We consider the problem of recognizing mentions of human senses in text. Our contribution is a method for acquiring labeled data, and a learning method that is trained on this data. Experiments show the effectiveness of our proposed data labeling approach and our learning model on the task of sense recognition in text.
View on arXiv