Deep Reinforcement Learning

Combination of reinforcement learning with deep neural networks — from Atari game agents to AlphaTensor's algorithm discovery.

Specifications & Metrics

type

Learning Methods & Systems

Core Principles & Analysis

• A subfield of machine learning that combines reinforcement learning (RL) and deep learning.
• It lets RL scale to high-dimensional perceptual inputs such as images and raw text.
• DeepMind's Atari-playing agents were its landmark early demonstrations.

Sources

• Wikipedia: Deep reinforcement learning