RiverSwim
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
RiverSwim
0.948Q^π(1, 0)
120
RiverSwim T=50 90% nominal coverage
91.3Q* (1, 0)
20
RiverSwim T=50 nominal 95% coverage
95.7Q* Recovery (s=1, a=0)
20
RiverSwim mostly-right target policy T=50
0.523V(s=1)
20
RiverSwim mostly-right target policy T=50
0.523Q-Value Estimate (s=1, a=0)
20
RiverSwim episode length T = 10 (nominal 95% coverage)
89.9Q* (1, 0)
20
RiverSwim mostly-left policy, T=50
55.7Qπ(1, 0) Coverage
20
RiverSwim
0.952Value Estimate State 1
20
RiverSwim
0.952Q-Value Estimate (s=1, a=0)
20
RiverSwim episode length T=100
94.9Q* Coverage (1, 0)
15
RiverSwim episode length T=100, nominal coverage 50%
0.529Q* (1, 0) Coverage
15
RiverSwim (T=100)
0.907V*(1)
15
RiverSwim T=100
0.907Q*(1,0)
15
RiverSwim tabular
100Returns
15
Ergodic RiverSwim
6.3Psi
6
RiverSwim
3.3Cumulative Reward
4