Reinforcement Learning on the Room (test)
128Average Total Reward per EpisodeEpisodic-only
Evaluation Results
| Method | Links | |
|---|---|---|
| Episodic-onlyMemory capacity=642022.12 | 128 | |
| Both, RL, semantic-pretrainedMemory capacity=642022.12 | 123.8 | |
| Both, RL, semantic-scratchMemory capacity=642022.12 | 117.9 | |
| Both, RL, semantic-pretrainedMemory capacity=322022.12 | 116.5 | |
| Both, RL, semantic-scratchMemory capacity=322022.12 | 109.1 | |
| Semantic-onlyMemory capacity=642022.12 | 92.2 | |
| Semantic-onlyMemory capacity=322022.12 | 91.6 | |
| Episodic-onlyMemory capacity=322022.12 | 90.6 | |
| Both, RL, semantic-pretrainedMemory capacity=162022.12 | 86.8 | |
| Both, RL, semantic-scratchMemory capacity=162022.12 | 85.3 | |
| Semantic-onlyMemory capacity=162022.12 | 84.2 | |
| Both, randomMemory capacity=642022.12 | 82.6 | |
| Both, randomMemory capacity=322022.12 | 76 | |
| Semantic-onlyMemory capacity=82022.12 | 61.6 | |
| Both, RL, semantic-scratchMemory capacity=82022.12 | 60.4 | |
| Both, RL, semantic-pretrainedMemory capacity=82022.12 | 59.2 | |
| Episodic-onlyMemory capacity=162022.12 | 56.8 | |
| Both, randomMemory capacity=162022.12 | 54.6 | |
| Both, randomMemory capacity=82022.12 | 39.4 | |
| Both, RL, semantic-pretrainedMemory capacity=42022.12 | 38.2 | |
| Both, RL, semantic-scratchMemory capacity=42022.12 | 37.6 | |
| Semantic-onlyMemory capacity=42022.12 | 36.8 | |
| Episodic-onlyMemory capacity=82022.12 | 35 | |
| Both, RL, semantic-pretrainedMemory capacity=22022.12 | 25.5 | |
| Both, randomMemory capacity=42022.12 | 25 | |
| Both, RL, semantic-scratchMemory capacity=22022.12 | 23.9 | |
| Episodic-onlyMemory capacity=42022.12 | 23.6 | |
| Semantic-onlyMemory capacity=22022.12 | 22.4 | |
| Episodic-onlyMemory capacity=22022.12 | 17.4 | |
| Both, randomMemory capacity=22022.12 | 15.4 |