TABLE 3.
Performance comparison of training strategies on the testing dataset.
Model | Average on test dataset | Median on test dataset | Trained episodes | Iterations | Population |
Seed-6 after initialization | 23.38 | 19.5 | 0 | – | |
Seed-6 after STDP-RL training | 144.67 | 130.5 | 4226 | – | |
Seed-6 after EVOL training | 499.42 | 500.0 | 80,000 | 1,600 | 10 |
Seed-3 after initialization | 23.04 | 19.0 | 0 | – | |
Seed-3 after STDP-RL training | 64.14 | 53.0 | 8,577 | – | |
Seed-3 after EVOL training | 499.09 | 500.0 | 80,000 | 1,600 | 10 |