From: Reliable knowledge graph fact prediction via reinforcement learning
Target relation | Model | Rule count | Length = 1 | Length = 2 | Length = 3 | Length > 3 |
---|---|---|---|---|---|---|
Athlete_plays_in_league | DeepPath | 24 | 0 | 3 | 6 | 15 |
EvoPath | 31 | 1 | 2 | 5 | 23 | |
Athlete_plays_sport | DeepPath | 21 | 0 | 3 | 4 | 14 |
EvoPath | 40 | 0 | 3 | 6 | 31 | |
filmWrittenBy | DeepPath | 4 | 4 | 0 | 0 | 0 |
EvoPath | 37 | 6 | 3 | 2 | 26 |