Geo-analytic Question Answering: Going Beyond Knowledge Retrieval

E. Nyamsuren, S. Scheider, H. Xu, J.F. Kruiger

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademic

Abstract

We present ongoing work in the ERC funded project Questionbased analysis of Geographic Information with Semantic Queries (QuAnGIS). The aim of the project is to develop a question answering system that is able to address geo-analytic questions. Such automation should make geographic information systems (GIS) not only more efficient but also more accessible to non-expert users. Current question-answering approaches are based on the retrieval of facts from knowledge sources. Such retrieval is not sufficient to address geo-analytic questions that require application of analytic tools on data sources to generate desired answers. Therefore, a research focus is how to identify and retrieve relevant tools and data sources for a given geo-analytic question. Our envisioned approach is to deconstruct a question query to its semantic constituents that can be mapped to semantic annotations of tools and data sources. In this work, we present preliminary results of our pilot study on analysis of the structure of geo-analytic questions and their semantic constituents. We also describe a working draft of a semantic schema for decomposing questions into GIS related concepts.
Original languageEnglish
Title of host publication 13th Workshop on Geographic Information Retrieval (GIR'19)
Publication statusPublished - 28 Nov 2019

Fingerprint

Dive into the research topics of 'Geo-analytic Question Answering: Going Beyond Knowledge Retrieval'. Together they form a unique fingerprint.

Cite this