A Robust Online Sequential Extreme Learning Machine

Hoang, Minh-Tuan T.; Huynh, Hieu Trung; Vo, Nguyen; Won, Yonggwan

doi:10.1007/978-3-540-72383-7_126

Cited by 13 publications

(9 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…But in the implementation of the GA-oselm algorithm, these parameters do not need manual settings. Because the genetic algorithm simulates natural selection and natural genetic processes in replication, crossover and mutation process, a set of candidate solutions is retained in each iteration [13] [14]. The genetic operators (selection, crossover and mutation) are used to generate a new generation of candidate solutions.…”

Section: The Selection Of Os-elm and Genetic Algorithm Algorithm Paramentioning

confidence: 99%

Optimizing Online Sequential Extreme Learning Machine Parameters and Application to Transformer Fault Diagnosis

Kang¹,

Chen²

2015

Proceedings of the 4th International Conference on Mechatronics, Materials, Chemistry and Computer Engineering 2015

View full text Add to dashboard Cite

Abstract. In order to solve the problem that the (OS-ELM) is used in the fault diagnosis of the transformer, the genetic algorithm (Algorithm Genetic) is applied to the on-line extreme learning machine, and a new method of transformer fault diagnosis is proposed. In this method, the number of hidden layer neurons of the Block L, the data set size N, and the hidden layer activation function are selected by the Algorithm Genetic optimization algorithm. Through simulation test, the fault diagnosis of transformer is 99.56%, and the test time is 0.0024 s. Compared with the optimization, the diagnostic accuracy and the test time of the transformer fault are improved obviously.

show abstract

Section: The Selection Of Os-elm and Genetic Algorithm Algorithm Paramentioning

confidence: 99%

Optimizing Online Sequential Extreme Learning Machine Parameters and Application to Transformer Fault Diagnosis

Kang¹,

Chen²

2015

Proceedings of the 4th International Conference on Mechatronics, Materials, Chemistry and Computer Engineering 2015

View full text Add to dashboard Cite

show abstract

“…The function also satisfies global Lipschitz's condition [18] which guarantees existence and uniqueness of solution to the differential equation (13).…”

Section: Lyapunov Based Parameter Update Algorithmmentioning

confidence: 99%

“…It has been reported that ELM might run into an ill-posed problem when it is over parameterized or when not properly initialized [11], [12], [13], [14]. A few attempts have been made to improve the regularization behavior or ELM [13] and OS-ELM [14].…”

Section: Stability Advantage Of L-elmmentioning

confidence: 99%

“…A few attempts have been made to improve the regularization behavior or ELM [13] and OS-ELM [14]. However, when the data is being processed 1-by-1 (as in the case of system identification), the regularization improvement suggested by [14] was not found to improve the situation.…”

Section: Stability Advantage Of L-elmmentioning

confidence: 99%

“…In spite of its known advantages (see section II), an over-parameterized ELM suffers from the ill-conditioning problem when recursive least squares type update is performed (as in OS-ELM). This results in poor regularization behavior [11], [12], [13], [14], and more importantly, leads to an unbounded growth of the model parameters and predictions. The goal of this paper is to address the above issue by developing a stable online learning algorithm for ELM using Lyapunov approach which guarantees the boundedness of the model parameters as well as predictions even when the model is over-parameterized.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A lyapunov based stable online learning algorithm for nonlinear dynamical systems using extreme learning machines

Janakiraman

2013

The 2013 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

Abstract-Extreme Learning Machine (ELM) is a promising learning scheme for nonlinear classification and regression problems and has shown its effectiveness in the machine learning literature. ELM represents a class of generalized single hidden layer feed-forward networks (SLFNs) whose hidden layer parameters are assigned randomly resulting in an extremely fast learning speed along with superior generalization performance. It is well known that the online sequential learning algorithm (OS-ELM) based on recursive least squares [1] might result in ill-conditioning of the Hessian matrix and hence instability in the parameter estimation. To address this issue, the stability theory of Lyapunov is utilized to develop an online learning algorithm for temporal data from dynamic systems and time series. The developed algorithm results in parameter estimation that is globally asymptotically stable. Simulations results of the developed algorithm compared against online sequential ELM (OS-ELM) and the offline batch learning ELM (O-ELM) show that the online Lyapunov ELM algorithm can perform online learning at reduced computation, comparable accuracy and with a guarantee on the boundedness of the estimated system.

show abstract

Advancing the incremental fusion of robotic sensory features using online multi-kernel extreme learning machine

Cao

Sun

et al. 2016

Front. Comput. Sci.

View full text Add to dashboard Cite

A Robust Online Sequential Extreme Learning Machine

Cited by 13 publications

References 5 publications

Optimizing Online Sequential Extreme Learning Machine Parameters and Application to Transformer Fault Diagnosis

Optimizing Online Sequential Extreme Learning Machine Parameters and Application to Transformer Fault Diagnosis

A lyapunov based stable online learning algorithm for nonlinear dynamical systems using extreme learning machines

Advancing the incremental fusion of robotic sensory features using online multi-kernel extreme learning machine

Contact Info

Product

Resources

About