Peter Lorenz

NTU

Adversarial Machine Learning and Model Stealing

I am currently a Postdoctoral Research Fellow at Nanyang Technological University (NTU) Singapore, ranked 2nd worldwide in AI research, focusing on trustworthy AI. I recently completed my Ph.D. with magna cum laude from Heidelberg University, Germany, specializing in adversarial machine learning for classification, generative diffusion models with focus on harmful outlier detection to increase the trustworthness of AI models and prepare them for open-world problems. During my Ph.D. (thesis), which I completed in 3.5 years, I conducted an internship at MIT-IBM Watson AI Lab and achieved notable results, including a top 3% paper award at the ICASSP conference, a top-20 ranking in the CVPR 2022 Art-of-Robustness Challenge, and two times Oxford summer school acceptance.

In addition to my research, I actively contribute to the machine learning community as a reviewer for top-tier conferences such as ICML, ICLR, and NeurIPS. My passion lies in advancing the field of machine learning, with a particular focus on its practical applications.

Previously, I worked in academia and industry on the CARLA simulator for pedestrian safety and classification models for mobile robotics, including vehicles and drones. Before that, I contributed to the vision system of autonomous robots (Team TEDUSAR @TU Graz, that won the robocup competition “best in autonomy” in 2016.) in terms of my Bachelor thesis.

Leave me an anonymous comment!

Tweets by cs_peter_lorenz

news

Jan 07, 2025	Check out my continuously updated reading list about model stealing!
Mar 18, 2024	I am happy to announce that I am a reviewer at the CVPR Workshop Robustness of Foundation Models 🎉
Jan 29, 2024	I am accepted for the Oxford Summer School - Representation Learning
Oct 18, 2023	I am happy to announce that I am reviewer at ICASSP on the topics federated / split learning and quantum privacy 😄
Aug 26, 2023	Check out my writeups from the Lakera Gandalf hackathon.

latest posts

Mar 28, 2025	Recommender Systems
Mar 28, 2025	Imbalanced Data
Mar 28, 2025	MLSD

selected publications

ICML

Deciphering the Definition of Adversarial Robustness for post-hoc OOD Detectors

Peter Lorenz, Mario Fernandez, Jens Mueller, and 1 more author

In ICML 2024 Workshop on the Next Generation of AI Safety , 2024

Bib HTML

@inproceedings{lorenz2024deciphering,
  title = {Deciphering the Definition of Adversarial Robustness for post-hoc OOD Detectors},
  author = {Lorenz, Peter and Fernandez, Mario and Mueller, Jens and Koethe, Ullrich},
  booktitle = {ICML 2024 Workshop on the Next Generation of AI Safety},
  year = {2024},
  url = {https://arxiv.org/pdf/2406.15104},
}

IJCNN

Adversarial Examples are Misaligned in Diffusion Model Manifolds

Peter Lorenz, Ricard Durall, and Janis Keuper

In IJCNN , 2024

Bib HTML

@inproceedings{lorenz2024manifold,
  title = {Adversarial Examples are Misaligned in Diffusion Model Manifolds},
  author = {Lorenz, Peter and Durall, Ricard and Keuper, Janis},
  booktitle = {IJCNN},
  year = {2024},
  url = {https://arxiv.org/pdf/2401.06637.pdf},
}

NeurIPS

Visual Prompting for Adversarial Robustness (top 3% @ ICASSP23)

Aochuan Chen*, Peter Lorenz*, Yuguang Yao, and 2 more authors

In NeurIPS WS TSRML, Safety ML WS, ICASSP23 , 2022

Bib HTML

@inproceedings{prompting,
  title = {Visual Prompting for Adversarial Robustness (top 3% @ ICASSP23)},
  author = {Chen*, Aochuan and Lorenz*, Peter and Yao, Yuguang and Chen, Pin-Yu and Liu, Sijia},
  booktitle = {NeurIPS WS TSRML, Safety ML WS, ICASSP23},
  year = {2022},
  url = {https://arxiv.org/pdf/2210.06284.pdf},
}

AAAI

Is RobustBench/AutoAttack a suitable Benchmark for Adversarial Robustness?

Peter Lorenz, Dominik Strassel, Margret Keuper, and 1 more author

In The AAAI-22 Workshop on Adversarial Machine Learning and Beyond , 2022

Bib HTML

@inproceedings{lorenz2022is,
  title = {Is RobustBench/AutoAttack a suitable Benchmark for Adversarial Robustness?},
  author = {Lorenz, Peter and Strassel, Dominik and Keuper, Margret and Keuper, Janis},
  booktitle = {The AAAI-22 Workshop on Adversarial Machine Learning and Beyond},
  year = {2022},
  url = {https://openreview.net/forum?id=aLB3FaqoMBs},
}

ICML

Detecting AutoAttack Perturbations in the Frequency Domain

Peter Lorenz, Paula Harder, Dominik Straßel, and 2 more authors

In ICML 2021 Workshop on Adversarial Machine Learning , 2021

Bib HTML

@inproceedings{lorenz2021detecting,
  title = {Detecting AutoAttack Perturbations in the Frequency Domain},
  author = {Lorenz, Peter and Harder, Paula and Stra{\ss}el, Dominik and Keuper, Margret and Keuper, Janis},
  booktitle = {ICML 2021 Workshop on Adversarial Machine Learning},
  year = {2021},
  url = {https://openreview.net/forum?id=8uWOTxbwo-Z},
}