Hierarchical Extreme Learning Machines
The techniques and theories of the Extreme Learning Machines (ELM) have been developing fast with the contributions from researchers over the world in the past 10 years. ELM originally aims to fill the gap between machine learning and biological learning. ELM as a common learning mechanism may play important roles in both machine learning and biological learning irrespective of whether they are implemented in silicon, proteins or other materials. ELM theories state that as long as neurons are nonlinear piecewise continuous (even without knowing their mathematical modeling), they can be randomly generated in both artificial and biological neural networks. Random neurons may play important roles in biological learning mechanism, that is, biological learning systems are globally structured but may be locally disordered.
G. Huang et al., "Hierarchical Extreme Learning Machines," Neurocomputing, vol. 277, Elsevier, Feb 2018.
The definitive version is available at https://doi.org/10.1016/j.neucom.2017.07.067
Electrical and Computer Engineering
International Standard Serial Number (ISSN)
© 2018 Elsevier, All rights reserved.