Steffen Eger

About me:

I am a Heisenberg grant recipient of the DFG, leading the Natural Language Learning Group (NLLG) at ~~Bielefeld University~~University of Mannheim, Germany, which is additionally funded by the BMBF through an interdisciplinary AI research group "Metrics4NLG". Previously, I was Interim Professor at Bielefeld University, substituting Philipp Cimiano in the Semantic Computing Group and Junior Research Group Leader (IRGL) at TU Darmstadt. I obtained my PhD in economics on the topic of Social Learning/Opinion Dynamics.

My research interests include Natural Language Processing (NLP), Machine Learning, and applications in the Digital Humanities and Social Sciences. I am particularly interested in

evaluation metrics (for text generation) and text generation
applications in the Social Sciences and Digital Humanities, e.g., change of social biases over time, detection of social solidarity, poetry (generation/evaluation), cross-temporal approaches (language change), cross-linguality, argumentation

Earlier research interests included alignments, opinion dynamics (in a social science/economic context) as well as applications in math (such as integer compositions and limits of powers of matrices).

Email: steffen.eger(at)uni-mannheim.de
Google Scholar | Github | Twitter |

This page is not actively maintained. For up to date information, check my group's page

News

October 2022: New EMNLP paper accepted (Reproducibility)
September 2022: New ACML paper accepted (Constrained Density Matching)
August 2022: New COLING paper accepted (Robust Metrics)
July 2022: New BMBF project "Metrics4NLG" accepted (3 new PhD positions)
June 2022: One paper accepted at the Journal of Social Computing (Measuring Social Solidarity during Crisis: The Role of Design Choices)
January 2022: I am Heisenberg grant recpient of the DFG!
September 2021: One paper accepted at EMNLP 2021 (Global Explainability)
August 2021: One paper accepted at JCDL 2021 (DeuParl)
May 2021: Three papers accepted at ACL 2021 (Changes in European Solidarity during COVID-19, Paired Evaluation, BERT-Defense)
April 2021: We received funding from the AIJ journal for the Eval4NLP-2021 shared task on explainable evaluation metrics
November 2020: We will organize the second Eval4NLP workshop at EMNLP 2021, including a shared task on Explainable Quality Estimation.
October 2020: Two papers accepted at COLING 2020 (Probing MBERT for Genetic Signals, Vec2Sent).
September 2020: One paper accepted at AACL 2020 (Zéroe).
September 2020: One paper accepted at CONLL 2020.
Nominated EACL Area Chair for Sentiment Analysis and Argumentation Mining
May 2020: Our team (David Rother, Thomas Haider, and me) ranked 1st and 2nd place in the SemEval-2020 Task 1: Unsupervised Lexical Semantic Change Detection for English data
April 2020: Two papers accepted at ACL 2020 (XMoverScore, SUPERT).
November 2019: We will organize the first Eval4NLP workshop at EMNLP 2020.
November 2019: FiF Grant from TU Darmstadt (joint with Petra Gehring) on the topic of "Doppelt annotierte Korpora: „Natur & Staat“ und Reichstagsprotokolle als Ressourcen für die innovative automatische Metaphernanalyse"
August 2019: One paper accepted at EMNLP 2019 (MoverScore).
July 2019: One paper accepted at ACL 2019 (Capsule Networks).
November 2018: Two papers accepted at NAACL 2019 (Like Humans, Rebuttal).

Selected Recent Publications

Measuring Social Solidarity during Crisis: The Role of Design Choices
Steffen Eger, Dan Liu, Alexandra Grunow
Journal of Social Computing (JoSoCo), 2022
[paper]
Global Explainability of BERT-based Evaluation Metrics by Disentangling along Linguistic Factors
Marvin Kaster, Wei Zhao, Steffen Eger
EMNLP, 2021
[paper]
The Eval4NLP Shared Task on Explainable Quality Estimation: Overview and Results
Marina Fomicheva, Piyawat Lertvittayakumjorn, Wei Zhao, Steffen Eger, Yang Gao
Eval4NLP, 2021
[paper]
Diachronic Analysis of German Parliamentary Proceedings: Ideological Shifts through the Lens of Political Biases
Tobias Walter, Celina Kirschner, Steffen Eger, Goran Glavaš, Anne Lauscher, Simone Paolo Ponzetto
JCDL, 2021
[paper]
BERT-Defense: A Probabilistic Model Based on BERT to Combat Cognitively Inspired Orthographic Adversarial Attacks
Yannik Keller, Jan Mackensen, Steffen Eger
ACL, 2021
[paper]
Better than Average: Paired Evaluation of NLP systems
Maxime Peyrard, Wei Zhao, Steffen Eger, Robert West
ACL, 2021
[paper]
Changes in European Solidarity Before and During COVID-19: Evidence from a Large Crowd- and Expert-Annotated Twitter Dataset
Alexandra Ils, Dan Liu, Alexandra Grunow, Steffen Eger
ACL, 2021
[paper]
From Hero to Zéroe: A Benchmark of Low-Level Adversarial Attacks
Steffen Eger, Yannik Benz
AACL, 2020
[paper]
How to Probe Sentence Embeddings in Low-Resource Languages: On Structural Design Choices for Probing Task Evaluation
Steffen Eger, Johannes Daxenberger, Iryna Gurevych
CONLL, 2020
[paper]
On the Limitations of Cross-lingual Encoders as Exposed by Reference-Free Machine Translation Evaluation
Wei Zhao, Goran Glavaš, Maxime Peyrard, Yang Gao, Robert West, Steffen Eger
ACL, 2020
[paper]
SUPERT: Towards New Frontiers in Unsupervised Evaluation Metrics for Multi-Document Summarization
Yang Gao, Wei Zhao, Steffen Eger
ACL, 2020
[paper]
PO-EMO: Conceptualization, annotation, and modeling of aesthetic emotions in German and English poetry
Thomas Haider, Steffen Eger, Evgeny Kim, Roman Klinger, Winfried Menninghaus
LREC, 2020
[paper]
MoverScore: Text Generation Evaluating with Contextualized Embeddings and Earth Mover Distance
Wei Zhao, Maxime Peyrard, Fei Liu, Yang Gao, Christian M. Meyer, Steffen Eger
EMNLP, 2019
[paper]
Text processing like humans do: Visually Attacking and Shielding NLP systems
Steffen Eger, Gözde Gül Sahin, Andreas Rückle, Ji-Ung Lee, Claudia Schulz, Mohsen Mesgar, Krishnkant Swarnkar, Edwin Simpson, Iryna Gurevych
NAACL, 2019
[paper]
Does my rebuttal matter? Insights from a major NLP conference
Yang Gao, Steffen Eger, Ilia Kuznetsov, Iryna Gurevych
NAACL, 2019
[paper]

Teaching

I am currently teaching Deep Learning for NLP at Uni Bielefeld as well as Statistical Natural Language Processing and Ethical, Social, and Legal Aspects of Data Science.
I taught the course Deep Learning for Natural Language Processing (DL4NLP) at TU Darmstadt from 2016 to 2020. I taught Foundations of Language Technology (FOLT) from 2020 to 2021. I also held the seminars Deep Learning and Digital Humanities as well as Meta-Science and Evaluation. More information at my TU page.