Students

Master thesis topics

Contribute to Mediverse, a platform for multi-modal clinical data harmonisation through unified graph-based representations aiming to transform clinical decision-making and information retrieval: native harmonisation across modalities and clinics, multi-modal cross-site predictive studies, and inference tuned with uncertainty control to infuse trust.

• Benchmarking retrieve and rerank methods for medical entity linking
• Exploring GNNs for medical entity linking
• Leveraging LLMs for multimodal information extraction
• Multivariate entity linking

To apply, send your CV, transcript and motivation for choosing the project to adio@kth.se

Master’s Students

Here you can find all my supervision work at TU Delft.

• 2026 - KTH
  • Matej Priesol - Leveraging LLMs for Synthetic Data Generation in the Finance Domain
    • Supervised in collaboration with Sonia Horchidan (KTH) and Evangelia Gogoulou (SEBx)
  • Abhinav Ramalingam - Comparative Analysis of Methods for Transforming EHR Data to Canonical Graph Elements, thesis
    • Student from Uppsala University
• 2024 - TU Delft
  • Zeger Mouw - Human Interaction in Tabular Data Augmentation in Data Science Workflows, thesis

  The research conducted by Zeger Mouw has been part of the following papers:

  • Human-in-the-Loop Feature Discovery for Tabular Data, Demo@CIKM 2024
  • Key Insights from a Feature Discovery User Study, HILDA@SIGMOD 2024
• 2022 - TU Delft
  • Wang Hao Wang - An exploratory journey to combine schema matchers for better relevance prediction, thesis

Bachelor’s Students

During my PhD at TU Delft, I have created and structured the research projects for the bachelor’s thesis, and lead the supervision of the following students:

• 2023 / Q4 – Project “Automatic feature discovery to improve Machine Learning performance”, responsible prof: Asterios Katsifodimos
  • Andrei Manastireanu – Automatic Feature Discovery: A comparative study between filter and wrapper feature selection techniques, paper, poster
  • Andrei Udila – Encoding Methods for Categorical Data: A Comparative Analysis for Linear Models, Decision Trees, and Support Vector Machines, paper, poster
  • Duyemo Anceaux – A comparative study for using PCA, LDA, GDA, and Lasso for dimensionality reduction before classification algorithms, paper, poster
  • Florena Buse – Data-Driven Empirical Analysis of Correlation-Based Feature Selection Techniques, paper, poster
  • Kiril Vasilev – Filtering Knowledge: A Comparative Analysis of Information-Theoretical-Based Feature Selection Methods, paper, poster

  The research conducted by Kiril Vasilev and Florena Buse has been part of the research paper: AutoFeat: Transitive Feature Discovery over Join Paths, ICDE 2024

• 2022 / Q4 – Project “Automatic feature discovery for machine learning”, responsible prof: Rihan Hai
  • Eduard Cruset Pla – From Feature Selection to Data Augmentation: the ADA Algorithm, paper, poster
  • Oliver Neut – Automatic feature augmentation ranking: XGBoost, paper, poster
  • Oskar Lorek – PCADA: Partial Correlation Aware Data Augmentation for random forest classifier, paper, poster