Experiments

DQN

DQN related experiments.

The simplest example to demonstrate how to use BasicDQN

card-cover-image

BasicDQN can also be applied to MountainCar

card-cover-image

BasicDQN can also be applied to discrete Pendulum

card-cover-image

A simple example to demonstrate how to use environments in GridWorlds.jl

card-cover-image

DQN applied to CartPole

card-cover-image

PrioritizedDQN applied to CartPole

card-cover-image

DQN can also be applied to MountainCar

card-cover-image

IQN applied to CartPole

card-cover-image

QRDQN applied to CartPole

card-cover-image

REMDQN applied to CartPole

card-cover-image

Rainbow applied to CartPole

card-cover-image

The simplest example to demonstrate how to use DQN to solve atari games.

card-cover-image

Use the Rainbow to play the atari game ms_pacman.

card-cover-image

Use the IQN to play the atari game breakout.

card-cover-image

Policy Gradient

Policy gradient related experiments.

A2C applied to CartPole

card-cover-image

A2CGAE applied to CartPole

card-cover-image

DDPG applied to Pendulum

card-cover-image

MADDPG applied to KuhnPoker

card-cover-image

MADDPG applied to SpeakerListenerEnv

card-cover-image

MAC applied to CartPole

card-cover-image

PPO applied to CartPole

card-cover-image

PPO applied to Pendulum

card-cover-image

SAC applied to Pendulum

card-cover-image

TD3 applied to Pendulum

card-cover-image

VMPO applied to CartPole

card-cover-image

VPG applied to CartPole

card-cover-image

Rlpyt A2C Atari

card-cover-image

Rlpyt PPO Atari

card-cover-image

Offline

Offline RL related experiments.

Collect CartPole dataset generated by BasicDQN

card-cover-image

Collect Pendulum dataset generated by SAC

card-cover-image

BC applied to CartPole

card-cover-image

BCQ applied to Pendulum

card-cover-image

BCQD applied to CartPole

card-cover-image

BEAR applied to Pendulum

card-cover-image

CRR applied to CartPole

card-cover-image

CRR applied to Pendulum

card-cover-image

FisherBRC applied to Pendulum

card-cover-image

PLAS applied to Pendulum

card-cover-image

FQE applied to CRR policy on PendulumEnv

card-cover-image

Searching related experiments.

Minimax applied to OpenSpiel(tictactoe)

card-cover-image

CFR

Counterfactual regret related experiments.

TabularCFR applied to OpenSpiel(kuhn_poker)

card-cover-image

DeepCFR applied to OpenSpiel(leduc_poker)

card-cover-image

NFSP

Neural Fictitious Self-play(NFSP) related experiments.

NFSP applied to KuhnPokerEnv

card-cover-image

play "kuhn_poker" in OpenSpiel with NFSP

card-cover-image

ED

Exploitability Descent algorithm related experiments.

play "kuhn_poker" in OpenSpiel with Exploitability Descent(ED) algorithm.

card-cover-image