news

Apr 30, 2026 Our paper on the Theoretical Limitations of Embedding-based Link Prediction is accepted at ICML’26! We show how linearity stops KGEs from scaling to larger graphs, then break the limitation with a Mixture of Softmaxes (borrowed from the LLM literature) at a low parameter cost.
Sep 09, 2025 Our paper on Rewarding Explainability for Drug Repurposing is accepted at IJCAI’25! We use a path-based RL model to generate explanations for predictions. We do so by rewarding path properties that make good explanations 〰️ I met Susana as I supervised her for an internship at Sony, it was a fantastic collaboration!
Jun 05, 2025 I am visiting Antonio Vergari’s APRIL lab at the University of Edinburgh for a summer to work on approximate weighted model counting in neurosymbolic AI.
Sep 18, 2024 I am giving a talk at LISEDA, Catia Pesquita’s AI for Science research group in the LASIGE department, and visiting the University of Lisbon for a few days.
Sep 09, 2024 Our paper on ULLER, a general-purpose NeSy language, is accepted as a spotlight at NeSy 2024!