Towards Compositional Tree Kernels
Paolo Annesi, Danilo Croce, Roberto Basili · Pisa University Press eBooks · 2014
Distributional Compositional Semantics (DCS) methods combine lexical vectors according to algebraic operators or func-tions to model the meaning of complex lin-guistic phrases. On the other hand, several textual inference tasks rely on supervised kernel-based learning, whereas Tree Ker-nels (TK) have been shown suitable to the modeling of syntactic and semantic simi-larity between linguistic instances. While the modeling of DCS for complex phrases is still an open research issue, TKs do not account for compositionality. In this pa-per, a novel kernel called Compositionally