Researcher 

ISSIA CNR – Genova 

SHORT BIO

Danilo Macciò (born March 30, 1980, in Genoa) received the M.Sc. Degree “summa cum laude” in Telecommunication Engineering and the Ph.D. Degree in Mathematical Engineering from the University of Genoa, Italy, in 2005 and 2009, respectively.

Currently, he works as a Researcher at the Genoa branch of the Institute of Intelligent Systems for Automation of the Italian National Research Council, where he has been involved in a series of funded research projects on diverse areas such as, e.g., fault detection of power plants, optimization of the energy chain of hybrid vehicles, prediction and decision support systems for cruise ships, etc.

His research activity focuses primarily in the areas of Learning Theory and Functional Optimization, where the main effort is dedicated to the analysis, both from a theoretical and experimental point of view, of local approximation methods based on kernel functions, combined with efficient sampling techniques (such as the ones employed in quasi-Monte Carlo integration) of the input space.
The research activity has led to the development of a method, based on semi-local approximation, for the solution of a general class of functional optimization problems, that are the paradigm of popular instances such as maximum likelihood/minimum distance estimation, dynamic programming, reinforcement learning, etc. More recently, the class of local linear regression models has been investigated by means the concept of discrepancy, a measure of uniformity of a set of points, leading to a unified analysis of the case in which the data are generated according to a given probability density function, and the case in which the data are  selected through a deterministic algorithm.
The proposed methodologies have successfully been applied to a number of cases of study such as, e.g., learning the control of a complex plant, approximate dynamic programming, inventory forecasting, etc.

He serves as a teaching assistant for the course in Operations Research held at the University of Genoa.

JOURNAL PUBLICATIONS

  1. C. Cervellera; M. Gaggero; D. Macciò

    Lattice point sets for state sampling in approximate dynamic programming, Optimal Control Applications & Methods, DOI: 10.1002/oca.2325, March 2017

  2. Cervellera C.; Maccio D.

    Low-Discrepancy Points for Deterministic Assignment of Hidden Weights in Extreme Learning Machines, IEEE Transactions on Neural Networks and Learning Systems, Vol. 27, No. 4, 2016

  3. Maccio D.

    Local linear regression for efficient data-driven control, Knowledge-based systems, Vol. 98, 2016

  4. Cervellera C.; Macciò D.

    F-Discrepancy for Efficient Sampling in Approximate Dynamic Programming, IEEE Transactions on Cybernetics, DOI 10.1109/TCYB.2015.2453123, 2016

  5. C., Cervellera; D., Macciò

    Local Linear Regression for Function Learning: An Analysis Based on Sample Discrepancy, IEEE Transactions on Neural Networks and Learning Systems, Vol. 25, No. 11, 2014

  6. C. Cervellera, M. Gaggero and D. Macciò

    Low-Discrepancy Sampling for Approximate Dynamic Programming with Local Approximators'', Computers and Operations Research, vol. 43, pp. 108-115, 2014

  7. C. Cervellera and D. Macciò

    Learning with Kernel Smoothing Models and Low Discrepancy Sampling, IEEE Transactions on Neural Networks and Learning Systems, vol. 24, no. 3, pp. 504-509, 2013

  8. C. Cervellera, M. Gaggero, D. Macciò

    Efficient Kernel Models for Learning and Approximate Minimization Problems, Neurocomputing, vol. 97, pp. 74-85, 2012

  9. D. Macciò and C. Cervellera

    Local Models for Data-Driven Learning of Control Policies for Complex Systems, Expert Systems with Applications, vol. 39, pp. 13399-13408, 2012

  10. C. Cervellera and D. Macciò

    A comparison of global and semi-local approximation in T-stage stochastic optimization, European Journal Of Operational Research, vol. 208, no. 2, pp. 109-118, 2011

  11. C. Cervellera and D. Macciò

    A numerical method for minimum distance estimation problems, Journal of Multivariate Analysis, vol. 102, pp. 789-800, 2011

  12. C. Cervellera, D. Macciò and M. Muselli

    Functional Optimization through Semi-Local Approximate Minimization, Operations Research, vol. 58, pp. 1491-1504, 2010

  13. A. Alessandri; C. Cervellera; D. Macciò; M. Sanguineti

    Optimization based on quasi-Monte Carlo sampling to design state estimators for non-linear systems, Optimization, Vol. 59, 2010

  14. C. Cervellera; D. Macciò; M. Muselli

    Functional optimization through semilocal approximate minimization, Operations research, Vol. 58, No. 5, 2010

  15. C. Cervellera; D. Macciò; M. Muselli

    Efficient global maximum likelihood estimation through kernel methods, Neural Networks, Vol. 23, No. 7, 2010

CONFERENCE PUBLICATIONS

  1. C. Cervellera; M. Gaggero; D. Macciò; R. Marcialis

    Lattice point sets for efficient kernel smoothing models, International Joint Conference on Neural Networks (IJCNN2015), Killarney, Ireland, 2015

  2. C. Cervellera; M. Gaggero; D. Maccio; R. Marcialis

    Efficient use of Nadaraya-Watson models and low-discrepancy sequences for approximate dynamic programming, International Joint Conference on Neural Networks (IJCNN2015), Killarney, Ireland, 2015

  3. Cervellera C.; Gaggero M.; Maccio D.; Marcialis R.

    (2014) "An approach to exploit non-optimized data for efficient control of unknown systems through neural and kernel models", IEEE International Joint Conference on Neural Network proceedings, Beijing, July 6-11, 2014

  4. Cervellera C.; Gaggero M.; Maccio D.

    (2014) "An analysis based on F-discrepancy for sampling in regression tree learning", IEEE International Joint Conference on Neural Network proceedings, Beijing, China, July 6-11, 2014

  5. Cervellera C.; Gaggero M.; Maccio D.; Marcialis R.

    Lattice sampling for efficient learning with Nadaraya-Watson local models, Proceedings of the International Joint Conference on Neural Networks, Beijing, China, July 6-11, 2014

  6. C. Cervellera; V. C P. Chen; D. Macciò

    "Global and semilocal estimation in multistage optimal control", Informs Annual Meeting, Charlotte, Novembre, 10-13, 2011

  7. C. Cervellera, D. Macciò and M. Muselli

    (2010)Semi-local approximation for the efficient solution of functional optimization problemsin INFORMS 2010 Annual Meeting, Austin, TX, U.S.A

  8. A. Alessandri, C. Cervellera, D. Macciò and M. Sanguineti

    (2009) Design of estimators via optimization based on quasi-Monte Carlo sampling, in AIRO Winter 2009, Cortina d'Ampezzo