About me
I am a researcher in the Machine Learning group, led by Prof. Robert Babuška, at the Czech Institute of Informatics, Robotics, and Cybernetics (CIIRC), Czech Technical University in Prague. My research focuses on robust, reliable, and safe AI systems, in particular on the safety and security of generative and multimodal AI, agentic and embodied AI, and AI applications in medicine, aiming to translate advanced AI technologies into systems that can be trusted in real-world, human-facing contexts. I continue to collaborate on the societal and ethical aspects of generative AI with the team of Dr. Nuria Oliver at the ELLIS Unit Alicante, where I previously worked as an ELLIS Postdoctoral Researcher. My areas of interest comprise human-centric artificial intelligence, large language models, robotics, computer vision, reinforcement learning, and genetic algorithms. I am a member of the ELLIS network.
Education
I hold a Ph.D. degree in Robotics and Machine Learning from the Czech Technical University in Prague. My Ph.D. thesis, defended in 2022, was honored with the Werner von Siemens Award in the Industry 4.0 category and the CTU FEE Dean's Award for a Prestigious Dissertation. Prior to my doctoral studies, I completed the selective B.Sc. and M.Sc. study program Open Informatics at the Faculty of Electrical Engineering, Czech Technical University in Prague, which gave me a solid background in machine learning and mathematics.
International experience
I find it very inspiring to work in an international environment. I have taken the opportunity to go for several study and research stays:
- ELLIS Unit Alicante, Alicante, Spain — Institute of Human-Centered AI (2023–2025, 2 years) — ELLIS Postdoctoral Researcher position
- ELLIS Unit Alicante, Alicante, Spain — Institute of Human-Centered AI (2022, 2 months) — research visit
- Carlos III University of Madrid (UC3M), Leganés (Madrid), Spain — Robotics Lab, Escuela Politécnica Superior (2018–2019 & 2021, 10 months in total) — Erasmus+ doctoral research stay
- TU Delft, Delft, The Netherlands — Cognitive Robotics, Faculty of 3mE (2017, 3 months) — research stay
- University of Ljubljana, Ljubljana, Slovenia — ViCoS Lab, Faculty of Computer and Information Science (2015, 1 semester) — Erasmus+ study & research stay
- Technical University of Denmark (DTU), Lyngby (Copenhagen), Denmark (2013–2014, 1 semester) — Erasmus+ study stay
Research
Projects
Currently, I am involved in the following projects:
- ELLIOT — a Horizon Europe flagship initiative developing open multimodal generalist foundation models (MGFMs) that generalize robustly to real-world conditions. I co-lead the task on the safety and security of MGFMs, focusing on red teaming and on the evaluation of their robustness, reliability, and safety. The project runs from 2025 to 2029 and brings together 34 partners from 13 countries; CIIRC CTU is the Czech national point of contact.
- ROBOPROX (Robotics and Advanced Industrial Production) — a project of the Excellent Research call of the Operational Programme Johannes Amos Comenius, coordinated by CIIRC CTU and running from 2023 to 2028. My work addresses safe generative-AI-powered human–robot interaction.
- CoreSense — a Horizon Europe project developing a hybrid cognitive architecture for deep understanding in autonomous robots (2022–2026). I contribute to CIIRC's part of the project, focusing on safe robot operation in dynamic environments and on physics-aware system modeling.
- CZAI (Czech AI Factory) — a national ecosystem of AI services built around the new KarolAIna AI supercomputer, with AI Campuses in Prague and Brno. My contribution focuses mostly on the use of generative AI in the health domain. CZAI is the Czech node of the pan-European network of AI Factories: national one-stop shops co-funded by the EuroHPC Joint Undertaking that combine AI-optimized supercomputing capacity with data, training, and expert support for researchers, startups, and industry. Nineteen AI Factories have been selected across Europe.
Previously, I was involved in the VIVES project, which formed a part of the PERTE of New Language Economy and aimed at building AI tools supporting the Spanish language as well as the co-official languages. I joined this project in September 2023. I also collaborated on developing safe and secure large language models within the Intel Research Center for Responsible Human-AI Systems (RESUMAIS).
Earlier, I worked on the Robotics for Industry 4.0 (R4I) project, which ran from June 2017 to June 2023. I contributed mostly to tasks from WP1 focused on data-efficient model learning methods and long-term autonomy of mobile robots. I have also contributed to the project Symbolic Regression for Reinforcement Learning (SR4RL).
Publications
Derner, E., Kučera, D., Gulati, A., Bagheri, R. A., & Oliver, N. (2026). Mind the Style: Impact of Communication Style on Human-Chatbot Interaction. Computers in Human Behavior: Artificial Humans (9), August 2026, 100362.
PDF
BIB
Arnaiz-Rodríguez, A., Baidal, M., Derner, E., Layton Annable, J., Ball, M., Ince, M., Pérez Vallejos, E., & Oliver, N. (2026). Between Help and Harm: An Evaluation Study of Mental Health Crisis Handling by Large Language Models. JMIR Mental Health (13), June 2026, e88435.
PDF
BIB
GIT
Hagendorff, T., Derner, E., & Oliver, N. (2026). Large Reasoning Models Are Autonomous Jailbreak Agents. Nature Communications 17 (1), February 2026, 1435.
PDF
BIB
Derner, E., & Batistič, K. (2025). Beyond Words: Multilingual and Multimodal Red Teaming of MLLMs. In Proceedings of the First Workshop on LLM Security (LLMSEC), ACL 2025, 198–206, Vienna, Austria.
PDF
BIB
GIT
Derner, E., & Batistič, K. (2025). Gender Representation Bias Analysis in LLM-Generated Czech and Slovenian Texts. In Proceedings of the 10th Workshop on Slavic Natural Language Processing (Slavic NLP 2025), ACL 2025, 124–135, Vienna, Austria.
PDF
BIB
GIT
Baidal, M., Derner, E., & Oliver, N. (2025). Guardians of Trust: Risks and Opportunities for LLMs in Mental Health. In Proceedings of the Fourth Workshop on NLP for Positive Impact (NLP4PI), ACL 2025, 11–22, Vienna, Austria.
PDF
BIB
Derner, E., Sansalvador de la Fuente, S., Gutiérrez, Y., Moreda, P., & Oliver, N. (2025). Leveraging Large Language Models to Measure Gender Representation Bias in Gendered Language Corpora. Proceedings of the 6th Workshop on Gender Bias in Natural Language Processing (GeBNLP), ACL 2025, 468–483, Vienna, Austria.
PDF
BIB
GIT
Derner, E., Batistič, K., Zahálka, J., & Babuška, R. (2024). A Security Risk Taxonomy for Prompt-Based Interaction With Large Language Models. IEEE Access (12), August 2024, 126176–126187.
PDF
BIB
Derner, E., Kučera, D., Oliver, N., & Zahálka, J. (2024). Can ChatGPT Read Who You Are? Computers in Human Behavior: Artificial Humans 2 (2), August 2024, 100088.
PDF
BIB
Vastl, M., Kulhánek, J., Kubalík, J., Derner, E., & Babuška, R. (2024). SymFormer: End-to-End Symbolic Regression Using Transformer-Based Architecture. IEEE Access (12), March 2024, 37840–37849.
PDF
BIB
WEB
GIT
Kubalík, J., Derner, E., & Babuška, R. (2023). Toward Physically Plausible Data-Driven Models: A Novel Neural Network Approach to Symbolic Regression. IEEE Access (11), June 2023, 61481–61501.
PDF
BIB
Derner, E., & Batistič, K. (2023). Beyond the Safeguards: Exploring the Security Risks of ChatGPT. arXiv preprint arXiv:2305.08005, May 2023.
PDF
BIB
Derner, E., & Zahálka, J. (2022). Benefits and Risks of AI Companions. Poster presented at the ELLIS Doctoral Symposium 2022, Alicante, Spain, September 2022.
PDF
BIB
Kulhánek, J., Derner, E., Sattler, T., & Babuška, R. (2022). ViewFormer: NeRF-Free Neural Rendering from Few Images Using Transformers. In 17th European Conference on Computer Vision, ECCV 2022, 198–216, Tel Aviv, Israel.
PDF
BIB
WEB
GIT
Derner, E. (2022). Data-Efficient Methods for Model Learning and Control in Robotics. Doctoral Thesis. Czech Technical University in Prague. Defended in May 2022.
PDF
BIB
Derner, E., Kubalík, J., & Babuška, R. (2021). Guiding Robot Model Construction with Prior Features. In 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 7112–7118, Prague, Czech Republic.
PDF
BIB
Kubalík, J., Derner, E., & Babuška, R. (2021). Multi-Objective Symbolic Regression for Physics-Aware Dynamic Modeling. Expert Systems with Applications (182), November 2021, 115210.
PDF
BIB
Kubalík, J., Derner, E., Žegklitz, J., & Babuška, R. (2021). Symbolic Regression Methods for Reinforcement Learning. IEEE Access (9), October 2021, 139697–139711.
PDF
BIB
Kulhánek, J., Derner, E., & Babuška, R. (2021). Visual Navigation in Real-World Indoor Environments Using End-to-End Deep Reinforcement Learning. IEEE Robotics and Automation Letters 6(3), 4345-4352.
PDF
BIB
WEB
GIT
Derner, E., Kubalík, J., & Babuška, R. (2021). Selecting Informative Data Samples for Model Learning Through Symbolic Regression. IEEE Access (9), January 2021, 14148–14158.
PDF
BIB
GIT
VIDEO
Derner, E., Gómez, C., Hernández, A. C., Barber, R., & Babuška, R. (2021). Change Detection Using Weighted Features for Image-Based Localization. Robotics and Autonomous Systems (135), January 2021, 103676.
PDF
BIB
Hernández, A. C., Derner, E., Gómez, C., Barber, R., Babuška, R. (2020). Efficient Object Search Through Probability-Based Viewpoint Selection. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 6172–6179, Las Vegas, NV, USA.
PDF
BIB
Gómez, C., Hernández, A. C., Derner, E., Barber, R., & Babuška, R. (2020). Object-Based Pose Graph for Dynamic Indoor Environments. IEEE Robotics and Automation Letters 5(4), 5401–5408.
PDF
BIB
Derner, E., Kubalík, J., Ancona, N., & Babuška, R. (2020). Constructing Parsimonious Analytic Models for Dynamic Systems via Symbolic Regression. Applied Soft Computing (94), September 2020, 106432.
PDF
BIB
GIT
Kubalík, J., Derner, E., & Babuška, R. (2020). Symbolic Regression Driven by Training Data and Prior Knowledge. In Proceedings of the Genetic and Evolutionary Computation Conference Companion (GECCO '20), 958–966, Association for Computing Machinery, New York, NY, USA.
PDF
BIB
Derner, E., Gómez, C., Hernández, A. C., Barber, R., & Babuška, R. (2019). Towards Life-Long Autonomy of Mobile Robots Through Feature-Based Change Detection. In 2019 European Conference on Mobile Robots (ECMR), 1–6, Prague, Czech Republic.
PDF
BIB
Kulhánek, J., Derner, E., de Bruin, T., & Babuška, R. (2019). Vision-based Navigation Using Deep Reinforcement Learning. In 2019 European Conference on Mobile Robots (ECMR), 1–8, Prague, Czech Republic.
PDF
BIB
WEB
GIT
Gómez, C., Hernández, A. C., Derner, E., & Barber, R. (2019). Semantic Localization Through Propagation of Scene Information in a Hierarchical Model. In 2019 European Conference on Mobile Robots (ECMR), 1–6, Prague, Czech Republic.
PDF
BIB
Hernández, A. C., Gómez, C., Derner, E., Barber, R. (2019). Indoor Scene Recognition Based on Weighted Voting Schemes. In 2019 European Conference on Mobile Robots (ECMR), 1–6, Prague, Czech Republic.
PDF
BIB
Derner, E., Kubalík, J., & Babuška, R. (2018). Reinforcement Learning with Symbolic Input-Output Models. In 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 3004–3009, Madrid, Spain.
PDF
BIB
VIDEO
Teaching
Courses
2017, 2021–2022 — lab sessions of the course Dynamics and Control of Networks at the Faculty of Electrical Engineering, Czech Technical University in Prague.
Student works
Feel most welcome to contact me if you are interested in working on topics that fall within the scope of my research. We can identify together a suitable student project, thesis, internship, etc.
Finished works:
- Risks and Opportunities for LLMs in Mental Health — Miguel Baidal Marco, internship (2025)
- Contextual Bias Detection Through Semantic Analysis Using Large Language Models — Elena Maestre Hernández, internship (2024)
- Automated Output Classification for Evaluating the Safeguards of Large Language Models — Taya Prymak, volunteering project (2024)
- Red-Teaming Large Language Models for Toxicity Evaluation in Low-Resource Languages — Juan José Bayona Reig, internship (2024)
- Automated Evaluation of Biases in Text Using Large Language Models — Sara Sansalvador de la Fuente, internship (2024)
- Co-Evolutionary Approach to Symbolic Regression — Přemysl Pilař, B.Sc. thesis (2023)
- Real-Time Assistance for Visually Impaired Individuals — Ernesto Iván Ochoa Hidalgo, M.Sc. thesis (2023)
- Visual Navigation Using Deep Reinforcement Learning — Jonáš Kulhánek, B.Sc. thesis (2019)
- Life-long Visual Localization of a Mobile Robot in Changing Environments — Kristýna Kumpánová, M.Sc. thesis (2018)
- Using Model Learning Actor-Critic (MLAC) for Reinforcement Learning with Symbolic Regression — Loi Do, summer internship (2017)