Select runs that logged average_score
to visualize data in this line chart.
to visualize data in this line chart.
Complete Mountain Car Online Experiments
Created on April 23|Last edited on September 10
Comment
Experiment 1: Regular DQN, Large Catastrophe Zone, Vary Penalty Values
Experiment 1A: Penalty Value = -4
Run set
0
Experiment 1B: Penalty Value = -6
Experiment 1C: Penalty Value = -8
Experiment 2: Duplicating Interventions, Large Catastrophe Zone
Experiment 3: Experiment 1 with Longer Period of Training
Experiment 4: Experiment 2 Sweep with Penalty = -4
Experiment 5: Parameter that Controls Proportion of Human Actions
Experiment 6: Appending Interventions
Experiment 7: Learned Blocker
Experiment 8: Replacing Robot Interventions with Human Actions
Experiment 8A: Learned Blocker; 50% Probability of Replacement
Regular
0
Control
0
Experiment 8B: Perfect (Algo) Blocker
Regular
0
Control
0
Experiment 9: Learned Blocker (with only State)
Add a comment