Alessandro Favero

Alessandro Favero is a Researcher and Doctoral Assistant at the École Polytechnique Fédérale de Lausanne (EPFL), affiliated with the Physics of Complex Systems Laboratory (PCSL) and the Signal Processing Laboratory 4 (LTS4). He is pursuing a Doctoral Program in Physics within the School of Basic Sciences (SB). His research focuses on the theoretical foundations of machine learning, particularly in diffusion models, neural network architectures, and computational complexity.

Key research interests include the study of hierarchical data structures, compositional generalization in models, and the interplay between model architecture and learning dynamics. His work bridges machine learning and statistical physics, analyzing phenomena such as phase transitions in diffusion models and the behavior of infinitely-wide networks.

Recent contributions explore topics like overparameterization effects, layer scaling techniques, and multimodal hallucination control. His publications emphasize understanding fundamental limitations and empirical phenomena in deep learning systems, with applications ranging from natural language processing to image analysis.

While no awards or grants are explicitly listed, his active involvement in multiple EPFL labs highlights collaborative research efforts. He maintains an academic presence through his GitHub profile and the PCSL lab website.