over time during training

Answer

To normalize reward values to fall between 0 and 1

Question 7

What is a Markov Decision Process (MDP) in reinforcement learning?

Accepted Answer

A mathematical framework for modeling sequential decision-making with states, actions, transitions, and rewards

Answer

A neural network architecture designed for sequential decision problems

Answer

A clustering algorithm for grouping similar time-series states

Answer

A method for handling missing or delayed reward signals