Replay Buffer

Definitions

1. A stored collection of past experiences or transitions used in reinforcement learning so training can reuse prior data instead of relying only on the newest samples. Replay buffers can improve sample efficiency and stabilize learning.

   In plain English: A stored pool of past training experiences reused during learning.

   Example: "The agent's training loop drew batches from a replay buffer so unusual but important failures were not immediately forgotten."

Related Terms

• Chinchilla Scaling
• Constitutional AI Method
• Context Distillation
• Expert Iteration
• Frozen Embedding
• Goal Conditioned
• Gradient-Free Optimization
• Machine Unlearning
• On-Policy
• Trajectory