IIT People Search

Research center

Saverio Salzo received a MSc degree in (pure) Mathematics from the University of Bari in 2001 (summa cum laude) and a PhD in Computer Science from the University of Genova in 2012. His main research interests are in nonsmooth optimization, proximal splitting methods, stochastic algorithms, hyperparameter optimization, bilevel optimization, optimization in probability spaces, support vector machines in Banach spaces, and tensor kernel methods. From 2016 to 2018 he was member of the Laboratory for Computational and Statistical Learning (LCSL), which is a joint initiative between the Italian Institute of Technology and the Massachusetts Institute of Technology. Since 2020 he is also honorary lecturer at University College London (UCL).

All Publications
Salzo S., Villa S.
Parallel random block-coordinate forward–backward algorithm: a unified convergence analysis
Mathematical Programming, Series B, vol. 193, (no. 1), pp. 225-269
Salzo S., Villa S.
Proximal Gradient Methods for Machine Learning and Imaging
Harmonic and Applied Analysis. From Radon Transforms to Machine Learning, Publisher: Birkhäuser
Salzo S., Suykens J.A.K.
Generalized support vector regression: Duality and tensor-kernel representation
Analysis and Applications, vol. 18, (no. 1), pp. 149-183
Grazzi R., Franceschi L., Pontil M., Salzo S.
On the iteration complexity of hypergradient computation
37th International Conference on Machine Learning, ICML 2020, vol. PartF168147-5, pp. 3706-3716
Conference Paper Conference
Frecon J., Salzo S., Pontil M.
Unveiling groups of related tasks in multi-task learning
Proceedings - International Conference on Pattern Recognition, pp. 7134-7141
Conference Paper Conference