CLASP
The Centre for Linguistic Theory and Studies in Probability

Neurosymbolic Models of Uncertainty and Logical Reasoning

Abstract

In this talk, I will discuss challenges with modeling uncertainty and logical reasoning in natural language processing, taking question answering in complex technical domains and solving logic puzzles as exemplary tasks. Generative language models fail when it comes to analyzing imprecise statements, modality, hedging, conditions, and mathematical deductions. To this end, I will present recent work on neurosymbolic models for Bayesian reasoning scenarios that are expressed in natural language text. Our results demonstrate that large language models mostly assume causal relations but fail on evidential and explaining-away reasoning scenarios, while our neuro-symbolic approach is robust across types of reasoning. I will also present recent results on solving logic puzzles by teaching LLMs to write code for answer set programming. Finally, I will give a sneak preview on ongoing work in the areas of digital humanities and corpus-linguistic studies of discourse.

References:

Timo Pierre Schrader, Lukas Lange, Tobias Kaminski, Simon Razniewski, Annemarie Friedrich. A Solver-in-the-Loop Framework for Improving LLMs on Answer Set Programming for Logic Puzzle Solving. In Proceedings of the 40th Annual AAAI Conference on Artificial Intelligence. January 2026. Singapore. (to appear)

Timo Schrader, Lukas Lange, Simon Razniewski, and Annemarie Friedrich. QUITE: Quantifying Uncertainty in Natural Language Text in Bayesian Reasoning Scenarios. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP).

Hanna Schmück, Michael Reder, Katrin Paula, and Annemarie Friedrich. A Case Study on Annotating and Analysing Situation Entity Types in Reddit Discussions on Democracy. Accepted to CMC Corpora. September 2025. Bayreuth, Germany.

Mini Bio: Annemarie Friedrich is Professor for Computational Linguistics at the Institute of Computer Science of the University of Augsburg. Her research interests are computational semantics, discourse processing, and linguistic annotation. Before joining the University of Augsburg in 2023, she worked as a Senior Expert for Natural Language Processing at the Bosch Center for Artificial Intelligence, where she focused on text mining for the scientific domain. Annemarie is currently serving as President of the German Society for Computational Linguistics and Language Technology (GSCL).