Google Research Football
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Adversarial Attack | Google Research Football 3 vs 1 | Reward5.43 | 12 | |
| Adversarial Attack | Google Research Football counterattack | Reward5.21 | 12 | |
| Multi-Agent Reinforcement Learning | Google Research Football CA_hard (test) | Win Rate62.6 | 7 | |
| Multi-Agent Reinforcement Learning | Google Research Football CA_easy (test) | Win Rate72.1 | 7 | |
| Multi-Agent Reinforcement Learning | Google Research Football 3_vs_1WK (test) | Win Rate94.5 | 7 | |
| Multi-agent reinforcement learning | Google Research Football (GRF) | Win Rate: Pass and Shoot97 | 7 | |
| Multi-Agent Reinforcement Learning | Google Research Football (GRF) Corner (test) | Mean Success Rate36 | 6 | |
| Multi-Agent Reinforcement Learning | Google Research Football (GRF) CA-Hard (test) | Mean Success Rate0.83 | 6 | |
| Multi-Agent Reinforcement Learning | Google Research Football (GRF) RPS (test) | Mean Success Rate75 | 6 | |
| Multi-agent reinforcement learning | Google Research Football Academy scenarios | Pass & Shoot Success Rate100 | 5 | |
| Multi-agent coordination | Google Research Football (GRF) 3-vs-1 with keeper | Goal Rate2.5 | 5 | |
| Multi-Agent Reinforcement Learning | Google Research Football academy_pass_and_shoot_with_keeper | Parameters (K)430.1 | 5 | |
| Multi-Agent Reinforcement Learning | Google Research Football academy_counterattack_easy | Parameters (K)430.36 | 5 | |
| Multi-Agent Reinforcement Learning | Google Research Football academy_3_vs_1_with_keeper | Params (K)430.36 | 5 | |
| Multi-Agent Reinforcement Learning | Google Research Football 3vs1 | Goal Rate100 | 3 | |
| Transfer Learning in MARL | Google Research Football 3vs1 with keeper to 5vs5 Full game | Environment Steps0.4 | 2 | |
| Transfer Learning in MARL | Google Research Football 5vs5 to 11vs11 | Environment Steps0.77 | 2 |
Showing 17 of 17 rows