Ege Erdogan

inc. PhD Student at the University of Amsterdam

pp.jpg

Hi! I am an incoming PhD student at the University of Amsterdam at the SIAS group advised by Ana Lucic, working on mechanistic interpretability. I am also interested in applications of ML in science, and deep generative models.

I previously did my MSc at the Technical University of Munich and wrote my thesis on geometric flow models over neural network weights, advised by David Rügamer and Bastian Rieck.

Feel free to contact me at ege at erdogan.dev.

news

Mar 2025 🎉 Paper accepted at the MLMP Workshop within ICLR ‘25!
FreeFlow: Latent Flow Matching for Free Energy Difference Estimation
Feb 2025 👨‍⚕️ Will be starting my PhD at UvA in May!
Jan 2025 🎓 Completed my MSc at TUM! Read my thesis here
Jun 2024 📕 Starting my masters thesis on efficient sampling of NN weights using generative models and geometric DL, with David Rügamer and Bastian Rieck.
Jun 2024 🎉 Paper accepted at CANS ‘24.
SplitOut: Out-of-the-Box Training-Hijacking Detection in Split Learning via Outlier Detection

selected publications

  1. MLMP ’25
    FreeFlow: Latent Flow Matching for Free Energy Difference Estimation
    Erdogan, Ege, Ralev, Radoslav, Rebensburg, Mika, Marquet, Céline, Klein, Leon, and Stark, Hannes
    Workshop on Machine Learning Multiscale Processes (at ICLR ’25) 2025
  2. CANS ’24
    SplitOut: Out-of-the-Box Training-Hijacking Detection in Split Learning via Outlier Detection
    Erdogan, Ege, Teksen, Unat, Celiktenyildiz, Mehmet Salih, Kupcu, Alptekin, and Cicek, A. Ercument
    International Conference on Cryptology And Network Security 2024
  3. GLFrontiers ’23
    Poisoning × Evasion: Symbiotic Adversarial Robustness for Graph Neural Networks
    Erdogan, Ege, Geisler, Simon, and Günnemann, Stephan
    New Frontiers in Graph Learning Workshop (at NeurIPS ’23) 2023
  4. (SRW) RANLP ’23
    Detecting ChatGPT: A Survey of the State of Detecting ChatGPT-Generated Text
    Dhaini, Mahdi, Poelman, Wessel, and Erdogan, Ege
    In Proceedings of the 8th Student Research Workshop associated with the International Conference Recent Advances in Natural Language Processing 2023
  5. WPES ’22
    SplitGuard: Detecting and Mitigating Training-Hijacking Attacks in Split Learning
    Erdogan, Ege, Kupcu, Alptekin, and Cicek, A Ercument
    The 21st Workshop on Privacy in the Electronic Society (at ACM CCS ’22) 2022
  6. WPES ’22
    UnSplit: Data-oblivious model inversion, model stealing, and label inference attacks against split learning
    Erdogan, Ege, Kupcu, Alptekin, and Cicek, A Ercument
    The 21st Workshop on Privacy in the Electronic Society (at ACM CCS ’22) 2022
  7. SRDS ’21
    (Demo) Zelig: Customizable Blockchain Simulator
    Erdogan, Ege, Aydin, Can Arda, Ozkasap, Oznur, and Gill, Waris
    40th International Symposium on Reliable Distributed Systems 2021