Hermann Kroll

Dr. Hermann Kroll

Computer Scientist

GitHub
YouTube
Google Scholar
ORCID
DBLP

Description

I am a Postdoc at the Institute for Information Systems, TU Braunschweig led by Wolf-Tilo Balke.


I am interested in research that boosts literature retrieval and search in digital libraries. My works focus on bridging the gap between unstructured information (natural language text) and structured representations (document graphs, knowledge graphs, etc.). Therefore, I proposed narrative information access, which allows users to formulate information needs as graph patterns while preserving result validity through context-compatible information fusion. My research interests include narratives, information extraction, entity linking, text classification, and query processing. My research aims to allow more sophisticated retrieval by adding structure to digital library collections.



PC Chair: JCDL Panels [2024] (Joint Conference on Digital Libraries)


PC Member: JCDL Poster Track [2024] (Joint Conference on Digital Libraries). TPDL [2024] (International Conference on Theory and Practice of Digital Libraries). ECIR [2024] (European Conference on Information Retrieval), DISCO [2021] (Digital Infrastructures for Scholarly Content Objects@JCDL). LWDA [2020] (Lernen. Wissen. Daten. Analysen.)

Journal Reviewer: TKDE (Transaction on Knowledge and Data Engineering), JODL (International Journal on Digital Libraries), IPM (International Journal on Information Processing and Management)

Projects

I am working for the specialized information service for Pharmacy (www.pubpharm.de). The project's overall goal is to support pharmaceutical literature search for the German academic landscape. I have been developing the Narrative Service for PubPharm. The Narrative Service allows users to formulate their information need as structured graph patterns. The service returns publications that match the corresponding pattern, i.e., they contain the desired information. Therefore, we transformed the literature into so-called document graphs through entity linking and information extraction. One special feature of the service is the support of queries with variables (e.g., which drugs to treat Diabetes Mellitus). To allow eased access, users can now formulate their information needs as keywords, and the service proposes possible patterns automatically. In addition to the Narrative Service, I have been developing the Drug Overview service which generates a structured literature overview about a drug by combining information from specialized databases and information extracted from research literature. The information is derived by the following steps: a set of narrative queries with variables is executed, the results are aggregated, and shown to the user. A click on some information will forward the user to corresponding text passages/sources to show from where the information was derived.

Publications


2024

2023


2022


2021


2020


2019


2018



Recorded Talks



Teaching


Lectures/Seminars

# Year Semester Course Name Role
1 2024 Winter Multimedia Databases Lecturer
1 2024 Summer Relational Database Systems 2 Lecturer
1 2023 Winter Relational Database Systems 1 Lecturer
2 2023 Summer Relational Database Systems 2 Assistant
3 2022 Winter Relational Database Systems 1 Assistant
4 2022 Summer Multimedia Databases Assistant
5 2021 Winter Seminar: Narrative Information Access Assistant
6 2020 Winter Deductive Databases and Knowledge-based Systems Assistant
7 2019 Winter Relational Database Systems 1 Assistant
8 2019 Summer Relational Database Systems 2 Assistant
9 2018 Winter Relational Database Systems 1 Assistant


Student Projects

# Year Course Name Type Role
1 2023 A pharmaceutical PDF table extraction pipeline Teamproject Supervisor
2 2022 Graph-based Exploration of Long COVID Literature Teamproject Supervisor
3 2021 Narrative Information Access for Pharmacy (Drug Overviews) Teamproject Supervisor
4 2021 SherloQL 2.0 Teamproject Supervisor
5 2020 SherloQL Software Development Project Supervisor
6 2019 A Narrative Query Designer Teamproject Supervisor
7 2019 Music Tinder Software Development Project Supervisor