From June 8 to 10, part of our team (Daniel Fuertes, David Aguado, and Leyre Encío) attended the 9th edition of the Graph Signal Processing Workshop (GSP 2026), held at the Quintana campus of Universidad Rey Juan Carlos and co-organized by the ELLIS units of Madrid and Delft. This event has established itself as a key meeting point for experts in Graph Signal Processing (GSP) and Graph Machine Learning, covering topics ranging from theoretical developments to applications in biological, social, and telecommunications networks.

We had the opportunity to play a very active role in the scientific program through different formats. On the one hand, David delivered an oral presentation in which he presented his work on the effectiveness of pretraining for graph combinatorial optimization, titled “On the Effectiveness of Pretraining for Graph Combinatorial Optimization.” On the other hand, the poster session featured double representation from our team: Leyre presented her research on improving Transformer-based routing with the work “Enhancing Transformer-Based Routing by Encoding Distance via Relative Positional Encoding,” while Daniel shared his advances toward breaking the speed–accuracy trade-off with “Shattering the Speed–Accuracy Dichotomy in Asymmetric Routing via Anisotropic GNNs.”

In addition to presenting our work, it was a privilege to attend the workshop’s plenary talks, with a special highlight being the keynote delivered by Antonio Ortega, one of the world’s leading experts in this field and a former colleague from the Grupo de Tratamiento de Imágenes.