I am an ELLIS Doctoral Student at AIDOS, supervised by Prof. Bastian Rieck and co-supervised by Prof. Søren Hauberg. My research focuses on geometrical and topological methods in machine learning.

Research Interests

Random principles that guide my scientific reasearch

News

Publications

On the Rademacher Complexity of GNNs: Unifying Expressivity and Geometry
M. Carrasco, V. A. Martirosyan, A. Mehrab, C. Netto, E. Okoyomon, and C. Graziani
Preprint, 2026
M. Carrasco, O. Zaghen, K. Sumaraj, E. Bekkers, and B. Rieck
Preprint, 2026
We propose a metric that captures both features and struture in attributed graphs.
K. Limbeck, N. Häusermann, M. Carrasco, G. Wolf, and B. Rieck
Preprint, 2026
We propose diversity curves, novel graph-level representations for unsupervised tasks that track the structural diversity, namely the metric space spread, of graphs across coarsening levels obtained through edge contractions.
J. S. Schmidt, M. Carrasco, E. Röell, G. Wolf, N. Blaser, and B. Rieck
Preprint, 2026
We find topological generalization is lacking in deep learning and present an evaluation framework to test for it.
M. Carrasco, V. A. Martirosyan, A. Mehrab, C. Netto, E. Okoyomon, and C. Graziani
New Perspectives in Graph Machine Learning, 2025
G. Bernardez, M. Montagna, L. Van Langendonck, M. Carrasco, A. Akbari, L. Cornelis, M. Papillon, P. Barlet-Ros, N. Miolane, and L. Telyatnikov
Preprint, 2025
L. Telyatnikov, G. Bernardez, M. Montagna, M. Hajij, M. Carrasco, P. Vasylenko, M. Papillon, G. Zamzmi, M. T. Schaub, J. Verhellen, P. Snopov, B. Miquel-Oliver, M. Gil-Sorribes, A. Molina, V. Guallar, T. Long, J. Suk, P. Rygiel, A. V. Nikitin, G. Escalona, M. Banf, D. Filipiak, L. Imasheva, M. Schattauer, A. L. Martinez, H. Fritze, M. Masden, V. Sanchez, M. Lecha, A. Cavallo, C. Battiloro, M. Piekenbrock, M. Tec, G. Dasoulas, N. Miolane, S. Scardapane, and T. Papamarkou
Journal of Data-centric Machine Learning Research, 2025
G. Bernárdez, L. Telyatnikov, M. Montagna, F. Baccini, M. Papillon, M. Ferriol-Galm\ś, M. Hajij, T. Papamarkou, M. S. Bucarelli, O. Zaghen, J. Mathe, A. Myers, S. Mahan, H. Lillemark, S. Vadgama, E. Bekkers, T. Doster, T. Emerson, H. Kvinge, K. Agate, N. K. Ahmed, P. Bai, M. Banf, C. Battiloro, M. Beketov, P. Bogdan, M. Carrasco, A. Cavallo, Y. Y. Choi, G. Dasoulas, M. Elphick, G. Escalona, D. Filipiak, H. Fritze, T. Gebhart, M. Gil-Sorribes, S. Goomanee, V. Guallar, L. Imasheva, A. Irimia, H. Jin, G. Johnson, N. Kanakaris, B. Koloski, V. Kovac, M. Lecha, M. Lee, P. Leroy, T. Long, G. Magai, A. Martinez, M. Masden, S. Meznar, B. Miquel-Oliver, A. Molina, A. Nikitin, M. Nurisso, M. Piekenbrock, Y. Qin, P. Rygiel, A. Salatiello, M. Schattauer, P. Snopov, J. Suk, V. S\'ćhez, M. Tec, F. Vaccarino, J. Verhellen, F. Wantiez, A. Weers, P. Zajec, B. Skrlj, and N. Miolane
Proceedings of Machine Learning Research, 2024
M. Carrasco, A. Berentzen, and A. Garcia
Geometry-grounded Representation Learning and Generative Modeling Workshop (GRaM) @ ICML, 2024
M. G. Palma, J. Gonzalez, M. Carrasco, R. Rubio-Noriega, K. Bergman, and R. Azevedo
IEEE Access, 2024

Talks & Presentations

Jul 2026
Symposium on Geometry Processing · Bern, Switzerland
Feb 2026
Graph Homomorphism Distortion: A Metric to Distinguish Them All and in the Latent Space Bind Them poster
Jornada de TDA · Seville, Spain
Aug 2025
Higher-Order Positional and Structural Encoder for Combinatorial Representations poster
ELLIS Doctoral Symposium · Warsaw, Poland
Jul 2025
Higher-Order Positional and Structural Encoder for Combinatorial Representations poster
Easter European Machine Learning Summer School · Sarajevo, Bosnia and Herzegovina
Jul 2025
London Geometry and Machine Learning (LOGML) · London, UK
Aug 2024
Learning on Graphs (LoG) · Amsterdam, Netherlands
Jul 2024
GRaM Workshop @ ICML 2024 · Vienna, Austria

Software

MANTRA Python
A benchmark dataset of triangulated manifolds for evaluating topological generalization in deep learning.
TopoBench Python
TopoBench is a Python library designed to standardize benchmarking and accelerate research in Topological Deep Learning
Code for our ICML 2024 Topological Deep Learning Challenge entry using TopoX and equivariant message passing on simplicial complexes.