/reinforcement-learning

Primary LanguageJupyter NotebookMIT LicenseMIT

RLADL Reinforcement Learning and advanced Deep Learning

Celina Hanouti & Imad Sidhoum (Equal Contribution)

This repository contains practical work of the course "Reinforcement Learning and Advanced Deep Learning" at Sorbonne Université, 2020-2021.

Contents:

  • TME 1 : Bandit algorithms (stochastic bandits, contextual bandits, ...)
  • TME 2 : Planning via Dynamic Programming (Value Iteration, Policy Iteration)
  • TME 3 : Value-Based methods (TD-lambda, Q learning, ...)
  • TME 4 : Deep value-based methodes (DQN, Prioritized Experience Replay,..)
  • TME 5 : Actor-Critic (A2C)
  • TME 6-7 : Advanced Actor-Critic (PPO)
  • TME8 : DDPG
  • TME 9 : Generative adversarial networks (GANs)
  • TME 10 : Variational auto-encoder (VAE)
  • TME 11 : Multi-agents DDPG
  • TME 12-13 : Imitation Learning (GAIL)
  • TME 14 : Curriculum Learning