Foundations
Reinforcement Learning
RLTraining a model through trial and error, rewarding good outcomes and penalizing bad ones until it learns a strategy.
In reinforcement learning an agent takes actions in some environment and receives rewards or penalties, then adjusts its behavior to collect more reward over time. There are no labeled right answers, just consequences, which makes it well suited to games, robotics, and control problems. It is also the backbone of RLHF, the step that tunes a chatbot to be helpful by rewarding responses people prefer.
Reference: Wikipedia: Reinforcement Learning
Related terms
Reinforcement Learning from Human Feedback
A training method that tunes a model to be helpful by rewarding the responses human reviewers prefer.
Supervised Learning
Training a model on labeled examples, where each input comes paired with the correct answer.
AI Alignment
The work of making AI systems pursue what people actually intend, including their values and safety.
Machine Learning
A way of building software that learns patterns from examples instead of being told exactly what to do.