Member role: PhD Student
| Simon Ging is a PhD candidate at the University of Freiburg, working with Prof. Thomas Brox on vision–language models. His current focus is radiology: getting models to connect what a scan shows with what a radiologist writes about it. Two papers at MICCAI 2026 come out of this line of work: disease-aware pretraining for 3D CT, and a scalable recipe for spatially grounded 2D radiology models. Before medicine, he worked on general-purpose vision–language problems: COOT (NeurIPS 2020) introduced a hierarchical transformer for video-text representation learning; an ICLR 2024 spotlight repurposed classification datasets and their label hierarchies into an open-ended VQA benchmark; a GCPR 2025 oral used knowledge graphs to harvest better CLIP training data. He holds an M.Sc. from Freiburg and, before that, several years of experience as a software developer and SAP consultant. |
| I am a PhD student at the Max Planck Institute for Intelligent Systems in Tübingen, working on astronomical imaging using machine learning under the supervision of Prof. Bernhard Schölkopf and Prof. Hans-Walter Rix. My first-author paper on combining continuous and discrete adversarial training for LLMs was published at NeurIPS 2025, based on research I conducted during an internship at INSAIT focused on AI safety. I hold an MPhil in Machine Learning from the University of Cambridge and a BSc in Computer Science with honours from Eötvös Loránd University. My broader research background spans generative material discovery, knot theory with ML, and medical AI. In high school, I won bronze medals at both the International Physics Olympiad (IPhO) and the European Physics Olympiad (EuPhO). |
Frederik Nolte is a DPhil candidate in Engineering Science at the University of Oxford, specialising in applied machine learning, with a focus on object-centric and modular approaches for robotics, representation learning, and world models. He is supervised by Ingmar Posner and Bernhard Schölkopf and is a member of the ELLIS PhD Programme.
He previously completed an MSc in Artificial Intelligence at the University of Amsterdam and a BSc in Business Informatics at the University of Mannheim.
I am a PhD researcher at Helmholtz Munich, affiliated with the Technical University of Munich and the ELLIS and ELIZA networks. I hold a bachelor’s degree in Physics and a master’s degree in Statistics and Data Science. My research focuses on generative AI for peptide and protein design, particularly machine learning methods for discovering antimicrobial and therapeutic peptides. I work with protein language models, geometric representation learning, generative modeling, and molecular optimization to design biologically active peptide candidates. Alongside these applications, I conduct research in machine learning theory, with a particular focus on neural collapse, feature learning, kernel methods, and the effects of data geometry and network architecture on learned representations. My work has been published at ICML, Bioinformatics Advances, TAG-DS, and TMLR. I am also a recipient of the Helmholtz AI Accelerator Award and the Deutschlandstipendium from LMU Munich, sponsored by Allianz.
Elias Kempf is an ELLIS PhD student in interpretability and AI safety at the University of Freiburg and Oxford, jointly supervised by Thomas Brox and Fazl Barez. His research studies how learned model behaviors emerge, generalize, and transfer, with a focus on safety-relevant failure modes. Recent work spans domain and compositional generalization in vision-language models, hidden-bias transfer via subliminal learning, and black-box model diffing for comparing language model behaviors. His master’s thesis on CLIP’s domain and compositional generalization received the Förderpreis Künstliche Intelligenz and led to a Spotlight paper at ICML 2025. He has also been a MATS Research Fellow, mentored by Arthur Conmy (prev. Google DeepMind), and previously worked on deep learning for radiotherapy dosimetry at PTW Freiburg.
| Ziheng Chen is an ELLIS PhD student in Information Engineering and Computer Science at the University of Trento, supervised by Prof. Nicu Sebe, and a visiting PhD student at the Max Planck Institute for Intelligent Systems, working with Prof. Bernhard Schölkopf. His research lies at the intersection of machine learning and differential geometry, with a focus on deep learning over non-Euclidean spaces such as Riemannian manifolds, hyperbolic spaces, matrix manifolds, and Lie groups. His work develops geometric counterparts of core neural-network components, including normalization, classification, attention, and hyperbolic layers. His research has appeared at ICLR, NeurIPS, CVPR, ICML, AAAI, and IEEE journals. He has received a DAAD Research Grant, ELSA and ELIAS Mobility Grants. |
I am a PhD candidate at the University of Freiburg working with Thomas Brox. I also hold a parallel position at Bosch Research in Renningen, Germany. My work deals with understanding representation learning in visual models through the lens of interpretability and explainability.
I did my Master’s at RIT, where I was an NSF AWARE-AI Fellow. I also worked at the RIT Center for Human Aware AI under the supervision of Prof. Andreas Savakis on unsupervised domain adaptation in large-scale foundation models.
I have had the pleasure of working alongside researchers and scientists from across the world. Prior to this, I have been associated with Microsoft Research, the Juno Mission at NASA JPL, IRIS-HEP at Princeton, the LIVIA Lab at ETS Montreal, ULR France, CamCann Smart Systems, and IIT Indore.
| Bhavyajeet Singh is a PhD student and doctoral researcher at the Ubiquitous Knowledge Processing Lab, Technical University of Darmstadt, advised by Prof. Iryna Gurevych. His research focuses on developing NLP methods for cybersecurity applications like penetration testing and secure code generation.
Previously, he was a Research Fellow at Microsoft Research India, where he worked on low-latency memory-augmented retrieval systems. This work has been published at ICML 2026 and was integrated into Bing Ads production pipelines. His broader research experience spans secure code generation, retrieval-augmented and memory-augmented NLP systems, and cross-lingual fact-to-text generation. One of his most influential earlier publications is XAlign: Cross-lingual Fact-to-Text Alignment and Generation for Low-Resource Languages, Tushar Abhishek, Shivprasad Sagare, Bhavyajeet Singh, Anubhav Sharma, Manish Gupta, and Vasudeva Varma, 2022. The work introduced a large-scale cross-lingual fact-to-text dataset with 0.45M pairs across eight low-resource languages, supporting research on grounded multilingual generation and cross-lingual alignment. The paper has gathered numerous citations, with the dataset being used my multiple downstream projects and models. |
| I am a PhD student in Multimodal Artificial Intelligence at TU Darmstadt under the supervision of Marcus and Anna Rohrbach and an ELLIS PhD candidate. My research focuses on trustworthy AI, with particular interests in adversarial robustness, AI security, multimodal learning, and vulnerabilities of unified and agent-based systems. I study how modern AI models can be manipulated and how such weaknesses can be detected and mitigated before deployment. Alongside my research, I lead teaching activities for the Statistical Machine Learning course at TU Darmstadt, one of the university’s largest machine learning courses with more than 500 students, and supervise student research projects. Before my PhD, I worked as an independent ML engineer and conducted more than 100 AI workshops across academia and industry. My most influential recent work is GEM: Geometric Erasure by Contrastive Velocity Matching in Rectified Flows (Grebe, Braun, Rohrbach & Rohrbach, ICML 2026 Spotlight), which introduces a novel geometric framework for robust concept erasure in generative models. I was also recognized as an ICML 2026 Gold Reviewer for outstanding reviewing contributions. |