Actors

• Off-Policy Actor
• On-Policy Actor
• Model-Based Actor

Actor Components

• Action prob. distributions
  • Categorical
  • Gaussian
  • Squashed Gaussian
• Feature Extractors
  • Multilayer Perceptron (MLP)
  • Convolutional Neural Network (CNN)
  • Residual CNN with Fixup initialization (Fixup)
  • Multimodal Neural Network (DictNet)
• Memory Networks
  • Gated Recurrent Unit (GRU)
• Noise
  • Gaussian Noise
  • Ornstein-Uhlenbeck Noise
• Reward Functions
  • Gym CartPole
  • Gym Pendulum
  • MuJoCO Inverted Pendulum
  • MuJoCo HalfCheetah
  • PyBullet HalfCheetah
• World Models
  • WorldModel