Yet Another Home Page

Matteo Gagliolo

IDSIA - Room F205
Galleria 2
6928 Manno-Lugano
Switzerland - CH

Tel. +41 58 666 65 68
Fax +41 58 666 66 61

matteo@idsia.ch

Ciao! I'm a PhD student at IDSIA and at the Faculty of Informatics of the University of Lugano, supervised by Jürgen Schmidhuber. I was recently hosted at IRIDIA (Université Libre de Bruxelles), and cooperated with the Statistics Institute of Université Catholique de Louvain, supported by an SNF grant for prospective researchers. My research activity was previously funded by the Hasler Foundation, with the project "Distributed Algorithm Portfolios", and by the SNF, under the project "General Methods for Search and Reinforcement Learning".

Events

Doctoral Symposium on Engineering Stochastic Local Search Algorithms (SLS-DS). September 3-5 2009, Brussels, Belgium.

Research interests

Static and dynamic algorithm portfolios, restart strategies, meta-learning, algorithm selection, parameter tuning, search in program space.
Survival analysis methods for algorithm performance modeling.
Bandit problem formulation of algorithm selection.
Theoretical foundations of model selection, denoising and regularization techniques: algorithmic complexity, statistical learning theory, information theory.
Evolutionary computation and reinforcement learning.
Recurrent artificial neural networks, iterated function systems, neuroevolution.

Publications (BibTeX)

Gagliolo, M., Legrand, C., Algorithm Survival Analysis. Invited book chapter, to appear in: T. Bartz-Beielstein et al., eds., Empirical Methods for the Analysis of Optimization Algorithms, Springer, Berlin.
Gagliolo, M., Legrand, C., Birattari, M., Mixed-Effects Modeling of Optimisation Algorithm Performance. In T. Stützle et al., eds., Engineering Stochastic Local Search Algorithms. SLS 2009, pp. 150-154, Springer LNCS 5752, Berlin, 2009.
Gagliolo, M. Schmidhuber, J., Towards Distributed Algorithm Portfolios. In J. M. Corchado et al., eds., International Symposium on Distributed Computing and Artificial Intelligence (DCAI 2008), pp. 634-643, Springer ASC 50, Berlin, 2008 (pdf - talk slides, pdf).
Gagliolo, M. Schmidhuber, J., Algorithm Selection as a Bandit Problem with Unbounded Losses. Tech. report IDSIA - 07 - 08, July 2008 (pdf - arXiv). To appear in LION4 Proceedings, Springer LNCS, 2010.
Gagliolo, M. Universal Search. Scholarpedia, 2(11):2575, 2007.
Gagliolo, M. Schmidhuber, J., Learning restart strategies. In M. Veloso, ed., IJCAI 2007 - Twentieth International Joint Conference on Artificial Intelligence - January 6-12, Hyderabad, India, vol. 1, pp.792-797, AAAI Press (pdf - talk slides, pdf).
Gagliolo, M., Schmidhuber, J., Learning Dynamic Algorithm Portfolios. AI & MATH 2006 Special Issue of the Annals of Mathematics and Artificial Intelligence, 47(3-4):295-328, August 2006, (Springer link - techrep, pdf). Presented at LION 2007 (talk slides, pdf).
Gagliolo, M. Schmidhuber, J., Gambling in a Computationally Expensive Casino: Algorithm Selection as a Bandit Problem. Online Trading of Exploration and Exploitation - NIPS 2006 Workshop, Whistler, British Columbia, Canada, December 2006 (pdf - poster, pdf).
Gagliolo, M. Schmidhuber, J., Impact of censored sampling on the performance of restart strategies. In F. Benhamou, ed., Principle and Practice of Constraint Programming - CP 2006, 12th International Conference, CP 2006, Nantes, France, September 25-29, 2006, Proceedings, pp. 167-181, Springer LNCS 4204, Berlin, 2006 (pdf - talk slides, pdf).
Gagliolo, M., Dynamic Meta-Learning. AI50 - 50th Anniversary Summit of Artificial Intelligence, Monte Verità, Switzerland, July 2006 - Summit proceedings (abstract, pdf - poster, pdf).
Schmidhuber, J., Wierstra, D., Gagliolo, M. Gomez, F., Training Recurrent Networks by Evolino. Neural Computation, 19(3), pp. 757-779, March 2007 (pdf).
Schmidhuber, J., Gagliolo, M., Wierstra, D., Gomez, F., Evolino for Recurrent Support Vector Machines. In M. Verleysen, ed., ESANN 2006, 14 th European Symposium on Artificial Neural Networks, Bruges, April 2006 - pp. 593-598, d-side, Evere, 2006. Techrep version available (pdf - arXiv - www).
Gagliolo, M., Schmidhuber, J., Dynamic Algorithm Portfolios. AI & MATH '06, Ninth International Symposium on Artificial Intelligence and Mathematics, Fort Lauderdale, Florida, January 2006 (pdf - talk slides, pdf).
Gagliolo, M., Schmidhuber, J., Towards Life-Long Meta Learning. Inductive Transfer : 10 Years Later - NIPS 2005 Workshop, Whistler, British Columbia, Canada, December 2005 (pdf).
Gagliolo, M., Schmidhuber, J., A Neural Network Model for Inter-Problem Adaptive Online Time Allocation. In W. Duch et al., eds., Artificial Neural Networks: Formal Models and Their Applications - ICANN 2005, 15th Int. Conf., Warsaw, Proceedings, Part 2, pp. 7-12, Springer LNCS 3697, Berlin, 2005 (pdf - talk slides, pdf).
Gagliolo, M., Zhumatiy, V., Schmidhuber, J., Adaptive Online Time Allocation to Search Algorithms. In J.F. Boulicaut et al., eds., Machine Learning: ECML 2004. Proceedings of the 15th European Conference on Machine Learning, Pisa, Springer LNCS 3201, Berlin, 2004 (techrep, ps.gz - talk slides, pdf).
Schmidhuber, J., Zhumatiy, V., Gagliolo, M., Bias-Optimal Incremental Learning of Control Sequences for Virtual Robots. In F. Groen et al., eds., Proc. 8th Conference on Intelligent Autonomous Systems IAS-8, pp. 658-665, IOS Press, Amsterdam, 2004.
Anguita, D., Gagliolo, M., MDL Based Model Selection for Relevance Vector Regression. In J. Dorronsoro, ed., Proceedings of Int. Conf. on Artificial Neural Networks (ICANN'02), Madrid, pp. 468-473, Springer LNCS 2415, Berlin, 2002 (pdf).

Last updated 1.3.2010.