Dr. Andrea CAPONNETTO

ORCID iD: 0000-0002-6311-0667
Scopus Author ID: 56572559000

Metrics for Academic and Research Impact

PhD – University of Genova, Italy

Dr Andrea Caponnetto's main research interests include Statistical Learning Theory, Inverse Problems and Empirical Processes. His recent research projects are aimed at the mathematical understanding of learning algorithms in the non-parametric framework, in the supervised, semi-supervised and multi-task settings. Another guiding line of his research activity is highlighting the strong connection between certain widely used learning algorithms and regularization techniques developed in the context of the ill-posed inverse problems.

Reports

1. De Vito E., Rosasco L., Caponnetto A., Piana M., Verri A.Representer Theorem for Convex Loss Fuctions.DISI-TR-03-13, DISI, Università di Genova, 2003.

2. De Vito E., Rosasco L., Caponnetto A., Piana M., Verri A.Minimization of Tikhonov Functional: the Continuous Setting.DISI-TR-03-14, DISI, Università di Genova, 2003.

3. Caponnetto A., Rosasco L.Non Standard Support Vector Machines and Regularization Networks. [pdf]DISI-TR-04-03, DISI, Università di Genova, 2004.

4. Caponnetto A., Carmeli C., De Vito E., Rosasco L., Toigo A.Discretization Error Analysis for Tikhonov Regularization. [pdf]DISI-TR-04-04, DISI, Università di Genova, 2004.

5. Caponnetto A., De Vito E.Fast Rates for Regularized Least-squares Algorithm. [pdf]CBCL Paper #248/AI Memo #2005-013, Massachusetts Institute of Technology, Cambridge, MA, April, 2005.

6. De Vito E., Caponnetto A.Risk Bounds for Regularized Least-Squares Algorithm with Operator-Valued Kernels. [pdf]CBCL Paper #249/AI Memo #2005-015, Massachusetts Institute of Technology, Cambridge, MA, May, 2005.

7. Caponnetto A., Rosasco L., De Vito E., Verri A.Empirical Effective Dimension and Optimal Rates for Regularized Least-Squares Algorithm. [pdf]CBCL Paper #252/AI Memo #2005-019, Massachusetts Institute of Technology, Cambridge, MA, May 2005.

8. Caponnetto A., Rakhlin A.Some Properties of Empirical Risk Minimization over Donsker Classes. [pdf]CBCL Paper #250/AI Memo #2005-018, Massachusetts Institute of Technology, Cambridge, MA, May 2005.

9. Caponnetto A.A Note on the Role of Squared Loss in Regression. [pdf]CBCL Paper, Massachusetts Institute of Technology, Cambridge, MA, June 2005.

10. Yao Y., Rosasco L., Caponnetto A.On Early Stopping in Gradient Descent Boosting.TR-2005-09, Department of Computer Science, The University of Chicago, Communicated by Partha Niyogi, 27 June 2005.

11. Caponnetto A.Optimal Rates for Regularization Operators in Learning Theory. [pdf]CBCL Paper #264/ CSAIL-TR #2006-062, Massachusetts Institute of Technology, Cambridge, MA, 2006.

12. Caponnetto A., Yao Y.Adaptation for Regularization Operators in Learning Theory. [pdf]CBCL Paper #265/ CSAIL-TR #2006-063, Massachusetts Institute of Technology, Cambridge, MA, 2006.

13. Smale S., Poggio T., Caponnetto A., Bouvrie J.Derived Distance: towards a mathematical theory of visual cortex. [pdf]CBCL Paper, Massachusetts Institute of Technology, Cambridge, MA, November, 2007.

14. Caponnetto A., Poggio T., Smale S.On a model of visual cortex: learning invariance and selectivity from image sequences. [pdf]CBCL Paper #272/ CSAIL-TR #2008-030, Massachusetts Institute of Technology, Cambridge, MA, April 4, 2008.

• Learning Theory
• Inverse Problems
• Empirical Processes