YZ Sözlüğü
Yapay Zekanın tam sözlüğü
Behavioral Cloning
Direct policy learning by minimizing the error between agent actions and expert demonstrations
Inverse Reinforcement Learning
Inferring the reward function from expert demonstrations to then learn the optimal policy.
Generative Adversarial Imitation Learning
Using adversarial networks to distinguish agent behaviors from expert demonstrations
DAgger Data Aggregation
Iterative data collection by querying the expert on states visited by the current policy
Offline Imitation Learning
Learning from a fixed set of demonstrations without additional interaction with the environment.
Apprentissage par Imitation en Ligne
Apprentissage continu avec interaction en temps réel et mises à jour basées sur les nouvelles démonstrations.
Observation-based Imitation
Learning by observing only states and trajectories without having access to expert actions.
Apprentissage par Imitation Hiérarchique
Décomposition des tâches complexes en sous-tâches avec apprentissage par imitation à différents niveaux d'abstraction.
One-Shot Imitation Learning
Ability to imitate a new task after observing a single demonstration.
Meta-Learning by Imitation
Learning to quickly learn new tasks by imitation through experience on multiple tasks.
Multimodal Imitation Learning
Handling demonstrations with multiple valid solutions and learning multimodal policies.
Imitation with Partial Observations
Imitation learning when demonstrations only partially cover the state space.