Women in Data Science (WiDS) Conference, FAU Erlangen-Nürnberg, 2025

Speaker: Prof. Dr. Andrea Bréard, Vice President Education, FAU

Speaker: Prof. Dr. Franziska Mathis-Ullrich, Department of Artificial Intelligence in Biomedical Engineering, FAU

Title: Cognitive Robotics for Surgery: Elevating Efficiency in the OR

Abstract: The integration of robotics into surgery has revolutionized the field, enabling greater precision and less invasiveness. As we move beyond rigid robotic systems like the Da Vinci and Senhance, the next frontier lies in flexible, sensorized continuum robots capable of navigating complex anatomical environments with minimal tissue damage. This talk explores how the future of surgical robotics hinges on cognitive capabilities—robots that perceive, learn from expert demonstrations, and adapt to dynamic surgical contexts. Leveraging machine learning, especially reinforcement and imitation learning, we develop systems that not only assist but also collaborate naturally with surgeons. Additionally, we address the challenges of controlling flexible instruments through a hybrid approach combining classical modeling, data-driven techniques, and real-time sensor feedback. Together, these advances lay the groundwork for context-aware, semi-autonomous surgical robots, fundamentally advancing the human-robot partnership in the operating room.

Speaker: Dr. Carolin Kaiser, Head of Artificial Intelligence

Title: The Human Touch in AI: Exploring Its Influence on Consumer Choices

Abstract: As AI technology continues to evolve, shopping interfaces are becoming not only smarter but also increasingly human-like. Chatbots now engage in sales conversations, voice assistants process verbal orders, and avatars and robots can even communicate non-verbally with customers through smiles or eye contact. This growing human-like quality of machines is changing how people perceive and interact with them. It raises important questions about the psychological impact of such relationships, including the potential to influence consumer decisions. This talk presents various studies on different shopping interfaces to show how interactions with artificial intelligence affect consumer perception, buying attitudes, and purchasing behavior. It also discusses the opportunities and risks for businesses, consumers, and society.

Speaker: Hannah Braun, Ph.D. Candidate, The Assistive Intelligent Robotics (AIROB) Laboratory, FAU

Title: Across Limbs and Interfaces: Advancing sEMG-Based Control for Intuitive Prosthetics

Abstract: Many myoelectric prostheses are rejected due to complexity, calibration demands, or lack of intuitive control. This keynote presents work across two domains — lower- and upper-limb prosthesis control — using surface EMG (sEMG) signals as a user-intent interface. In the lower-limb system, we explore real-time ankle prosthesis control via ridge regression with a 9-second calibration and evaluate performance using the Target Achievement Control (TAC) test. In the upper limb, we examine the signal quality and movement classification potential of an Agonist-Antagonist Myoneural Interface (AMI) using high-density sEMG and Linear Discriminant Analysis (LDA). These insights highlight how signal separability, training time, and control strategy can shape the usability of prostheses.

Speaker: Dr. Tanja Kaiser, Research Group Leader, Artificial Intelligence and Robotics Lab

Title: Multi-Robot Learning: Towards Intelligent Robot Groups

Abstract: No longer confined to structured environments such as factory floors, robots are becoming more ubiquitous in our world, including where we live and work, and their numbers will continue to grow. Robots will inevitably need to interact with each other, whether in cooperation or competition. In many applications, multi-robot systems (MRSs) offer even greater efficiency and robustness than single-robot systems. However, coordinating multiple robots is challenging due to our ever-changing world. In this talk, I will use examples from our research to show how machine learning and generative AI techniques can help us overcome remaining challenges in multi-robot systems to make them ready for our everyday environments.

Speaker: Dr.-Ing. Siming Bayer, Research Group Leader, Chair of Computer Science 5 (Pattern Recognition), FAU

Title: Multimodal Data Analysis with Machine Learning – a journey between departments and across borders

Abstract: This presentation illustrates interdisciplinary and international approaches to applying advanced machine learning methodologies across healthcare and energy sectors. It outlines the development and application of digital twins using multimodal data and knowledge graphs for utility management. Further, the focus shifts to healthcare, highlighting AI-based frameworks for enhanced patient-centered care, specifically in medical screening, diagnostic support, and surgical planning. A detailed probabilistic registration framework is introduced for multimodal image alignment, primarily targeting vascular structures. Clinical validations of these methodologies, including intraoperative brain shift compensation and longitudinal pulmonary studies, demonstrate substantial improvements in accuracy and robustness. Further outlooks for multimodal data analysis are given considering the current development of advanced AI techniques.

Speaker: Melanie B. Sigl, Managing Consultant at PRODATO and Ph.D. Candidate at Chair of Computer Science 6 (Data Management), FAU

Title: Scalable Optimization of Large-Scale Systems Using MLOps

Abstract: Optimizing large systems is essential for long-term cost reduction and improved product efficiency across supply chains. However, implementing plant-specific Advanced Process Controls can be expensive, labor-intensive, and time-consuming. Machine learning techniques can significantly reduce this effort. The resulting model serves as a digital twin of the plant, enabling simulation and optimization of various plant settings. The goal, following a successful proof of concept, is to fully automate and scale the operationalization of the entire infrastructure and data pipeline. This talk explores the migration of the machine learning project from on-premise to the Cloud. It discusses the challenges of implementing a scalable MLOps process and how these challenges were addressed in the project.

Speaker: Dr. Emmanuelle Salin, Group Leader, Department of Artificial Intelligence in Biomedical Engineering, FAU

Title: Deep Learning Approaches for Clinical Data

Abstract: The increasing digitalization of clinical data offers significant opportunities for the development of deep learning approaches to support clinical decision-making. However, several challenges hinder their development and implementation in clinical practice, including the heterogeneity of clinical data, the complexity of biomedical interactions, and the scarcity of annotated datasets. Combining machine learning with clinical expertise can help better understand and address these challenges. The Erlangen Center for AI in Medicine, established in 2024, aims to foster interdisciplinary research collaboration between clinicians and machine learning researchers. In this talk, I will present examples of ongoing research projects in various medical fields.