Lexical Coherence Graph Modeling Using Word Embeddings, Mesgar+, NAACL'16

[AkihikoWatanabe]

https://aclanthology.org/N16-1167/

[AkihikoWatanabe]

__translate: Coherence is established by semantic connections between sentences of a text which can be modeled by lexical relations. In this paper, we introduce the lexical coherence graph (LCG), a new graph-based model to represent lexical relations among sentences. The frequency of subgraphs (coherence patterns) of this graph captures the connectivity style of sentence nodes in this graph. The coherence of a text is encoded by a vector of these frequencies. We evaluate the LCG model on the readability ranking task. The results of the experiments show that the LCG model obtains higher accuracy than state-of-the-art coherence models. Using larger subgraphs yields higher accuracy, because they capture more structural information. However, larger subgraphs can be sparse. We adapt Kneser-Ney smoothing to smooth subgraphs’ frequencies. Smoothing improves performance.