### Resumen

The Mixture of Experts (ME) model is a type of modular artificial neural network (MANN) specially suitable when the search space is stratified and whose architecture is composed by different kinds of networks which compete to learn several aspects of a complex problem. Training a ME architecture can be treated as a maximum likelihood estimation problem, where the Expectation Maximization (EM) algorithm decouples the estimation process in a manner that fits well with the modular structure of the ME architecture. However, the learning process relies on the data and so is the performance. When the data is exposed to outliers, the model is affected by being sensible to these deviations obtaining a poor performance as it is shown in this work. This paper proposes a Robust Expectation Maximization algorithm for learning a ME model (REM-ME) based on M-estimators. We show empirically that the REM-ME for these architectures prevents performance deterioration due to outliers and yields significantly faster convergence than other approaches.

Idioma original | English |
---|---|

Páginas (desde-hasta) | 238-245 |

Número de páginas | 8 |

Publicación | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |

Volumen | 2686 |

Estado | Published - 2003 |

### Huella dactilar

### ASJC Scopus subject areas

- Computer Science(all)
- Biochemistry, Genetics and Molecular Biology(all)
- Theoretical Computer Science

### Citar esto

*Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)*,

*2686*, 238-245.

}

*Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)*, vol. 2686, pp. 238-245.

**Robust expectation maximization learning algorithm for mixture of experts.** / Torres, Romina; Salas, Rodrigo; Allende, Hector; Moraga, Claudio.

Resultado de la investigación: Article

TY - JOUR

T1 - Robust expectation maximization learning algorithm for mixture of experts

AU - Torres, Romina

AU - Salas, Rodrigo

AU - Allende, Hector

AU - Moraga, Claudio

PY - 2003

Y1 - 2003

N2 - The Mixture of Experts (ME) model is a type of modular artificial neural network (MANN) specially suitable when the search space is stratified and whose architecture is composed by different kinds of networks which compete to learn several aspects of a complex problem. Training a ME architecture can be treated as a maximum likelihood estimation problem, where the Expectation Maximization (EM) algorithm decouples the estimation process in a manner that fits well with the modular structure of the ME architecture. However, the learning process relies on the data and so is the performance. When the data is exposed to outliers, the model is affected by being sensible to these deviations obtaining a poor performance as it is shown in this work. This paper proposes a Robust Expectation Maximization algorithm for learning a ME model (REM-ME) based on M-estimators. We show empirically that the REM-ME for these architectures prevents performance deterioration due to outliers and yields significantly faster convergence than other approaches.

AB - The Mixture of Experts (ME) model is a type of modular artificial neural network (MANN) specially suitable when the search space is stratified and whose architecture is composed by different kinds of networks which compete to learn several aspects of a complex problem. Training a ME architecture can be treated as a maximum likelihood estimation problem, where the Expectation Maximization (EM) algorithm decouples the estimation process in a manner that fits well with the modular structure of the ME architecture. However, the learning process relies on the data and so is the performance. When the data is exposed to outliers, the model is affected by being sensible to these deviations obtaining a poor performance as it is shown in this work. This paper proposes a Robust Expectation Maximization algorithm for learning a ME model (REM-ME) based on M-estimators. We show empirically that the REM-ME for these architectures prevents performance deterioration due to outliers and yields significantly faster convergence than other approaches.

KW - Expectation Maximization

KW - Mixtures of Experts

KW - Modular Neural Networks

KW - Robust Learning Algorithm

UR - http://www.scopus.com/inward/record.url?scp=21644462102&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:21644462102

VL - 2686

SP - 238

EP - 245

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -