GECKO: A Question Answering System for Official Statistics

Lucas Lageweg, Jonas Kouwenhoven, Benno Kruit

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

This paper presents GECKO, a knowledge graph-based statistical question answering system currently in beta deployment.GECKO aims to facilitate the retrieval of single statistical values from an extensive database containing over a billion values across more than 4,000 tables.The system integrates a comprehensive framework including data augmentation, entity retrieval, and large language model (LLM)-based query generation.A key feature of the beta deployment is the collection of user feedback, which is critical for improving system performance and accuracy.This feedback mechanism allows users to report issues directly, ensuring continuous improvement based on real-world use.

Original languageEnglish
Title of host publicationSEMANTiCS-PDWT 2024 Posters, Demos, Workshops, and Tutorials of the 20th International Conference on Semantic Systems (SEMANTiCS 2024)
Subtitle of host publicationJoint Proceedings of Posters, Demos, Workshops, and Tutorials of the 20th International Conference on Semantic Systems co-located with 20th International Conference on Semantic Systems (SEMANTiCS 2024) Amsterdam, The Netherlands, September 17-19, 2024
EditorsDaniel Garijo, Anna Lisa Gentile, Anelia Kurteva, Andrea Mannocci, Francesco Osborne, Sahar Vahdati
PublisherCEUR-WS.org
Number of pages5
Publication statusPublished - 2024
EventJoint of Posters, Demos, Workshops, and Tutorials of the 20th International Conference on Semantic Systems, SEMANTiCS-PDWT 2024 - Amsterdam, Netherlands
Duration: 17 Sept 202419 Sept 2024

Publication series

NameCEUR Workshop Proceedings
Volume3759
ISSN (Electronic)1613-0073

Conference

ConferenceJoint of Posters, Demos, Workshops, and Tutorials of the 20th International Conference on Semantic Systems, SEMANTiCS-PDWT 2024
Country/TerritoryNetherlands
CityAmsterdam
Period17/09/2419/09/24

Bibliographical note

Publisher Copyright:
© 2024 Copyright for this paper by its authors.

Fingerprint

Dive into the research topics of 'GECKO: A Question Answering System for Official Statistics'. Together they form a unique fingerprint.

Cite this