VADIS -- a VAriable Detection, Interlinking and Summarization system

TL;DR

VADIS tackles the challenge of making survey variables and their links to publications and datasets readily discoverable in the social sciences. It combines data preprocessing, extractive/abstractive cross-lingual summarization, and an SV identification component within an API-driven architecture that indexes results in ElasticSearch and exposes them through a web UI. The work provides a demonstrator, exploration capabilities, and a test bed for user studies, with a corpus of 607 publications and 7086 sentence-variable linkages. The approach promises improved reproducibility and data-to-publication traceability by enabling precise search, contextual variable inspection, and integrated summaries across English and German sources.

Abstract

The VADIS system addresses the demand of providing enhanced information access in the domain of the social sciences. This is achieved by allowing users to search and use survey variables in context of their underlying research data and scholarly publications which have been interlinked with each other.

Figures (2)

• Figure 1: VADIS Framework
• Figure 2: VADIS User Interface