EVENT

Event News

Talk on "From Models to Decisions: A Holistic NLP Perspective on Intelligent Decision Support" by Dr. Vera Schmitt from TU Berlin

We are pleased to inform you about the upcoming seminar by Dr. Vera Schmitt from TU Berlin titled:"From Models to Decisions: A Holistic NLP Perspective on Intelligent Decision Supports" Everyone interested is cordially invited to attend!

Title:

From Models to Decisions: A Holistic NLP Perspective on Intelligent Decision Supports

Abstract:

Large language models and multimodal NLP systems are increasingly used to support human decision-making in high-stakes domains such as fact-checking, political communication, and public discourse. Despite strong advances in model capabilities, current decision support pipelines remain fragmented: language models are often optimized independently of explanation methods, evaluation metrics, and human interaction requirements, limiting their practical usefulness and trustworthiness. In this talk, Vera Schmitt present the research agenda of the XplaiNLP group, which investigates intelligent decision support from a holistic perspective. Our work integrates representation learning, explainable NLP, and human-centered evaluation across tasks including claim verification, disinformation and propaganda analysis, narrative and framing detection, and multimodal evidence aggregation. Rather than treating explanations and evaluation as post-hoc components, we study how linguistic structure, model adaptation, and evaluation criteria jointly shape interpretability, robustness, and human trust.

Speaker:

Dr. Vera Schmitt from TU Berlin

Dr. Vera Schmitt build up and is leading the XplaiNLP group at the QUL, based on the aquired funding from third-party projects. With her group she is exploring core NLP topisc, xAI, HCI and legal aspects of AI systems in the domain of disinformation detection and medical data processing. Vera completed her undergraduate studies in Politics and Public Administration (B.A.) at the University of Konstanz. During this time, she developed a keen interest in statistics and co-founded CorrelAid, a non-profit community of data science enthusiasts. Afterward, she pursued her passion for data science by enrolling in a Master's program in Data Science at Leuphana University in Lüneburg. As a member of the ChangemakerXchange she actively contributed to projects of CorrelAid in various countries, including Malaysia, Japan, and Singapore. Following this, she started a Ph.D. at the Q&U Labe at the TU Berlin, concerning the topic of economic aspects of privacy. During her PhD, she aquired a BMBF research group funding (KI-Nachwuchsgruppen unter Leitung von Frauen) of 1.4 Million € to significantly extend her already existing group.