TY - GEN
T1 - Experimental design in dynamical system identification
T2 - European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2014
AU - Llamosi, Artémis
AU - Mezine, Adel
AU - D'Alché-Buc, Florence
AU - Letort, Véronique
AU - Sebag, Michèle
PY - 2014/1/1
Y1 - 2014/1/1
N2 - This study focuses on dynamical system identification, with the reverse modeling of a gene regulatory network as motivating application. An active learning approach is used to iteratively select the most informative experiments needed to improve the parameters and hidden variables estimates in a dynamical model given a budget for experiments. The design of experiments under these budgeted resources is formalized in terms of sequential optimization. A local optimization criterion (reward) is designed to assess each experiment in the sequence, and the global optimization of the sequence is tackled in a game-inspired setting, within the Upper Confidence Tree framework combining Monte-Carlo tree-search and multi-armed bandits. The approach, called EDEN for Experimental Design for parameter Estimation in a Network, shows very good performances on several realistic simulated problems of gene regulatory network reverse-modeling, inspired from the international challenge DREAM7.
AB - This study focuses on dynamical system identification, with the reverse modeling of a gene regulatory network as motivating application. An active learning approach is used to iteratively select the most informative experiments needed to improve the parameters and hidden variables estimates in a dynamical model given a budget for experiments. The design of experiments under these budgeted resources is formalized in terms of sequential optimization. A local optimization criterion (reward) is designed to assess each experiment in the sequence, and the global optimization of the sequence is tackled in a game-inspired setting, within the Upper Confidence Tree framework combining Monte-Carlo tree-search and multi-armed bandits. The approach, called EDEN for Experimental Design for parameter Estimation in a Network, shows very good performances on several realistic simulated problems of gene regulatory network reverse-modeling, inspired from the international challenge DREAM7.
KW - Active learning
KW - Monte-Carlo tree search
KW - Upper Confidence Tree
KW - e-science
KW - experimental design
KW - gene regulatory network
KW - ordinary differential equations
KW - parameter estimation
U2 - 10.1007/978-3-662-44851-9_20
DO - 10.1007/978-3-662-44851-9_20
M3 - Conference contribution
AN - SCOPUS:84907048405
SN - 9783662448502
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 306
EP - 321
BT - Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2014, Proceedings
PB - Springer Verlag
Y2 - 15 September 2014 through 19 September 2014
ER -