Semih Cantürk

PhD Student, Mila & UdeM DIRO • ML Engineer, Zetane Systems

prof_pic.png

I’m a second year PhD student at Mila – Quebec AI Institute and Université de Montréal DIRO, in Guy Wolf’s group. My interests cover theory and applications of machine learning with a focus on graph representation learning and spectral graph theory. In addition, I am interested in explainable AI (XAI), domain adaptation, out-of-distribution generalization & reinforcement learning. More recently, I have been working on applying GRL on molecular data, solving optimization problems via GRL, improving the scalability of graph neural networks, and self-supervised learning on graphs.

I’m also an ML Engineer at Zetane Systems, where I focus on XAI and lead ML projects with clients and industry partners.

I completed my MSc also at Mila & UdeM, and previously obtained my BEng from the University of Pennsylvania (Penn). I’ve previously interned at the Imperial College Data Science Institute (distributed computing), SAS Analytics (data science) and InfoTRON (AR/VR).

news

Nov 26, 2024 Towards a General Recipe for Combinatorial Optimization with Multi-Filter GNNs has been accepted to LoG 2024 as a Spotlight! See you next week online in poster session 1 (Nov 27, 14:00 EST) or our talk on the 29th (14:30 EST)!
Nov 18, 2024 Two new pre-prints are up! Check out the selected publications below for ‘Towards Graph Foundation Models: A Study on the Generalization of PSEs’, a follow-up paper on GPSE, and ‘OpenQDC’, a large collection of open-source quantum molecular datasets, collated and developed by Valence Labs.
May 1, 2024 GPSE has been accepted to ICML 2024! I will be attending the conference in Vienna between July 21-27, so feel free to reach out if you want to meet up.
Jan 24, 2024 I have been awarded the Université de Montréal PhD Scholarship in Artificial Intelligence (Bourse en Intelligence Artificielle 2023-2024 des ESP)!

selected publications

  1. LoG
    Towards a General Recipe for Combinatorial Optimization with Multi-Filter GNNs
    Frederik Wenkel, Semih Cantürk, Stefan Horoi, and 2 more authors
    2024
  2. arXiv
    Towards Graph Foundation Models: A Study on the Generalization of Positional and Structural Encodings
    Billy Joe Franks, Moshe Eliasof, Semih Cantürk, and 4 more authors
    2024
  3. arXiv
    OpenQDC: Open Quantum Data Commons
    Cristian Gabellini, Nikhil Shenoy, Stephan Thaler, and 5 more authors
    2024
  4. ICML
    Graph Positional and Structural Encoder
    Semih Cantürk, Renming Liu, Olivier Lapointe-Gagné, and 4 more authors
    21–27 jul 2024
  5. LoG
    Taxonomy of Benchmarks in Graph Representation Learning
    Renming Liu, Semih Cantürk, Frederik Wenkel, and 9 more authors
    In Learning on Graphs Conference, 09–12 dec 2022