Quiz dossier / Reinforcement Learning

Reinforcement Learning: Comprehensive Foundations

Comprehensive self-assessment on RL fundamentals, PPO, AlphaStar and league training, distributed RL with Ray/RLlib, MLOps, and simulation-based learning.

27 questionsEstimated 50 min0 flagged

Session status

Draft

Answered: 0/27
Score: 0/27

Question 1

1 point(s)

What formally defines a Markov Decision Process (MDP)?

RL problems are commonly framed as MDPs. Start by identifying the formal components and objective.

A tuple of states

S

, actions

A

, transition probabilities

P(s' \mid s, a)

, a reward function

R

, and a discount factor

\gamma

A labelled dataset of state-action pairs with ground-truth target actions

A directed acyclic graph of neural-network layers and their weights

A fixed lookup table mapping every state to one deterministic action