Neural Networks: A Comprehensive Foundation

By Simon Haykin. 1994

How researchers see this paper

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

In Model Architectures

Citations in the corpus

The Graph Neural Network Model (2009, neutral): “Traditional machine learning applications cope with graph structured data by using a preprocessing phase which maps the graph structured information to a simpler representation, e.g., vectors of reals [16].”
Integrating Learning and Reasoning with Deep Logic Models (2019, methodology): “A possible approximation [12, 13] relies on replacing the expected values with the corresponding value at the MAP solution, assuming that most of the probability mass of the distribution is centered around it.”
On the Global Convergence of Gradient Descent for Over-parameterized Models using Optimal Transport (2018, background): “Common choices are the sigmoid function or the rectified linear unit [18, 14], see more details in Section 4.”
A High-Throughput Screening Approach to Discovering Good Forms of Biologically Inspired Visual Representation (2009, neutral): “[1–6]), biologically-inspired models are routinely among the highest-performing artificial vision systems in practical tests of object and face recognition [7–12].”
An approach to reachability analysis for feed-forward ReLU neural networks (2017, background): “Feed-forward neural networks (FFNN) are the simplest class of neural networks [10].”
Toward an Integration of Deep Learning and Neuroscience (2016, background): “The ﬁeld has advanced from simple linear systems (Minsky and Papert, 1972), to nonlinear networks (Haykin, 1994), to deep and recurrent networks (LeCun et al., 2015; Schmidhuber, 2015).”
A survey on feature selection methods (2014, methodology): “A Radial Basis Function Network [6,77] (RBFN or RBF) is a type of feed-forward Neural Network.”
Iterative kernel principal component analysis for image modeling (2005, background): “0 and eigenvectors v i 2 IR N n f0g (cf., e.g., [21]).”
Extreme learning machine: RBF network case (2004, background): “…that ˜ =1 o j − t j = 0, i.e., there exist β i , µ i and σ i such that The above N equations can be written compactly as: where Similar to SLFNs [9], [4], H is called the hidden layer output matrix of the RBF network; the i th column of H is the output of the i th kernel with respect to inputs”
A review on design, modeling and applications of computer experiments (2006, background): “ANN models have been very popular for modeling a variety of physical relationships (for a general introduction see Lippmann (1987) or Haykin (1999), for statistical perspectives see White (1989), Barron et al. (1992), Ripley (1993), or Cheng and Titterington (1994)).”
Laplacian Eigenmaps for Dimensionality Reduction and Data Representation (2003, unknown): “Most of them, such as self-organizing maps and other neural network–based approaches (e.g., Haykin, 1999), set up a nonlinear optimization problem whose solution is typically obtained by gradient descent that is guaranteed only to produce a local optimum; global optima are difﬁcult to attain by…”
Pattern recognition in the automatic inspection of aluminium castings (2003, methodology): “There are two solutions for this problem: the number of samples of the 'regular structure' class can be reduced (but thus losing the network’s capacity to generalise), or duplicate the samples from the smaller class thus diminishing the dominant tendency of the larger class (Haykin, 1994).”
Pedestrian Detection via Classification on Riemannian Manifolds (2008, unknown): “Ç”
Recent developments in consumer credit risk assessment (2007, background): “The final function yields values which can be compared with a cut-off for classification (Haykin, 1999). A large choice of activation functions is available (logistic, exponential etc.). Each training case is submitted to the network, the final output compared with the observed value (0 or 1 in this”
FUNDAMENTALS OF NEURAL NETWORKS (2001, methodology): “The trained network is then used to apply what it has learned to approximate or predict the corresponding output [ 8 ].”
Survey of clustering algorithms (2005, background): “They are different in the meaning of an individual in the population, encoding methods, ﬁ tness function de ﬁ nition, and evolutionary operators [67], [195], [273].”