machine learning/deep learning/reinforcement learning/autoML
Up to now, implementing Q-learning\DQN\DDQN\Actor-Critic\DDPG.
Further more, approving a general theory, The Worst is the Best, which improved DQN\DDQN\Actor-Critic on openai gym environment.
This theory is for NIPS-2018, and the code is under the directory 'ML-DL-RL-AutoML/RL/RL-extend-for-NIPS2018-paper'.
Up to now, implementing traditional bayesian and gaussion process regressor by sklearn and skopt