This is a brief summary of paper for me to study and arrange it, Jointly Learning Word Representations and Composition Functions Using Predicate-Argument Structures (Hashimoto et al., EMNLP 2014) I read and studied.
They focus on the dependecy embedding with predictate-arguemtne structure as follows:
In order for them to resolve the word disambiguation, they used POS information by combine word with POS taggging such as cause_NN.
Note(Abstract):
They introduce a novel compositional language model that works on PredicateArgument Structures (PASs). their model jointly learns word representations and heir composition functions using bagof-words and dependency-based contexts. Unlike previous word-sequencebased models, their PAS-based model composes arguments into predicates by using he category information from the PAS. his enables our model to capture longrange dependencies between words and o better handle constructs such as verbobject and subject-verb-object relations.