Erik Derner

Erik Derner is an ELLIS Postdoctoral Researcher at the ELLIS Unit Alicante, focusing on human-centric AI and ethical LLMs. He collaborates with Prof. Robert Babuška's Machine Learning team at the Czech Institute of Informatics, Robotics, and Cybernetics (CIIRC) and contributes to the VIVES project under the PERTE of New Language Economy. His work addresses biases in LLMs, AI safety, and security, while also exploring symbolic regression for robotics and reinforcement learning.

• Education: Ph.D. in Robotics and Machine Learning (2022) from Czech Technical University in Prague, with B.Sc. and M.Sc. in Open Informatics.

Erik's research spans human-centric AI, large language models, robotics, computer vision, and reinforcement learning. He emphasizes ethical and secure AI development, particularly for underrepresented languages. His work integrates symbolic regression with neural networks to create interpretable and efficient models for robotics.

Recent publications highlight his focus on LLM security, gender bias analysis, and symbolic regression for dynamic systems. Notable trends include ethical AI, multimodal red-teaming, and physics-informed modeling.

Scientific Awards:

• Werner von Siemens Award (2023) for his Ph.D. thesis.
• CTU FEE Dean's Award (2023) for a prestigious dissertation.

Erik supervises student projects in robotics, AI ethics, and LLM security, including internships and theses on topics like mental health risks, toxicity evaluation, and real-time assistance for the visually impaired.