Representing Verbs with Rich Contexts: an Evaluation on Verb Similarity
This work addresses a specific issue in computational linguistics for improving verb similarity modeling, but it is incremental as it builds on existing distributional semantic models.
The paper tackled the problem of representing verbs with rich contexts to capture word expectations, by proposing a distributional semantic model that uses joint syntactic dependencies as contexts, and achieved comparable or better performance on verb similarity tasks across two datasets while overcoming data sparsity issues.
Several studies on sentence processing suggest that the mental lexicon keeps track of the mutual expectations between words. Current DSMs, however, represent context words as separate features, thereby loosing important information for word expectations, such as word interrelations. In this paper, we present a DSM that addresses this issue by defining verb contexts as joint syntactic dependencies. We test our representation in a verb similarity task on two datasets, showing that joint contexts achieve performances comparable to single dependencies or even better. Moreover, they are able to overcome the data sparsity problem of joint feature spaces, in spite of the limited size of our training corpus.