A Survey on Extreme Learning Machine and Evolution of Its Variants

Subhasmita Ghosh, Himadri Mukherjee, Sk Md Obaidullah, Kc Santosh, Nibaran Das, Kaushik Roy

Abstract

Extreme Learning Machine (ELM) is most popular emerging learning algorithm that modify classical ‘Generalized’ single hidden layer feed forward network. Though some traditional gradient based learning algorithm like variant Levenberg-Marguardt (LM) and Back propagation (BP) are widely utilized for training in multi layer FFNN but some drawbacks of this mechanism are the most prime issue to promote ELM. It imparts efficient learning solutions for different practices of classification and regression under supervise learning. ELM sharply deals with the messes arise from the gradient based learning algorithm like stopping criteria, learning rate, learning epoch, local minimum etc. But due to some fallibility the different concepts of variants of ELM are presented. This paper clarifies about Extreme Learning Machine along with different types of variants.
BibTeX
@inproceedings{ghosh2018survey,
  author = {Subhasmita Ghosh and Himadri Mukherjee and Sk Md Obaidullah and KC Santosh and Nibaran Das and Kaushik Roy},
  title = {{A Survey on Extreme Learning Machine and Evolution of Its Variants}},
  booktitle = {RTIP2R},
  year = {2018},
  doi = {10.1007/978-981-13-9181-1\_50},
}

Loading related content...

Figure