Through interdisciplinary collaboration and cutting-edge research methodologies, we strive to advance the state-of-the-art in data science and language technologies, with a focus on addressing real-world problems and societal challenges.
For more information about our ongoing projects, publications, and opportunities for collaboration, please explore the respective sections of our website or reach out to our team members directly. We welcome inquiries from researchers, students, and industry partners interested in joining us on our mission to push the boundaries of knowledge and innovation in these exciting research areas.
We study how foundation models represent people, values, perspectives, disagreement, and social context. We develop methods and benchmarks for human modeling, value-sensitive behavior, social intelligence, pluralism, and alignment stability under personalization and post-training.
We develop efficient, explainable methods for training, adapting, compressing, and evaluating foundation models. Our work spans LLM architectures, multilingual representations, tokenizers, model distillation and domain adaptation, reasoning approaches, and benchmarking.
We build reliable domain-specialized machine learning frameworks and LLM-powered tools for scientific discovery. Our work focuses on uncertainty-aware AI for particle physics, astrophysics, and biomedicine, scientific agents, robust RDM systems, simulation reliability, and statistically valid AI-assisted workflows.
The past decade of AI was largely driven by one question: how to make large language models work at all. How to scale them, stabilize them, and push their capabilities far enough to be usable.
A major success for the b-it: The European Research Council (ERC) is providing millions of euros for projects in the fields of computer science, economics, and evolutionary biology. Prof. Dr. Lucie Flek from Data Science & Language Technologies Group has been awarded with the popular EU funding to continue her research on social parameters in AI.
In the field of Data Perspectivism, perspective has emerged as an umbrella term encompassing annotators’ points of view and culturally shaped worldviews. When modeling annotators, researchers have explored a variety of potential predictors, with demographics receiving particular attention, especially following the rise of techniques such as sociodemographic prompting. In this talk, 1 will examine the field’s strong emphasis on annotators’ sociodemographic information and highlight the limitations of this approach. I will focus on challenges in annotator modeling and the complexities of addressing highly subjective linguistic phenomena, going through data collection, modeling and evaluation.
The study of harms in NLP is a fast-evolving field of research, which in a few years has seen the need of considering the subjectivity that characterizes this phenomenon. In this talk | present two complementary research projects that address this topic from two different perspectives. First, I discuss the systematic presence of bias against women and people with non-Western origin in data filtering strategies for harm reduction in pretraining datasets (Stranisci, & Hardmeier, C., 2025). Then, 1 describe the results of our study on canceling attitudes, whose perception appears to strongly rely on individuals' moral stance rather than sociodemographic features (Lo, et al, 2025).
The emergence of large language models has transformed the landscape of conversational systems, but our understanding of how users interact with these systems and what they seek to accomplish remains limited. This talk presents findings from two empirical studies investigating real-world interactions with LLM-based and voice-based conversational systems. The first study analyses over 15,000 prompts submitted to Google Gemini, revealing how users formulate structured, often imperative inputs that go well beyond traditional informational, navigational, transactional search intents. This analysis highlights the expanding role of LLMs in supporting complex tasks such as content creation and information extraction. The second study examines over 600,000 interactions with Google Assistant across 173 users, offering insight into voice-based conversational systems' everyday utility and limitations. The data reveal a predominance of simple instructions and a lack of deeper information-seeking behaviours. Together, these studies offer a nuanced account of user intent, interaction styles, and the evolving role of conversational systems in supporting diverse and situated information needs.
Parkinson's disease (PD) is a progressive neurodegenerative disorder with a lengthy prodromal phase that remains difficult to capture using traditional clinical tools. Most monitoring begins only after diagnosis, limiting insight into early symptoms and the lived experience of disease progression. In this talk, I will present work evaluating Facebook as a novel, longitudinal data source for studying PD-related disclosures across the disease timeline -from years before diagnosis to later stages.
The increasing burden of mental health disorders-including depression, anxiety, OCD, and suicidal ideation-necessitates the development of advanced Al frameworks capable of interpreting complex emotional signals from language. Our research focuses on context-aware large language models (LLMs) that capture nuanced emotional and psychological patterns embedded in long, unstructured text. These models are designed to preserve semantic coherence and context across sequences, enabling more accurate detection of early mental health risk factors. We introduce a multi-task representation learning approach that integrates subject specific and context-specific features for detecting a range of mental health conditions from both psychiatric and social media texts. This strategy allows for task-specific adaptation while maintaining shared representations, enhancing generalization across related emotional and behavioral tasks. A key aspect of our work involves Hierarchical Explainable Al (XAI), where we employ layered attention mechanisms and graph-based interpretability techniques to identify critical risk-inducing patterns in suicidal and emotionally volatile texts. The framework not only highlights word-level and sentence-level importance but also models higher-order semantic dependencies across text segments, offering transparency in sensitive decision-making contexts. Our current direction explores the use of Explainable Graph Attention Networks and Deep Q-Learning to identify high-risk emotional states and generate context-aware intervention strategies. We further envision the integration of generative Al for producing personalized, real-time supportive responses. Future extensions involve multimodal LLMs that combine text, image, and genetic data for a more holistic understanding of mental health.
Amid the recent uptake of Generative Al, sociotechnical scholars and critics have traced a multitude of resulting harms, with analyses largely focused on values and axiology (e.g., bias). While value- based analyses are crucial, we argue that ontologies-concerning what we allow ourselves to think or talk about-is a vital but under-recognized dimension in analyzing these systems. Proposing a need for a practice-based engagement with ontologies, we offer four orientations for considering ontologies in design: pluralism, groundedness, liveliness, and enactment. We share examples of potentialities that are opened up through these orientations across the entire LLM development pipeline by conducting two ontological analyses: examining the responses of four LLM-based chatbots in a prompting exercise, and analyzing the architecture of an LLM-based agent simulation. We conclude by sharing opportunities and limitations of working with ontologies in the design and development of sociotechnical systems.
Adversarial text-carefully crafted inputs designed to mislead or degrade the performance of NLP systems-poses a growing challenge across a range of language technologies. In this talk, I will present my work on adversarial text detection and methods for improving the quality and stability of such texts once identified. / will discuss the linguistic and structural characteristics of adversarial inputs, outline current approaches for automatic detection, and introduce techniques for refining adversarial examples to make them more semantically coherent. While the primary focus will be on traditional NLP systems, / will also reflect on how these techniques might evolve to address the emerging complexities of large language models (LLMs). Looking ahead, / will highlight how adversarial methods could be leveraged not only for defence but also as diagnostic tools for probing and improving LLM robustness, interpretability, and trustworthiness.
When people comprehend, interpret, or communicate about their environment, they draw on "mental schemata" that encode common knowledge and associations based on experiences, moral values, or beliefs. New information that aligns with existing mental schemata is much more readily understood and accepted. This talk will present two projects that explore the manifestation of media framing, and moral understanding in humans in LLMs. First, / will introduce "narrative media framing," a conceptualization of framing grounded in the social sciences that links media framing devices with cognitively salient narrative representations. Secondly, I will present our recent work where we propose a robust method for probing representations of morality in LLMs through word associations.
See you next year in Utrecht, Netherlands.
Joining LLM superpowers along one DB speed train line! The release of Teuken7B was just a start. In Europe, we need to build not only bigger, but smarter. I am convinced that joining forces with the MCML NLP teams at Ludwig-Maximilians-Universität München and Technical University of Munich, combined with our existing collaborations with hessian.AI and many more, are the right way forward to a new generation of models for EU AI sovereignty.
Don't miss our new paper on forecasting radicalization: "Unifying the Extremes: Developing a Unified Model for Detecting and Predicting Extremist Traits and Radicalization"
Our proposal on an open ML Challenge on Adversarial Learning for Particle Physics ("Colliding with Adversaries") got accepted to ECML 2025. We will set up the CERN LHC data on CodaBench within a month from now, and people are welcome to develop their adversarial attack systems until June 23.
Technology of tomorrow, problems of yesterday: What can be done about the gender gap in AI? How can we inspire more young women to study AI? Why is equal gender participation in the development of Artificial Intelligence so important? And how can we address the gender gap in AI?
Prof. Lucie Flek was invited to the third edition of the HumanCLAIM workshop to talk about stereotypes in multilingual models. This event, organised by Prof. Lisa Beinborn from the University of Göttingen, brought together leading scholars from linguistics, cognitive science, and computer science to develop a more diverse and human-centered perspective on cross-lingual models.
