Disentangling Sources of Risk for Distributional Multi-Agent Reinforcement Learning

Benchmark Model Rank Results
smac-on-smac-def-armored-parallelDRIMAMedian Win Rate: 60.0
smac-on-smac-def-armored-sequentialDRIMAMedian Win Rate: 100
smac-on-smac-def-infantry-parallelDRIMAMedian Win Rate: 100.0
smac-on-smac-def-infantry-sequentialDRIMAMedian Win Rate: 100
smac-on-smac-def-outnumbered-parallelDRIMAMedian Win Rate: 70.0
smac-on-smac-def-outnumbered-sequentialDRIMAMedian Win Rate: 100
smac-on-smac-off-complicated-parallelDRIMAMedian Win Rate: 100
smac-on-smac-off-complicated-sequentialDRIMAMedian Win Rate: 96.9
smac-on-smac-off-distant-parallelDRIMAMedian Win Rate: 95.0
smac-on-smac-off-distant-sequentialDRIMAMedian Win Rate: 100
smac-on-smac-off-hard-parallelDRIMAMedian Win Rate: 80.0
smac-on-smac-off-hard-sequentialDRIMAMedian Win Rate: 93.8
smac-on-smac-off-near-parallelDRIMAMedian Win Rate: 95.0
smac-on-smac-off-near-sequentialDRIMAMedian Win Rate: 93.8
smac-on-smac-off-superhard-parallelDRIMAMedian Win Rate: 0.0
smac-on-smac-off-superhard-sequentialDRIMAMedian Win Rate: 15.6