AI 용어집
인공지능 완전 사전
Online Imitation Learning
Learning method where the agent continuously learns from real-time demonstrations, adjusting its behavior as new interactions occur.
Interactive Continual Learning
Continuous knowledge acquisition process where the agent constantly interacts with the environment while learning from new demonstrations.
Real-Time Demonstration
Process where the expert provides behavioral examples during the agent's active execution, enabling immediate learning.
Online Feedback
Corrective information provided to the agent during its operation, allowing for instantaneous behavioral adjustments based on demonstrations.
Continuous Data Collection
Ongoing process of acquiring new demonstrations during the system's operation, constantly enriching the training dataset.
Adaptive Knowledge Transfer
Mechanism for integrating new skills from demonstrations while preserving pre-existing knowledge.
Continuous Observational Learning
Technique where the agent learns passively by observing the expert's continuous demonstrations without interrupting its own operation.
Evolving Imitation Policy
Behavioral strategy that progressively refines itself through the continuous integration of new online demonstrations.
Online Human-Agent Interaction
Continuous dialogue between the human operator and the AI agent, where demonstrations are provided and integrated in real-time.
Continuous Behavioral Correction
Process of continuous rectification of the agent's actions based on corrective demonstrations provided continuously by the expert.
Interactive Learning by Example
Method where the agent learns through dynamically provided examples during its interaction with the environment.
Online Trajectory Optimization
Continuous adjustment of the agent's action paths based on new demonstrations of optimal trajectories.
Continuous Episodic Memory
Permanent and organized storage of new episodic demonstrations for future reference and incremental learning.
Progressive Learning by Demonstration
Approach where skills are acquired gradually through demonstrations of increasing complexity provided continuously.
Imitative Feedback Loop
Continuous cycle where the agent's action is evaluated against demonstrations, generating corrections that improve future performance.