Reinforcement Learning and Decision Making

Explore agents, environments, rewards, Markov Decision Processes (MDP), Q-learning, and Deep RL.

En anglais Ce paquet est en anglais : il se joue quand le jeu est réglé dans cette langue.

Questions

20, en 5 parties : Reinforcement Learning Core Concepts · Markov Decision Processes (MDP) · Trick Questions · Q-Learning and Value Functions · Deep RL and Frontier Applications

Chapitre

Artificial Intelligence

Jeu

Citadel

Langue du paquet

English

Pour vous faire une idée

1. Machine learning paradigm where a software entity learns optimal behavior through trial-and-error interaction is:Reinforcement learning (RL)Rule-based expert systemsSupervised learningUnsupervised learning
2. Formal mathematical framework used to model decision making in RL environments is a:CNNDecision TreeGANMDP (Markov Decision Process)
3. Does a reinforcement learning agent require pre-labeled dataset examples of correct decisions to learn?No, RL agents learn by trial and errorNo, RL uses zero hardwareYes, RL needs labeled target pairsYes, humans label every action