Biased Generalization in Diffusion Models
ICML 2026 · Spotlight, top 2.2%
I’ve just graduated cum laude from Bocconi University’s MSc in Artificial Intelligence (Computer Science, class LM-18), with the thesis “Bias and Memory in Diffusion Models: Training-Data Dependence and Image Reconstruction through Learned Conditioning”, supervised by Prof. Luca Biggio.
I recently studied memorisation and generalisation in diffusion models, and generation efficiency in diffusion language models.
I also recently worked as a technical consultant at Silverstream AI, doing AI agent evaluation.
I’m looking for roles in machine learning research and/or engineering, and applying to PhD programmes starting in 2027.
Updated

ICML 2026 · Spotlight, top 2.2%
NeurIPS 2026 DiffuLM Workshop · Poster
The Count of Monte Cristo
The Wind Rises
Paris and Chantilly