I received my PhD in applied mathematics in February 2020 from the University of Modena and Reggio Emilia (UniMoRe), Italy, with a thesis on "A framework for the analysis of machine learning systems". While at UniMoRe, I have contributed to create the machine learning division of the high-performance real-time (HiPeRT) laboratory led by professor Marko Bertogna. I have been working on data science and machine learning for a while, fields in which I have also had some industrial experiences (Tetra Pak, Maserati).
I am interested in everything that can help me finding and understanding the patterns hidden in this world. Since I found out that boosting statistical models with parallel computers can be pretty efficient at this, I have made it my job to understand how machine learning systems work, how to combine them and how to apply them to real-world problems. My current research interests lie at the intersection of mathematical analysis, stochastic optimisation and parallel programming. In particular, I am now working on quantized neural networks, applying neural architecture search (NAS) algorithms to improve their topology and function approximation properties, and studying new algorithms to improve their learning process.
- Office: ETZ J76.2
- telephone: (+41 44 63) 384 70
- e-mail: firstname.lastname@example.org
- www: Matteo Spallanzani (ETH page)