Maria-Irina Nicolae

Lead Researcher in Machine Learning and Security at Bosch Center for Artificial Intelligence

Robert Bosch Campus 1
71272 Renningen, Germany
Email: irina(dot)nicolae(at)bosch(dot)com

Research interests

• Security for machine learning, adversarial attacks
• LLM evaluation, alignment and trust
• Machine learning for security: fuzzing, intrusion detection, AI agents

Short bio

I am a machine learning scientist Bosch Center for Artificial Intelligence. Previously, I worked as a research scientist in the AI & Machine Learning team at IBM Research. I obtained my Ph.D. from the University of Saint-Etienne in 2016 under the supervision of Marc Sebban and Éric Gaussier. I graduated from Politehnica University of Bucharest in Computer Science in 2011, and from ENSIMAG in Information Systems in 2013.

I am the editor of the Security Garten blog on security, privacy and safety.

[Resume]

Talks

• New Securing LLMs Against Extraction Attacks
  Women in AI Security Workshop, Alan Turing Institute, 2025
• Fuzz testing with machine learning
  Dagstuhl Seminar on Intelligent Security: Is "AI for Cybersecurity" a Blessing or a Curse?, 2022
• Arsenal demo at BlackHat Europe, 2019
• Arsenal demo at BlackHat USA & Europe, 2018 [details]
• Tutorial at Nemesis 2018 Workshop on Adversarial Machine Learning, co-located with ECML/PKDD
• RSA 2018
  
• DefCamp 2017
  
• GreHack 2017 [video]

Publications

[Google Scholar profile] [dblp profile]

2025

• New Y. Shen, Z. Zhuang, K. Yuan, M.-I. Nicolae, N. Navab, N. Padoy, M. Fritz
  Medical Multimodal Model Stealing Attacks via Adversarial Domain Alignment. AAAI, 2025.
  [arXiv preprint] [blog post]

2024

• Z. Zhuang, M.-I. Nicolae, M. Fritz
  Stealthy Imitation: Reward-guided Environment-free Policy Stealing. ICML, 2024.
  [PDF] [arXiv preprint]

2023

• M.-I. Nicolae, M. Eisele, A. Zeller
  Revisiting Neural Program Smoothing for Fuzzing. ESEC/FSE, 2023.
  [PDF] [arXiv preprint] [MLFuzz repo]

2018

• F. Pollok, S. Boag, M.-I. Nicolae
  Open Fabric for Deep Learning Models. MLOSS Workshop: Sustainable Communities at NeurIPS, 2018.
  [PDF]
• M.-I. Nicolae, M. Sinn, M. N. Tran, A. Rawat, M. Wistuba, V. Zantedeschi, I. Molloy, B. Edwards
  Adversarial Robustness Toolbox.
  [PDF]

2017

• A. Rawat, M. Wistuba, M.-I. Nicolae
  Harnessing Model Uncertainty for Detecting Adversarial Examples. NIPS Workshop on Bayesian Deep Learning, 2017.
  [PDF][arXiv preprint]
• V. Lonij, A. Rawat, M.-I. Nicolae
  Extending Knowledge Bases Using Images. NIPS Workshop on Automated Knowledge Base Construction, 2017.
• V. Zantedeschi, M.-I. Nicolae, A. Rawat
  Efficient Defenses Against Adversarial Attacks. ACM CCS Workshop on Artificial Intelligence and Security (AISec), 2017.
  [arXiv preprint]
• V. Lonij, A. Rawat, M.-I. Nicolae
  Open-World Visual Recognition Using Knowledge Graphs, arXiv:1708.08310, 2017.
  [PDF]
• M. Sinn, A. Rawat, M.-I. Nicolae
  Rigorous Analysis of Adversarial Training with Empirical Distributions. ICML Workshop on Implicit Generative Modeling, 2017.
• M. Sinn, A. Rawat, M.-I. Nicolae
  Practical Adversarial Training with Empirical Distributions. ICML Workshop on Implicit Generative Modeling, 2017.

2016

• M.-I. Nicolae
  Learning Similarities for Linear Classification: Theoretical Foundations and Algorithms. Ph.D. thesis, Université Jean Monnet de Saint-Étienne, 2016.
  [PDF]
• M.-I. Nicolae, E. Gaussier, A. Habrard, M. Sebban
  Similarity Learning for Time Series Classification. arXiv:1610.04783, 2016.
  [PDF]
• M.-I. Nicolae, E. Gaussier, A. Habrard, M. Sebban
  Apprentissage de Similarités pour la Classification de Séries Temporelles Multivariées. French Conference on Machine Learning (CAp'16), 2016.
  

2015

• M.-I. Nicolae, M. Sebban, A. Habrard, E. Gaussier, M.-R. Amini
  Algorithmic Robustness for Semi-Supervised (ε, γ, τ)-Good Metric Learning. ICONIP, 2015.
  
• M.-I. Nicolae, M. Sebban, A. Habrard, E. Gaussier
  Joint Semi-Supervised Similarity Learning for Linear Classification. ECML/PKDD, 2015.
  [PDF] [slides] [poster] [bibtex]
• M.-I. Nicolae, M. Sebban, A. Habrard, E. Gaussier
  Apprentissage Joint Semi-Supervisé de Bonnes Similarités. French Conference on Machine Learning (CAp'15), 2015.
  
• M.-I. Nicolae, M. Sebban, A. Habrard, E. Gaussier, M.-R. Amini
  Algorithmic Robustness for Learning via (ε, γ, τ)-Good Similarity Functions. ICLR Workshop, 2015.
  [PDF] [bibtex]