Extreme learning machine: Theory and applications

By G. Huang, Qin-Yu Zhu, Chee Kheong Siew. 2006

12,768 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Electrical and Electronic Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Electrical and Electronic Engineering

Citations in the corpus

EMNIST: an extension of MNIST to handwritten letters (2017, methodology): “A simple three-layer ELM network, as described in [20], was used to perform the classiﬁcation and it is expected that deeper and more sophisticated networks will provide better classiﬁcation performance.”
On the Equivalence between Kernel Quadrature Rules and Random Feature Expansions (2015, background): “…of two random one-dimensional (typically non-linear) features computed on these two points, the full kernel (and hence its feature space) may be approximated by sufﬁciently many random samples, replacing the expectation by a sample average (Neal, 1995; Rahimi and Recht, 2007; Huang et al., 2006).”
No bad local minima: Data independent training error guarantees for multilayer neural networks (2016, background): “However, such wide networks are easy to optimize, since by training the last layer we get to a global minimum (zero training error) from almost every random initialization [11, 16, 23].”
Fix your classifier: the marginal value of training the last weight layer (2018, background): “Past work noted the fact that predefined (Park & Sandberg, 1991) and random (Huang et al., 2006) projections can be used together with a learned affine transformation to achieve competitive results on several tasks.”
Exponentially vanishing sub-optimal local minima in multilayer neural networks (2017, background): “For any y ∈ { 0 , 1 N and X ∈ R d 0 × N almost everywhere 4 we ﬁnd matrices (Nilsson, 1965; Baum, 1988; Yu, 1992; Huang et al., 2006; Livni et al., 2014; Shen, 2016)), while Theorem 8 here requires much less: approximately d ∗ 1 ≈ 2 N/d 0 .”
NPENAS: Neural Predictor Guided Evolution for Neural Architecture Search (2022, methodology): “Besides the above search strategies, the extreme learning machine (ELM) [37], [38] and broad learning system (BLS) [39] can also be used to design the structure of neural networks automatically.”
On the Equivalence between Quadrature Rules and Random Features (2015, background): “If a certain positive definite kernel between two points may be represented as the expectation of the product of two random one-dimensional (typically non-linear) features computed on these two points, the full kernel (and hence its feature space) may be approximated by sufficiently many random samp”
Machine Learning Algorithms in Civil Structural Health Monitoring: A Systematic Review (2020, methodology): “For example, ELM was first proposed by in [52–55] as a tool that is faster in the training phase, which may result in better interpolation, but did not necessarily produce more precise and accurate results.”
Multifractal analysis of the time series of daily means of wind speed in complex regions (2017, methodology): “The choice of ELM is motivated by the fact that it is fast (with only one parameter to optimise, which is the number of hidden nodes N̂), gives accurate predictions, and it was proven to be a universal modelling algorithm [35].”
Flood Prediction Using Machine Learning Models: Literature Review (2018, methodology): “Back-percolation learning algorithms are used to individually calculate the propagation error in hidden network nodes for a more advanced modeling approach.”
Neural learning of stable dynamical systems based on data-driven Lyapunov candidates (2013, background): “The figure shows a single hidden layer feed-forward neural network called extreme learning machine (ELM) [18] comprising three layers of neurons: x ∈ Rd denotes the input, h ∈ RR the hidden, and v̂ ∈ Rd the output neurons.”
Human face recognition based on multidimensional PCA and extreme learning machine (2011, methodology): “[34] proposed ELM to train a single-hidden layer feedforward neural network (SLFNN) as shown in Fig.”
Market impact analysis via deep learned architectures (2018, methodology): “, the internal weights and hidden layer biases are randomly assigned by ELM instead of a fully tuning process of all the internal parameters [5].”
Hierarchical Coding Vectors for Scene Level Land-Use Classification (2016, background): “…(HA) [12], Soft Assignment (SA) [6], Local Coordinate Coding (LCC) [13], Locality-constrained Linear Coding (LLC) [14], Vector of Locally Aggregated Descriptors (VLAD) [15], FV [7]), and (3) more complex classiﬁers (e.g., Support Vector Machine (SVM) [16], Extreme Learning Machine (ELM) [17]).”
Deep Convex Net: A Scalable Architecture for Speech Pattern Classification (2011, positive): “While the basic module of the DCN reported in this paper is similar to the extreme learning machine in the literature (e.g., [14]), any simple or weak classifier can be embedded in the DCN architecture to make it stronger.”