Learning capability and storage capacity of two-hidden-layer feedforward networks

By G. Huang. IEEE Trans. Neural Networks, 2003

823 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Mechanical Engineering

View Paper: https://www.semanticscholar.org/paper/064f1e85984b207c1eb3c53ac8b68037089b7a0b

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 Mechanical Engineering

Citations in the corpus

Application of the extreme learning machine algorithm for the prediction of monthly Effective Drought Index in eastern Australia (2015, background): “The studies of Tamura and Tateishi (1997) and Huang (2003) showed in their work that the SLFNs with randomly chosen input weights efficiently learn distinct training examples with minimum error.”
Extreme learning machine: algorithm, theory and applications (2013, methodology): “Extreme learning machine is used in the fields such as signal processing (Mao 2002), automatic control (Mao and Huang 2005), image processing (Huang et al. 2011), market analysis, aviation and aerospace and medical diagnosis (Shang et al. 2005), etc. A new human face recognition algorithm based on b”
Regularized Extreme Learning Machine (2009, background): “It has been shown that SLFNs with arbitrarily assigned input weights and with almost any nonzero activation function can universally approximate any continuous functions on any compact input sets[ 4 ][5].”
Universal Approximation using Incremental Constructive Feedforward Networks with Random Hidden Nodes (2006, methodology): “In order to learn ﬁnite training samples, Huang [5] constructs a speciﬁc class of TLFNs which need much fewer additive nodes than SLFNs by letting separate learning machines (SLFNs locally trained on different data regions) share common additive nodes in the ﬁrst hidden layer.”
Error Minimized Extreme Learning Machine With Growth of Hidden Nodes and Incremental Learning (2009, background): “is the transpose of a matrix or vector . is called the hiddenlayer output matrix of the network [ 20 ]; the th column of is the th hidden node’s output vector with respect to inputs and the th row of is the output vector of the hidden layer with respect to input .”
Evaluation of data-driven models (SVR and ANN) for groundwater-level prediction in confined and unconfined systems (2019, background): “Huang (2003) proved that the number of hidden nodes sufficient to learn Ni samples with a negligibly small error is identified as follows: whereNH denotes the maximum size of hidden nodes, and Ni and No are the size of input and output nodes, respectively.”
Sales forecasting using extreme learning machine with applications in fashion retailing (2008, background): “In Kuo and Xue [27], the authors reported that the ANNs are better than many conventional statistical forecasting methods (see [3,16] for more details).”
Extreme learning machines: a survey (2011, background): “H is called the hidden layer output matrix of the SLFN [13, 26]; the ith column of H is the ith hidden node output with respect to inputs x1; x2; .”
Extreme learning machine: Theory and applications (2006, methodology): “Furthermore Based on Theorems 2.1 and 2.2 we can propose in this section an extremely simple and efﬁcient method to train SLFNs.”
Auto-encoder based dimensionality reduction (2016, background): “Some researchers committed themselves to investigating several fundamental issues of neural networks with one or two hidden layers [8] [1] [22] [23] [3] [9] [17] [25] [21] [27].”
Review on Methods to Fix Number of Hidden Neurons in Neural Networks (2013, background): “In 2003, Huang [15] formulated the following:𝑁 ℎ = √(𝑚 + 2)𝑁+2√𝑁/(𝑚 + 2) in single hidden layer and𝑁 ℎ = 𝑚√𝑁/(𝑚 + 2) in two hidden layer.”
Power Utility Nontechnical Loss Analysis With Extreme Learning Machine Method (2008, methodology): “Simulations on numerous real-world datasets (with noise) [22], [ 25 ], [26] have shown ELM well in different cases.”
Extreme learning machine: RBF network case (2004, background): “The output of a RBF network with ˜ N kernels for an input vector x ∈ R n is given by T is the weight vector connecting the i th kernel and the output neurons and φ ( x ) is the output of the i th kernel, which is usually gaussian: T is the i th kernel’s center and σ i is its impact width.”
A novel hybrid extreme learning machine–grey wolf optimizer (ELM-GWO) model to predict compressive strength of concrete with partial replacements for cement (2020, methodology): “ELM was originated from the observations which proved that an SLFN with random weights and biases could approximate any continuous function on any compact input set [61].”
Extreme learning machine: a new learning scheme of feedforward neural networks (2004, background): “As analyzed in pervious works[4], [5], [6], if the number of hidden neurons is equal to the number of distinct training samples, Ñ = N , matrix H is square and”