Graph Neural Networks for Knowledge Base QA, Daniil Sorokin (TU Darmstadt)

Title: Graph Neural Networks for Knowledge Base Question Answering

Abstract:
Knowledge base question answering aims to provide a natural language interface to factual knowledge. It requires precise modeling of the question meaning through the entities and relations available in the knowledge base in order to retrieve the correct answer.
In this talk, we present a semantic parsing approach to knowledge base question answering. We address the problem of processing semantic structures that consist of multiple entities and relations. Previous work has largely focused on selecting the correct semantic relations for the question and disregarded the connections between entities and the directions of the relations. We propose to use Gated Graph Neural Networks to encode the semantic structures.
We will present a variant of Gated Graph Neural Networks for labeled knowledge base subgraphs and show how it can be used in a question answering pipeline. Empirically, we demonstrate that the graph networks lead to an improved performance against the baseline models that do not explicitly model the semantic structure.

Bio:
Daniil Sorokin is a final year PhD student in Natural Language Processing at the UKP Lab, Technische Universität Darmstadt. He works on linking texts to knowledge bases and has recently published on the topics of relation extraction and semantic parsing for question answering.
Before the PhD studies, Daniil had completed a master degree in Computational Linguistics at the University of Tübingen and had worked as a developer at a machine translation company.