Neural networks and Deep Learning

By Bradley Efron, Trevor J. Hastie. 2016

2,763 citations recorded by the source corpus.

3 Logistic Regression as a Neural Network 7 3.1 Derivatives (optional) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3.2 Binary classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.3 Logistic regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.4 Cost function in logistic regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 3.5 Gradient descent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 3.6 Computation graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.6.1 Derivatives with a computation graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.6.2 The chain rule for derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.7 Computation graphs for logistic regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.7.1 Gradient descent for a single training example . . . . . . . . . . . . . . . . . . . . . . . . 11

Subfields: Artificial Intelligence

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

Citations in the corpus

Quantitative measures to evaluate neural network weight initialization strategies (2017, background): “Most beginner texts today indicate that a ”good enough” weight initialization for sigmoid activation neurons is a uniformly distributed random value in the interval (−1, 1) [1][2][3].”
Predicting academic performance of students from VLE big data using deep learning models (2020, neutral): “increasing the training split leads to a more robust accuracy (Nielsen, 2015).”
Digital twin, physics-based model, and machine learning applied to damage detection in structures (2020, background): “plicating the history of performance and degradation of its twin. Some pretty well established disciplines are part of a digital twin framework; for example, ﬁnite element model [19], neural networks [5], digital signal processing [20], structural health monitoring [21], uncertainty quantiﬁcation [2”
Discovering physical concepts with neural networks (2018, methodology): “ated arbitrarily well by a feedforward network with just one hidden layer by using suciently many hidden neurons. For a mathematical statement of the result, see [68,69]. A visualization is given in [32]. Training. The weights and the biases of the neural network are not tuned by hand; instead, they”
A Survey of Deep Learning Applications to Autonomous Vehicle Control (2019, methodology): “These types of weaknesses in DNNs could be exploited and pose a security concern for any technology using DNNs.”
Quantization Networks (2019, background): “However, the problem is that it is not immediately obvious how to learn the perceptron networks [26].”
Review and Comparison of Commonly Used Activation Functions for Deep Neural Networks (2020, methodology): “Backpropagation algorithm[11], [12] multiplies the derivatives of the activation function.”
An Introductory Review of Deep Learning for Prediction Models With Big Data (2020, methodology): “for the backpropagation algorithm can be formulated as shown in Figure 9B (Nielsen, 2015).”