Publications

Below are some of my research papers. If you have questions or comments please feel free to contact me.
An up-to-date list of my publications can be found at Google Scholar here.
Back to home.

2023

• Multi-task representation learning with stochastic linear bandits. L. Cella, K. Lounici, G. Pacreau, and M. Pontil. AISTATS 2023.
• Bilevel optimization with a lower-level contraction: optimal sample complexity without warm-start. R. Grazzi, M. Pontil, and S. Salzo. J. Machine Learning Research 24(167):1-37, 2023.
• Reexamining low rank matrix factorization for trace norm regularization. C. Ciliberto, M. Pontil, and D. Stamos. Mathematics in Engineering 5(3):1-22, 2023.

2022

• Distribution Regression with Sliced Wasserstein Kernels. D. Meunier, M. Pontil, and C. Ciliberto. ICML 2022.
• Group meritocratic fairness in linear contextual bandits. R. Grazzi, A. Akhavan, J. I. T. Falk, L. Cella, and M. Pontil. NeurIPS 2022.
• Conditional meta-learning of linear representations. G. Denevi, M. Pontil, and C. Ciliberto. NeurIPS 2022.
• Learning dynamical systems via Koopman operator regression in reproducing kernel Hilbert spaces. V. Kostic, P. Novelli, A. Maurer, C. Ciliberto, L. Rosasco, and M. Pontil. NeurIPS 2022.
• A gradient estimator via L1-randomization for online zero-order optimization with two point feedback. A. Akhavan, E. Chzhen, M. Pontil, and A. Tsybakov. NeurIPS 2022.

2021

• Convergence properties of stochastic hypergradients. R. Grazzi, M. Pontil, and S. Salzo. AISTATS 2021.
• The role of global labels in few-shot classification and how to infer them. R. Wang, M. Pontil, and C. Ciliberto. NeurIPS 2021.
• A gang of adversarial bandits. M. Herbster, S. Pasteris, F. Vitale, and M. Pontil. NeurIPS 2021.

2020

• Sinkhorn barycenters with free support via Frank-Wolfe algorithm. G. Luise, S. Salzo, M. Pontil, and C. Ciliberto. NeurIPS 2020.
• On the iteration complexity of hypergradient computation. R. Grazzi, L. Franceschi, M. Pontil, and S. Salzo. ICML 2020.
• Exploiting MMD and Sinkhorn divergences for fair and transferable representation learning. L. Oneto, M. Donini, G. Luise, C. Ciliberto, A. Maurer, and M. Pontil. NeurIPS 2020.
• The advantage of conditional meta-learning for biased regularization and fine tuning. G. Denevi, M. Pontil, and C. Ciliberto. NeurIPS 2020.

2019

• Learning-to-learn stochastic gradient descent with biased regularization. G. Denevi, C. Ciliberto, R. Grazzi, and M. Pontil. ICML 2019.
• Online-within-online meta-learning. G. Denevi, D. Stamos, C. Ciliberto, and M. Pontil. NeurIPS 2019.
• Sinkhorn barycenters with free support via frank-wolfe algorithm. G. Luise, S. Salzo, M. Pontil, and C. Ciliberto. NeurIPS 2019.
• Uniform concentration and symmetrization for weak interactions. A. Maurer and M. Pontil. COLT 2019.

2018

• Incremental learning-to-learn with statistical guarantees. G. Denevi, C. Ciliberto, D. Stamos, and M. Pontil. UAI 2018.
• Empirical bounds for functions with weak interactions. A. Maurer and M. Pontil. COLT 2018.
• Empirical risk minimization under fairness constraints. M. Donini, L. Oneto, S. Ben-David, J. Shawe-Taylor, and M. Pontil. NeurIPS 2018.

2017

• Reexamining low rank matrix factorization for trace norm regularization. C. Ciliberto M. Pontil, and D. Stamos. arXiv:1706.08934.
• Forward and reverse gradient-based hyperparameter optimization. L. Franceschi, M. Donini, P. Frasconi, and M. Pontil. ICML 2017.
• Sparsity is better with stability: combining accuracy and stability for model selection in brain decoding. L. Baldassarre, M. Pontil, and J. Mourao-Miranda. Frontiers in Neuroscience 11(62), 2017.
• Consistent multitask learning with nonlinear output relations. C. Ciliberto, A. Rudi, L. Rosasco, and M. Pontil. NIPS 2017.
• Regret bounds for lifelong learning. P. Alquier, T. T. Mai, and M. Pontil. AISTATS 2017.

2016

• Mistake bounds for binary matrix completion. M. Herbster, S. Pasteris, and M. Pontil. NIPS 2016.
• Bounds for vector-valued function estimation. A. Maurer and M. Pontil. arXiv:1606.01487.
• Learning with optimal interpolation norms: properties and algorithms. C. L. Combettes, A. M. McDonald, C. A. Micchelli, and M. Pontil. arXiv:1603.09273.
• Fitting sparsity and spectral decay with the spectral (k,p)-support norm. A. M. McDonald, M. Pontil, and D. Stamos. AISTATS 2016.
• New perspectives on k-support and cluster norms. A. M. McDonald, M. Pontil, and D. Stamos. J. Machine Learning Research 17:1-38, 2016.
• The benefit of multitask representation learning. A. Maurer, M. Pontil, and B. Romera-Paredes. J. Machine Learning Research 17(81):1-32, 2016.

2015

• Predicting a switching sequence of graph labelings. M. Herbster, S. Pasteris, and M. Pontil. J. Machine Learning Research, 16:2003-2022, 2015.
• Learning with dataset bias in latent subcategory models. D. Stamos, S. Martelli, M. Nabi, A. M. McDonald, V. Murino, and M. Pontil. CVPR 2015.

2014

• Spectral k-support norm regularization. A. M. McDonald, M. Pontil, and D. Stamos. NIPS 2014.
• Large margin local metric learning. J. Bohne', Y. Ying, S. Gentric, and M. Pontil. ECCV 2014.
• An inequality with applications to structured sparsity and multitask dictionary learning. A. Maurer, M. Pontil, and B. Romera-Paredes. COLT 2014.
• Lower bounds for sparse coding. A. Maurer, M. Pontil, and L. Baldassarre. In Measures of Complexity: Festschrift for Alexey Chervonenkis (V. Vovk et al. eds.), 2014.
• New perspectives on k-support and cluster norms. A. M. McDonald, M. Pontil, and D. Stamos. arXiv:1403.1481.

2013

• A new convex relaxation for tensor completion. B. Romera-Paredes and M. Pontil. NIPS 2013.
• Regularized robust portfolio estimation. T. Evgeniou, M. Pontil, D. Spinellis, R. Swiderski, and N. Nassuphis, 2013.
• Multilinear multitask learning. B. Romera-Paredes, H. Aung, N. Bianchi-Berthouze, and M. Pontil. ICML 2013.
• Sparse coding for multitask and transfer learning. A. Maurer, M. Pontil, and B. Romera-Paredes. ICML 2013.
• On sparsity inducing regularization methods for machine learning. A. Argyriou, L. Baldassarre, C. A. Micchelli, and M. Pontil. In Empirical Inference, Festschrift in Honor of Vladimir N. Vapnik, (B. Scholkopf et al. eds.)
• Excess risk bounds for multitask learning with trace norm regularization. A. Maurer and M. Pontil. COLT 2013.
• Transfer learning to account for idiosyncrasy in face and body expressions. B. Romera-Paredes, H. Aung, M. Pontil, N. Bianchi-Berthouze, A. C. Williams, and P. Watson. FG 2013.
• Regularizers for structured sparsity. C. A. Micchelli, J. M. Morales, and M. Pontil. Advances in Computational Mathematics 38(3):455-489, 2013.

2012

• Optimal kernel choice for large-scale two-sample tests. A. Gretton, B. Sriperumbudur, D. Sejdinovic, H. Strathmann, S. Balakrishnan, M. Pontil, and K. Fukumizu. NIPS 2012.
• Conditional mean embeddings as regressors. S. Grünewälder, G. Lever, L. Baldassarre, S. Patterson, A. Gretton, M. Pontil. ICML 2012.
• Modelling transition dynamics in MDPs with RKHS embeddings. G. Lever, S. Grünewälder, L. Baldassarre, M. Pontil, and A. Gretton. ICML 2012.
• Structured sparsity models for brain decoding from fMRI data. L. Baldassarre, J. Mourao-Miranda, and M. Pontil. PRNI 2012.
• A tale of many cities: universal patterns in human urban mobility. A. Noulas, S. Scellato, R. Lambiotte, M. Pontil, and C. Mascolo. PLoS One 7(5):e37027, 2012.
• Exploiting unrelated tasks in multi-task learning. B. Romera-Paredes, A. Argyriou, N. Bianchi-Berthouze, and M. Pontil. AISTATS 2012.
• A general framework for structured sparsity via proximal optimization. L. Baldassarre, J. M. Morales, A. Argyriou, and M. Pontil. AISTATS 2012.
• Structured sparsity and generalization. A. Maurer and M. Pontil. J. Machine Learning Research, 13:671-690, 2012.
• PSICOV: precise structural contact prediction using sparse inverse covariance estimation on large multiple sequence alignements. D. T. Jones, D. W. A. Buchan, D. Cozzetto, and M. Pontil. Bioinformatics, 28(2):184-190, 2012.

2011

• Exploiting semantic annotations for clustering geographic areas and users in location-based social networks. A. Noulas, S. Scellato, C. Mascolo, and M. Pontil. SMW 2011.
• An empirical study of geographic user activity patterns in foursquare. A. Noulas, S. Scellato, C. Mascolo, and M. Pontil. ICWSM 2011.
• Efficient first order methods for linear composite regularizers. A. Argyriou, C. A. Micchelli, M. Pontil, L. Shen, and Y. Xu. arXiv:1104.1436.
• Oracle inequalities and optimal inference under group sparsity. K. Lounici, M. Pontil, A. B. Tsybakov, and S. van de Geer. Annals of Statistics 39(4):2164-2204, 2011.
• Predictions of hot spot residues at protein-protein interfaces using support vector machines. S. Lise, D. Buchan, M. Pontil, and D. T. Jones. PLoS One 6(2):e16774, 2011.

2010

• A family of penalty functions for structured sparsity. C. A. Micchelli, J. M. Morales, and M. Pontil. NIPS 2010.
• K-dimensional coding schemes in Hilbert spaces. A. Maurer and M. Pontil. IEEE Transactions on Information Theory, 56(11):5839-5846, 2010.
• On spectral learning. A. Argyriou, C. A. Micchelli, and M. Pontil. J. Machine Learning Research, 11:935-953, 2010.

2009

• Prediction of hot spot residues at protein-protein interfaces by combining machine learning and energy-based methods. S. Lise, C. Archambeau, M. Pontil, and D. T. Jones. BMC Bioinformatics, 10:365-382, 2009.
• Inferring interests from mobility and social interactions. A. Noulas, M. Musolesi, M. Pontil, and C. Mascolo. NIPS Workshop on Analyzing Networks and Learning With Graphs, 2009.
• When is there a representer theorem? Vector versus matrix regularizers. A. Argyriou, C. A. Micchelli, and M. Pontil. J. Machine Learning Research, 10:2507-2529, 2009.
• Taking advantage of sparsity in multi-task learning. K. Lounici, M. Pontil, A. B. Tsybakov, and S. van de Geer. COLT 2009.
• Empirical Bernstein bounds and sample-variance penalization. A. Maurer and M. Pontil. COLT 2009.
• Entropy conditions for Lr-convergence of empirical processes. A. Caponnetto, E. De Vito, and M. Pontil. Advances in Computational Mathematics, 30(4):355-373, 2009.

2008

• Online prediction on large diameter graphs. M. Herbster, G. Lever, and M. Pontil. NIPS 2008.
• Fast prediction on a tree. M. Herbster, M. Pontil, and S. Rojas-Galeano. NIPS 2008.
• A uniform lower error bound for half-space learning. A. Maurer and M. Pontil. ALT 2008.
• Generalization bounds for K-dimensional coding schemes in Hilbert spaces. A. Maurer and M. Pontil. ALT 2008.
• An algorithm for transfer learning in a heterogeneous environment. A. Argyriou, A. Maurer and M. Pontil. ECML 2008.
• Universal multi-task kernels. A. Caponnetto, C. A. Micchelli, M. Pontil, and Y. Ying. J. Machine Learning Research, 9:1615-1646, 2008.
• Convex multi-task feature learning. A. Argyriou, T. Evgeniou, and M. Pontil. Machine Learning 73(3):243-272, 2008.
• Online gradient descent learning algorithms. Y. Ying and M. Pontil. Foud. Comp. Math. 8(5):561-596, 2008.

2007

• A spectral regularization framework for multi-task structure learning. A. Argyriou, C. A. Micchelli, M. Pontil, and Y. Ying. NIPS 2007.
• A convex optimization approach to modeling heterogeneity in conjoint estimation. T. Evgeniou, M. Pontil, and O. Toubia. Marketing Science, 26:805-818, 2007.
• Feature space perspectives for learning the kernel. C. A. Micchelli and M. Pontil. Machine Learning, 66:297-319, 2007.
• Conditional graphical models. F. Perez-Cruz, Z. Ghahramani, M. Pontil. In Predicting Structured Data, edited by G. Bakir et al., MIT Press, 2007.

2006

• Multi-task feature learning. A. Argyriou, T. Evgeniou, and M. Pontil. NIPS 2006.
• Prediction on a graph with the perceptron. M. Herbster and M. Pontil. NIPS 2006.
• A DC-programming algorithm for kernel selection. A. Argyriou, R. Hauser, C. A. Micchelli, and M. Pontil. ICML 2006.
• Error analysis for online gradient descent algorithms in reproducing kernel Hilbert spaces. M. Pontil, Y. Ying and D.-X. Zhou. Preprint, April 2006.

2005

• Combining graph Laplacians for semi-supervised learning. A. Argyriou, M. Herbster, and M. Pontil. NIPS 2005.
• Error bounds for learning the kernel. C. A. Micchelli, M. Pontil, Q. Wu, and D.-X. Zhou. Research Note RN/05/09, Dept of Computer Science, UCL, June, 2005.
• Online learning over graphs. M. Herbster, M. Pontil, and L. Wainer. ICML 2005.
• Learning convex combinations of continuously parameterized basic kernels. A. Argyriou, C. A. Micchelli, and M. Pontil. COLT 2005.
• Learning the kernel function via regularization. C. A. Micchelli and M. Pontil. J. Machine Learning Research, 6:1099-1125, 2005.
• Learning multiple tasks with kernel methods. T. Evgeniou and C. A. Micchelli, and M. Pontil. J. Machine Learning Research, 6:615-637, 2005.
• Wide coverage natural language processing using kernel methods and neural networks for structured data. S. Menchetti, F. Costa, P. Frasconi, and M. Pontil. Pattern Recognition Letters, 26(12):1896-1906, 2005.
• On learning vector-valued functions. C. A. Micchelli and M. Pontil. Neural Computation, 17:177-204, 2005.
• Stability of randomized learning algorithms. A. Elisseeff, T. Evgeniou, and M. Pontil. J. Machine Learning Research, 6:55-79, 2005 (See also the longer version: Stability of randomized learning algorithms with an application to bootstrap methods).

2004

• Kernels for multi-task learning. C. A. Micchelli and M. Pontil. NIPS 2004.
• Regularized multi-task learning. T. Evgeniou and M. Pontil. SIGKDD 2004
• A function representation for learning in Banach spaces. C. A. Micchelli and M. Pontil. COLT 2004.
• New results on error correcting output codes of kernel machines. A. Passerini, M. Pontil, and P. Frasconi. IEEE Trans. on Neural Networks, 15(1):45-54, 2004.
• Leave-one-out error, stability, and generalization of voting combination of classifiers. A. Elisseeff, M. Pontil, and T. Evgeniou. Machine Learning, 55(1):71-97, 2004.

2003

• A note on different covering numbers in learning theory. M. Pontil. Journal of Complexity, 19:665-671, 2003.
• Leave-one-out error and stability of learning algorithms with applications. A. Elisseeff and M. Pontil. In Advances in Learning Theory: Methods, Models and Applications, NATO Science Series III: Computer and Systems Sciences, Vol. 190, J. Suykens et al. Eds., IOS press, 2003.
• Learning in reproducing kernel Hilbert spaces: a guide tour. M. Pontil. Bull. of the Italian Artificial Intelligence Association - AI*IA Notizie, 2003.
• On different ensembles of kernel machine. M. Yamana, H. Nakahara, M. Pontil, and S. Amari. ESANN 2003.
• Image representations and feature selection for multimedia database search. T. Evgeniou, M. Pontil, C. Papageorgiou, and T. Poggio. IEEE Trans. on Knowledge and Data Engineering, 15(4):911-920, 2003.
• Combining flat and structured representations for fingerprint classification with recursive neural networks and support vector machines. Y. Yao, G. Marcialis, M. Pontil, P. Frasconi, and F. Roli. Pattern Recognition, 36(2):397-406, 2003.
• Full body person recognition. C. Nakajima, M. Pontil, B. Heisele, and T. Poggio. Pattern Recognition, 36:1997-2006, 2003.

2002

• A simple algorithm for learning stable machines. S. Andonova, A. Elisseeff, T. Evgeniou, and M. Pontil. ECAI 2002.
• Regularization and statistical learning theory for data analysis. T. Evgeniou, T. Poggio, M. Pontil, and A. Verri. Computational Statistics and Data Analysis, 38:421-432, 2002.
• Support vector machines with clustering for training with very large datasets. T. Evgeniou, and M. Pontil. SETN 2002.

2001

• Categorization by learning and combining object parts. B. Heisele, T. Serre, M. Pontil, T. Vetter and T. Poggio. NIPS 2001.
• Component-based face detection. B. Heisele, T. Serre, M. Pontil, and T. Poggio. CVPR 2001.
• Fingerprint classification with combination of support vector machines. Y. Yao, P. Frasconi, and M. Pontil. AVBPA 2001.

2000

• On the noise model of support vector machine regression. M. Pontil, S. Mukherjee, and F. Girosi. ALT 2000.
• Face detection in still gray images. B. Heisele, T. Poggio, and M. Pontil. AI Memo 1687, Massachusetts Institute of Technology, May 2000.
• Statistical learning theory: a primer. T. Evgeniou, M. Pontil, and T. Poggio. Int. Journal of Computer Vision, 38(1):9-13, 2000.
• Bounds on the generalization performance of kernel machines ensembles. T. Evgeniou, L. Perez-Breva, M. Pontil, and T. Poggio. ICML 2000.
• Regularization networks and support vector machines. T. Evgeniou, M. Pontil, and T. Poggio. Advances in Computational Mathematics, 13(1):1-50, 2000.
• Feature selection for SVMs. J. Weston, S. Mukherjee, O. Chapelle, M. Pontil, T. Poggio, and V. Vapnik. NIPS 2000.

1999

• From regression to classification in support vector machines. M. Pontil, R. Rifkin and T. Evgeniou. ESANN 1999.
• On the V-gamma dimension for regression in reproducing kernel Hilbert spaces. T. Evgeniou and M. Pontil. ALT 1999.

1998

• Properties of support vector machines. M. Pontil and A. Verri. Neural Computation, 10:955-974, 1998.
• Support vector machines for 3D object recognition. M. Pontil and A. Verri. IEEE Trans. on PAMI, 20(6):637-646, 1998.