KTS develops innovative digital services and research data infrastructures for the social sciences as a joint effort with and for other scientific GESIS departments. Taking a cross-cutting role within GESIS, the department develops infrastructures across all stages of the research data cycle, provides central and user-friendly interfaces and ensures their long-term sustainability and compliancy with state-of-the-art standards and techniques. A particular focus is given to providing integrated access to GESIS data and services.
In order to ensure quality and innovation of GESIS services, KTS conducts research in applied computer science, in particular in areas such as information retrieval, information extraction & NLP, semantic technologies and human computer interaction. Research at KTS aims at ensuring access and use of social sciences research data along the FAIR principles, for instance, through interlinking of research data, established vocabularies and knowledge graphs and by facilitating semantic search across distinct platforms and datasets. One recent challenge is the support of novel forms of social sciences research data, such as social web crawls and Web archives, which call for efficient techniques for indexing, linking and retrieval, accounting for their scale and heterogeneity. Due to the increasing importance of Web- and W3C standards as well as Web-based platforms, in addition to traditional research data portals, findability and interoperability of research data across the Web constitutes another current challenge.
The overarching goal of the department is the support of all infrastructural challenges related to all GESIS research data along the research data cycle, in order to support open science in the social sciences.
We would particularly like to highlight the following offers from our department:
The team “Information & Data Retrieval” focusses on service and research activities which deal with the efficient search for research data, especially in the social sciences. Users should be supported efficiently and effectively throughout all stages of the information search. To do so, the team conducts research in various IR areas, e.g. interactive information retrieval, entity-based and semantic search or personalization. We concentrate on vertical search in different types of information, such as research data or literature data. The information types are interlinked with each other in order to present the user relations between the data and to enable semantic or entity-related searches across all data (see GESIS-wide search). A current challenge is temporal search in large amounts of web data and digital behavioral data, important forms of data for social science research. The goal of the “Information & Data Retrieval” team is to provide suitable state-of-the-art search technologies for all infrastructures at GESIS and to offer innovative search solutions which base on research in the above mentioned areas.
The team Information Extraction & Linking investigates methods and develops services for the integration and linking of research data and research information. The core task of the team is the establishment of a Knowledge Graph infrastructure using established vocabularies and standards, in order to link GESIS research data with each other and with established vocabularies, knowledge graphs and databases. The team develops innovative methods for extracting and linking research data and information, and conducts research in information extraction and natural language processing (NLP), entity resolution and interlinking, and data fusion. Current challenges arise e.g. by mining research data and linking to/from research data references and other data on the web or linking to large web archives containing e.g. Twitter data. In addition to linking research data and its referencing publications, it also advances the linking and homogenization of established vocabularies and annotation frameworks that support a holistic view of research data along the FAIR principles and facilitate cross-search, e.g. as part of the GESIS-wide search and on the Web.
The team FAIR Data aims at advancing the usability of research data along the FAIR Principles by research-based consolidation and integration of GESIS’ digital services and by improving the visibility and findability of GESIS data on the Web. To this end, the team conducts research at the intersection of information extraction, knowledge graphs, data integration, and information retrieval, and consistently applies these to the improvement of research data and services. As part of the ongoing development of GESIS' Integrated Information Services, the team continuously collaborates with other teams at KTS and departments at GESIS. In addition, collaboration with national and international research data infrastructure initiatives (especially NFDI) is a central task, ensuring transfer to the community and connectivity to standards and innovations from the community (e.g., on FAIR assessment). The team is involved in three NFDI consortia.
The main goal of our team Data & Service Engineering is the development of a common and sustainable software architecture on the basis of a state-of-the-art technology stack for all digital services and infrastructures of GESIS. The team concentrates the software development expertise of GESIS based on service-oriented methods and develops and runs the range of research data portals offered by GESIS. Besides state-of-the-art backend infrastructures, efficient and effective user-interfaces are a main focus. In order to fulfil these tasks, we constantly monitor and evaluate uprising technologies, the lead technical development of services for the social sciences and the IT governance to support the GESIS FAIR strategy.
The team Big Data Analytics works on research and services related to large scale computing. The goal is to make the (scalable) analytics FAIR in terms of responsible data science alongside being (Findable, Accessible, Interoperable and Reproducible). Our activities range from the design of scalable infrastructure for large scale data processing to the development of analytics methods for the social sciences. The aim is to make these methods publicly available, reproducible and explainable wherever possible. The team works at the intersection of the knowledge graphs, distributed analytics and big data.
The Human Information Interaction team works at the intersection of Human Computer Interaction and Machine Learning. By applying methods of user-centered design, the team actively involves users in the development process and thus contributes to the development of digital services with convincing usability and user experience. One example is the GESIS Search, our integrated search system for different types of information. Another goal is to use machine learning to generate insights about cognitive processes and personality traits from digital behavioural data and to incorporate them in user models. In research, the team is looking at online discourse behavior and how this can be improved, what eye tracking and physiological data (e.g. EEG) reveal about cognitive processes in reading and comprehending text (e.g. Web content or survey items), and how users with more vague information needs can be better supported in their information search process.