We test whether distributional models can do one-shot learning of definitional properties from text only. Using Bayesian models, we find that first learning overarching structure in the known data, regularities in textual contexts and in properties, helps one-shot learning, and that individual context items can be highly informative. Our experiments show that our model can learn properties from a single exposure when given an informative utterance.

PDF

In In Proceedings of the 8th International Joint Conference on Natural Language Processing (IJCNLP-17), Taipei, Taiwan, November 2017.

@inproceedings{wang:ijcnlp17, title={Distributional modeling on a diet: One-shot word learning from text only}, author={Su Wang and Stephen Roller and Katrin Erk}, booktitle={In Proceedings of the 8th International Joint Conference on Natural Language Processing (IJCNLP-17)}, month={November}, address={Taipei, Taiwan}, url="http://www.cs.utexas.edu/users/ai-labpub-view.php?PubID=127668", year={2017} }