Definition Frames: Using Definitions for Hybrid Concept Representations
This addresses the need for more interpretable word representations in NLP, though it appears incremental as it builds on existing distributional semantic approaches.
The paper tackled the problem of word representations lacking semantic interpretability by introducing Definition Frames, a matrix representation extracted from definitions with semantically interpretable dimensions based on Qualia structure relations, achieving competitive performance on word similarity tasks.
Advances in word representations have shown tremendous improvements in downstream NLP tasks, but lack semantic interpretability. In this paper, we introduce Definition Frames (DF), a matrix distributed representation extracted from definitions, where each dimension is semantically interpretable. DF dimensions correspond to the Qualia structure relations: a set of relations that uniquely define a term. Our results show that DFs have competitive performance with other distributional semantic approaches on word similarity tasks.