InverseRL-Pytorch

Agent Algorithm

• PPO
• SAC

Discriminator Algorithm

• GAIL(Generative Adversarial Imitation Learning)
• VAIL(Variational Adversarial Imitation Learning)
• SQIL(Imitation Learning via Reinforcement Learning with Sparse Rewards)
• AIRL(Adversarial Inverse Reinforcement Learning)
  • Two value functions can be merged into one.
  • Extremely unstable
• EAIRL(Empowerment-regularized Adversarial Inverse Reinforcement Learning)
  • Two value functions can be merged into one.
  • Extremely unstable
• VAIRL(Variational Adversarial Inverse Reinforcement Learning)
  • Joint gaussian distribution kl-divergence yet.

TODO

• add more environments(ant and disabled ant)
• build setup file
• make expert
• make trajectories by expert