Learning by Stretching Deep Networks

23 citations recorded by the source corpus.

How researchers see this paper

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

Deep Fried Convnets (2014, methodology): “…and we use Rectified Linear units (ReLU’s), φ(x) = √ 1/nmax(0,Wx), (3) we obtain the rotationally invariant arc-cosine kernel introduced in Cho & Saul (2009); Pandey & Dukkipati (2014) k(x,x′) = Ew[φ(x)>φ(x′)] = 1 2π ||x||||y||(sin(θ) + (π − θ) cos(θ)), (4) where θ is the angle between x and x′.”
Convolutional Gaussian Processes (2017, background): “Computational issues Similar kernels have been proposed in various forms [29, 30], but have never been applied directly in GPs, probably due to the prohibitive costs.”
Nonparametric Neural Networks (2017, background): “Pandey & Dukkipati (2014) transform a trained small network into a larger network by multiplying weight matrices with large, random matrices.”
A Survey of Machine Learning Techniques Applied to Software Defined Networking (SDN): Research Issues and Challenges (2019, background): “In recent years, the development of hardware data processing capabilities (e.g., GPU and TPU) and the evolved activation functions (e.g., ReLU) make it possible to train deep NNs [96].”