Green Internet of Things and Machine Learning by Unknown

Author:Unknown
Language: eng
Format: epub
ISBN: 9781119793120
Published: 2021-03-21T00:00:00+00:00

The single-layer network is similar to the organizing regression model, commonly used in statistical modeling The derivative of this function is easily calculated:

If its activation function is modulo 1, then it can solve XOR problem with exactly one neuron.

6.3.1.2 Multi-Layer Perceptron

In this type of networks, there are many computational units’ layers that are connected to each other in a feedforward way. Every neuron in a layer has straight connections to neurons of the next layer. The units of these networks have a sigmoid function as a starting function. It uses a variety of learning approaches, and the most popular is back-propagation. In this, the output values are then contrasted with the actual right answer to compute the value of some already defined error-function, and after that, the error is then fed back again through the network. After repetition of this process for multiple times, the network normally converges to a state where the chances of mistake are less. To alter weights correctly, one applies a general method for non-linear development that is called “gradient descent.” The issue of the back-propagation algorithm is the speed of merging and the chances of ending up in a local minimum of the error function. Today back-propagation in multi-layer perceptron’s the tool of choice for various machine learning tasks.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.