Playing Atari with few neurons: Improving the efficacy of reinforcement learning by decoupling feature extraction and decision making

Computer Science

• Feature Extraction 100%
• Neural Network 100%
• Reinforcement Learning 100%
• Decision-Making 100%
• Approximation (Algorithm) 50%
• Image Processing 50%
• Sparsity 50%
• Online Learning 50%
• Vector Quantization 50%
• Reconstruction Error 50%
• Evolution Strategy 50%
• Decision Information 50%
• Deep Neural Network 50%
• Deep Reinforcement Learning 50%
• Natural Evolution 50%
• Dictionary Vector 50%
• Dictionary Size 50%

Keyphrases

• Feature Decoupling 100%
• Direct Policy Search 33%