Deep Learning Embeddings for Discontinuous Linguistic Units

TL;DR

This paper explores the use of deep learning embeddings for discontinuous linguistic units, demonstrating their superior performance over word form embeddings in coreference resolution tasks.

Contribution

It introduces the concept of embeddings for discontinuous linguistic units and empirically evaluates their effectiveness in coreference resolution.

Findings

Discontinuous unit embeddings outperform word form embeddings in coreference resolution.

Embeddings for various linguistic units can enhance NLP task performance.

The approach broadens the scope of embedding applications in NLP.

Abstract

Deep learning embeddings have been successfully used for many natural language processing problems. Embeddings are mostly computed for word forms although a number of recent papers have extended this to other linguistic units like morphemes and phrases. In this paper, we argue that learning embeddings for discontinuous linguistic units should also be considered. In an experimental evaluation on coreference resolution, we show that such embeddings perform better than word form embeddings.