MACHINE LEARNING GROUP

RPTU KAISERSLAUTERN-LANDAU

Jun.-Prof. Dr. Sophie Fellenz (née Burkhardt)

Junior professor

Bio

Since 2020 Sophie Fellenz is a junior professor at RPTU's department of computer science. Previously she was a research group leader at Johannes Gutenberg University Mainz (07/2020-10/2020) and before that a PostDoc in the research group of Stefan Kramer. She obtained a PhD from Uni Mainz in 2018 under the supervision of Stefan Kramer and a Magister in Philosophy and Computer Science in 2013 with a thesis on mental representation.

Research interests

Sophie Fellenz is interested in topic models, and, more generally, generative models and deep generative models. More recently, she is also interested in how to insert prior knowledge and constraints into (generative) neural models. In particular, her interest is in combining physical or biological models such as differential equations, known biological pathways, or boundary conditions with experimental data to achieve hybrid, grey-box models for tasks such as time series generation and forecasting, scRNA data generation, or the prediction of molecular properties. Her PhD was about multi-label classification, nonparametric Bayesian models and online models.

Appointments and scientific matters
=YWZsxWZupHQjNnL15WatsGbuQWZ
Office
RPTU, Building 36, Room 331 - 67663 Kaiserslautern

Curriculum Vitae

Education

2018
Doctoral Degree in Computer Science, Johannes Gutenberg University Mainz, Germany
2013
Magister in Philosophy and Computer Science, Johannes Gutenberg University Mainz, Germany

Professional Experience

since 2020
Junior Professor, RPTU Kaiserslautern-Landau, Kaiserslautern, Germany
2017-2020
Research Associate, Johannes Gutenberg University Mainz, Germany

Activities, functions, and honors

Ongoing
Serving regularly as Area Chair for top AI venues, including ECML, ICML, and NeurIPS
Co-organizer DFG SPP 2331 (2024-2029)
Co-organizer ML4CCE Workshop at ECML 2024
PI, DFG KI-FOR 5359 (since 2022)
PI, coordinated BMBF project PIAD (2025-2028)
PI, Carl Zeiss project Artifcial Intelligence for treating cancer therapy resistance AICare (2024-2030)
PI, Collaborative Research Centre / Transregio (TRR) 375 Multifunctional High-Performance Components made of hybrid porous materials (2024-2027)
2020-2024
Junior research group leader of BMBF research group Semantic Disentanglement: Differentiating Style and Topic in Text Data
2019
Best Dissertation Award, Uni Mainz
2013-2017
PhD Scholarship, PRIME Research
2010-2011
Scholarship German Academic Exchange Service (DAAD)

Key publications

  • M. Nagda, P. Ostheimer, and S. Fellenz. Tethering Broken Themes: Aligning Neural Topic Models with Labels and Authors.
    Findings of the Association for Computational Linguistics: NAACL, 2025.
  • M. Nagda and S. Fellenz. Putting Back the Stops: Integrating Syntax with Neural Topic Models.
    Proceedings of International Joint Conference on Artificial Intelligence (IJCAI), 6424-6432, 2024.
  • Burkhardt, S., Kramer, S. (2019) Decoupling Sparsity and Smoothness in the Dirichlet Variational Autoencoder Topic Model. Journal of Machine Learning Research 20.131, pp. 1-27.
  • Burkhardt, S. and Kramer, S. (2018) Online Multi-Label Dependency Topic Models for Text Classification. Machine Learning 107.5, pp. 859-886.
  • Burkhardt, S. and Kramer, S. (2017) Online Sparse Collapsed Hybrid Variational-Gibbs Algorithm for Hierarchical Dirichlet Process Topic Models, ECML-PKDD. Skopje, Macedonia, pp. 189-204.