Advanced Learning in NLP

While many NLP tasks can be tackled easily with standard supervised unsupervised machine learning, some task and data settings raise the need for more advanced learning techniques. The NLP Group has extensive experiences on the use and adaptation of techniques such as multitask learning, contrastive learning, reinforcement learning, adapters, kernel-based learning, and Gaussian mixture models.

• Ziegenbein et al. (2026). Timon Ziegenbein,  Maja Stahl, and Henning Wachsmuth. Teaching LLMs Human-Like Editing of Inappropriate Argumentation via Reinforcement Learning. ACL 2026, to appear.
• Spliethöver et al. (2024). Maximilian Spliethöver, Sai Nikhil Menon, and Henning Wachsmuth. Disentangling Dialect from Social Bias via Multitask Learning to Improve Fairness. ACL Findings 2024.
• Ziegenbein et al. (2024). Timon Ziegenbein, Gabriella Skitalinskaya, Alireza Bayat Makou, and Henning Wachsmuth. LLM-based Rewriting of Inappropriate Argumentation using Reinforcement Learning from Machine Feedback. ACL 2024.
• Sengupta et al. (2023). Meghdut Sengupta, Milad Alshomary, Ingrid Scharlau, and Henning Wachsmuth. Modeling Highlighting of Metaphors in Multitask Contrastive Learning Paradigms. EMNLP Findings 2023.
• Chen et al. (2020). Wei-Fan Chen, Khalid Al Khatib, Benno Stein, and Henning Wachsmuth. Detecting Media Bias in News Articles using Gaussian Bias Distributions. EMNLP Findings 2020.

• DFG-TRR 318 Subproject C04 (since 2021): Metaphors as an Explanation Tool
• DFG-CRF 901 Subproject B1 (2019–2023): Parameterized Service Configurations