List your company

ResearchFormerly WizWand

Walker2D

Benchmarks

TaskDatasetSOTA resultTrendResults

Goal-conditioned Reinforcement Learning (G3_efficient)

walker2d medium v2

0.374Goal-conditioned Return

5

Goal-conditioned Reinforcement Learning (G2_balanced)

walker2d medium v2

1.863Goal-conditioned Return

5

Goal-conditioned Reinforcement Learning (G1_speed)

walker2d medium v2

2.904Goal-conditioned Return

5

Reinforcement Learning

Walker2d

4,595.3Max Return

5

Continuous Control

Walker2d v4

98Number of Interactions (10^4 steps)

5

Continuous Control

Walker2d v4

4,106.21Average Return

5

Multi-task locomotion

walker2d

1,057.9Score (Forward, medium-replay)

5

Reinforcement Learning

Walker2d discrete-regime exponential dwell

374Mean Return

4

walker2d expert v2

115.1Normalized Return

4

Dynamics Prediction

Walker2d

16.35MAE

4

Offline Reinforcement Learning

walker2d kine-e

95.8Normalized Score

4

Offline Reinforcement Learning

walker2d kine-me

69.2Normalized Score

4

Offline Reinforcement Learning

walker2d kine-m

65.1Normalized Score

4

Reinforcement Learning

Walker2d 1.5 density (target domain)

3,383Reward

4

Reinforcement Learning

Walker2d 0.5 gravity (target)

2,830Reward

4

Reinforcement Learning

Walker2d 1.5 gravity MuJoCo (test)

4,409Reward

4

Reinforcement Learning

Walker2d MuJoCo (broken source)

4,366Reward

4

Reinforcement Learning

Walker2D MuJoCo (unseen states)

3,200Reward

4

Continuous Control

Walker2d (test)

2.96Best Test Reward

3

Walker2d non-stationary MuJoCo

363Final Reward

3

Robot Co-design

Walker2D search phase

1.36D2C Score

3

Offline Reinforcement Learning

walker2d medium-replay v2

375Return

3

Noisy Action Robust MDP

Walker2D high randomness MuJoCo Noisy Action Robust MDP

502.6Worst-case Return

3

Noisy Action Robust MDP

Walker2D medium randomness MuJoCo Noisy Action Robust MDP

508.4Worst-case Return

3

Noisy Action Robust MDP

Walker2D low randomness MuJoCo Noisy Action Robust MDP

429.6Worst-case Return

3