معرفی
Eleni Straitouri is a Professor at the Max Planck Institute for Software Systems (MPI-SWS) in Kaiserslautern, Germany, where she leads research at the intersection of artificial intelligence safety and human-centered decision systems. Her work bridges theoretical machine learning with practical applications in high-stakes environments requiring rigorous uncertainty quantification and causal reasoning.
Her primary research focuses on causal inference in large language models, counterfactual prediction sets, and human-AI complementarity. She develops methods to control algorithmic harm through counterfactual reasoning, enhances model evaluation via conformal prediction, and designs decision support tools that leverage human expertise. Current projects address critical challenges in LLM safety, including bias detection through coupled token generation and harm mitigation in medical/operational decision systems.
Analysis of her 13 publications (2021-2025) reveals an evolving trajectory from reinforcement learning foundations toward cutting-edge LLM safety frameworks. Her work consistently integrates causal methods with human factors, producing practical tools for reliable AI deployment across healthcare and operational domains. Recent NeurIPS/ICML publications demonstrate increasing influence in trustworthy AI research.
- Best Paper Award, Workshop on AI and HCI (ICML), 2023
- Spotlight Paper (top 3.5%), ICML 2024
Dr. Straitouri maintains active collaborations with Manuel Gomez-Rodriguez's research group, evidenced by co-authorship on 100% of her recent publications. While specific grant funding isn't detailed in available materials, her work aligns with MPI-SWS's focus on software systems for societal benefit. No information about academic advising or laboratory infrastructure was found in the provided text.
