Measuring Thematic Fit with Distributional Feature Overlap
This work addresses the problem of improving thematic fit modeling in computational linguistics, offering an incremental advancement over existing unsupervised methods.
The paper tackles the problem of modeling predicate-argument thematic fit judgments by introducing a distributional method that uses syntax-based distributional semantic models to build verb-specific role prototypes and compute thematic fit as weighted feature overlap. The result is that the method consistently outperforms a baseline state-of-the-art system and achieves competitive results with other unsupervised systems, while providing explicit feature representations for semantic roles.
In this paper, we introduce a new distributional method for modeling predicate-argument thematic fit judgments. We use a syntax-based DSM to build a prototypical representation of verb-specific roles: for every verb, we extract the most salient second order contexts for each of its roles (i.e. the most salient dimensions of typical role fillers), and then we compute thematic fit as a weighted overlap between the top features of candidate fillers and role prototypes. Our experiments show that our method consistently outperforms a baseline re-implementing a state-of-the-art system, and achieves better or comparable results to those reported in the literature for the other unsupervised systems. Moreover, it provides an explicit representation of the features characterizing verb-specific semantic roles.