Michael Färber/en

Dr. Michael Färber
Substitute Professor
Phone: +49 721 608 465 92
Email: michael faerber∂kit edu
Room: 5A-20 (Building: 05.20)

Research Group: Web Science
Consultation hour by appointment.
Office: Nicoletta Joanni
vCard

Michael Färber is the deputy professor (W3) of the research group Web Science at the KIT-institute AIFB since October 2020. He is also a Helmholtz AI Associate.

News:

April 27, 2023: Our workshop at the Girls' Day with reserved GPU Cluster: AI & Language Models
Feb -- Mar, 2023: Research stay at the NII (Tokyo)
Oct 27, 2022: Best Poster award (300 USD) for The Green AI Ontology at ISWC'22.
Aug 01 -- Sep 30, 2022: Research stay/guest professorship at the Digital Science Center (DiSC), University of Innsbruck.
Jul 12, 2022: NAACL Paper: Few-Shot Document-Level Relation Extraction

Michael Färber on

his personal website (incl. CV and list of publications),
Google Scholar,
DBLP,
Wikidata (Scholia, Reasonator),
ORCID,
ResearchGate,
Semantic Scholar,
YouTube.

Research

Profile: As W3 Deputy Full Professor until 2025, I am the head of the research group Web Science at the Karlsruhe Institute of Technology (KIT). Together with my team of seven PhD students and one postdoc, I work on the development and application of artificial intelligence (AI) methods. Specifically, my focus is in the triangle of knowledge representation, machine learning, and natural language processing. Since my postdoc phase, I have made it my mission to find solutions to the increasingly urgent problem of the rising flood of information in science (see "tsunami of publications") and to develop novel methods of communication in science. To this end, I conduct research on the extraction and modeling of scientific knowledge, in particular on the construction of large knowledge graphs. I am also developing search and recommendation systems that can exploit the explicitly modeled knowledge while explaining the results and recommendations to the user. I have published more than 80 papers in prestigious international conferences (e.g., CIKM, ISWC, ECIR, NAACL) with international researchers. In addition, I am leading several projects (e.g. TruthfulLM, ChemKB, IIDI, KIGLIS, KIWI) as Principal Investigator (PI) at the KIT Institute AIFB.

➜ publications

Research Interests

natural language processing,
machine learning, and
knowledge representation (e.g., knowledge graphs).

Among other things, Michael Färber pursues research on scholarly data mining (e.g., scientific impact quantification), scholarly recommender systems (e.g., recommending citations, papers, data sets, and neural networks), and scholarly knowledge graphs (e.g., modeling papers, authors, methods, and datasets). Furthermore, he develops AI solutions for peace mediation (see AI4Peace).

More information can be found at his homepage and on Google Scholar.

Online Demo Systems

Recently developed demonstration systems:

RefBee: http://refbee.org/
- ...shows for an author which publications are in which bibliographic databases.
C-Rex: http://c-rex.org
- ... recommends citations for given texts.
PaperHunter: http://paperhunter.net
- ... provides, among other things, the sentences in which searched papers are cited.
ScholarSight: http://scholarsight.org
- ... allows the exploration of trends from scientific concepts.
Linked Crunchbase: http://linked-crunchbase.org
- ... allows to query information about startups and innovative companies in the Semantic Web format RDF.

Data Sets

Recently created data sets:

DSKG: http://dskg.org
- ...a knowledge graph representing datasets.
unarXive: http://unarxive.org
- ... contains the full texts of all papers on arXive.org with further annotations.
Microsoft Academic Knowledge Graph: https://makg.org
- ... a knowledge graph containing the metadata of almost all publications in all scientific disciplines.
FAIRnets: https://doi.org/10.5281/zenodo.3885249
- ...a knowledge graph with metadata about neural networks.

Code, Data, and Presentations

Code: GitHub
Data: Zenodo
Presentations: YouTube

Open Positions and Theses

Open student assistant jobs (Hiwis)

in machine learning, natural language processing, or knowledge graphs: [1]
in Semantic MediaWiki or PHP: [2].

Current calls for Bachelor/Master thesis

	Titel
Thema5060	Transforming Academic Content Creation: Exploring the Power of Language Models
Thema4864	Quantum Computing for Natural Language Processing
Thema4977	Knowledge Graphs for Robots’ Situational Awareness
Thema4939	Chronik 2050: Automatische Extraktion von erwarteten Ereignissen aus Webseiten
Thema4420	Wie fair sind Forscher? Eine Analyse von Zerrungen bzgl. Zitaten in wissenschaftlichen Publikationen
Thema4909	Scalable Graph Neural Networks on Knowledge Graphs
Thema4910	Performance Analysis of Graph Neural Diffusion via Fourier Decomposition
Thema4648	Creating a Large Knowledge Graph about Scientific Publications for Innovation Forecast
Thema4574	Deep Learning + Knowledge Graphs
Thema4772	GPT-3, BERT & Co.: When to use which language model?
Thema4423	Automatically Recommending Citations for Texts Using Neural Networks

I have supervised more than 50 Bachelor/Master theses (➜ List of all completed theses directly supervised).
All topics are open to English and German-speaking students.

Theses Abroad

Many of the thesis topics can also be written at a partner institution abroad (e.g. in Japan, USA) and funded by the DAAD. More information under Web_Science/DAAD-Stipendium/en.

Publications

Active Projects