JAIR Analogical Mapping (State of the art)
- analogical mapping between two sets of words
- 20 mapping problems: 10 science analogies, 10 common metaphors
- typically about 7 words per set
- available from JAIR
- task described here
Sample question
Problem: Map {planet, attracts, revolves, sun, gravity, solar system, mass} to {revolves, atom, attracts, electromagnetism, nucleus, charge, electron}
Solution: {solar system -> atom, sun -> nucleus, planet -> electron, mass -> charge, attracts -> attracts, revolves -> revolves, gravity -> electromagnetism}
Table of results
| Algorithm | Reference for algorithm | Reference for experiment | Type | Correct |
|---|---|---|---|---|
| LIN+POS | Lin (1998) | Turney (2008) | WordNet-based | 76.8% |
| human | human | Turney (2008) | human | 87.6% |
| LRME | Turney (2008) | Turney (2008) | corpus-based | 91.5% |
References
Lin, D. (1998). An information-theoretic definition of similarity. Proceedings of the 15th International Conference on Machine Learning (ICML-98), Madison, WI, pp. 296-304.
Turney, P.D. (2008), The latent relation mapping engine: Algorithm and experiments, Journal of Artificial Intelligence Research (JAIR), 33, 615-655.