Saruhan Mete Gürbüz

  

Education

Sorbonne University & Université Paris Cité

MSc Cog-SUP — Computational Neuroscience & AI Track

2025–2027
Paris, France

Galatasaray University

BSc Computer Engineering

  • Honors: Best Research Award (2025)
Graduated 2025
Istanbul, Turkey

Skills

Neuroscience: Population dynamics analysis (jPCA, tensor decomposition), dimensionality reduction, neural decoding, Bayesian inference
Machine Learning: Graph Neural Networks (PyTorch Geometric), variational autoencoders, representation learning
Programming: Python (NumPy, SciPy, pandas, scikit-learn), PyTorch, C/C++, Git, Linux
Languages: Turkish (Native), English (Fluent), French (Intermediate)

Publications

An Edge Feature Inclusive Variational Graph Autoencoder for PET-Driven Alzheimer's Diagnosis

Gürbüz, S. M., Adel, M. — 2025 14th International Conference on Image Processing, Theory, Tools & Applications (IPTA), Istanbul, Türkiye, pp. 1–4. DOI: 10.1109/IPTA66025.2025.11222021

2025
IPTA

Research & Projects

Motor Cortex Population Dynamics Analysis

Personal project

  • Built analysis pipeline for MC_RTT dataset to study rotational dynamics in motor cortex populations.
  • Implemented noise injection experiments to test robustness of jPCA-extracted manifolds.
  • Applied sliceTCA to demix covariability classes before trajectory analysis.
Dec 2025 – Present
Paris, France

FSCV Neuromodulator Decoding

Research Intern — INM/UNICOG

  • Developing deep learning models for FSCV to decode monoamine concentrations from noisy time series recordings.
  • Supervisors: Joao Barbosa, Philippe Domenech
Jan 2026 – Present
Paris, France

GINE–VGAE for FDG-PET in Alzheimer's Disease

Bachelor's Thesis, Conference Paper — Advisor: Prof. Dr. Mouloud Adel

  • Developed edge-informed Variational Graph Autoencoder for Alzheimer's classification from FDG-PET scans (93.8% accuracy, 0.937 F1-score).
  • Best Research Award (2025).
  • Published: Gürbüz & Adel, IEEE IPTA 2025, DOI: 10.1109/IPTA66025.2025.11222021.
Oct 2024 – Jun 2025
Istanbul, Turkey

Bayesian Agent Learning & Inference

Course Project — Bayesian modeling of brain and behavior

  • Inferred Bayesian agent priors using analytical and maximum likelihood methods.
  • Developed algorithms to infer learning rules governing belief updates.
Sept 2025 – Dec 2025
Paris, France