Neural Network

Neural Network

An artificial neural network (ANN) is a computational method inspired in biologic processes to solve problems that are very difficult for computers or humans. An ANN can be seen as a black box with inputs and outputs as shown below. Artificial Neural Networks are special because they can adjust themselves to perform a specific task. One of the key features of ANN is that they can adapt to any kind of problems where a mathematic equation or model is missing.

Problem 1
ANNs are used for weather prediction. In some cases, they are used as tools to assist on medical diagnosis. Search the Internet for three more applications where ANNs can be used.


Neural Network Software

An ANN can be implemented using software or hardware. Neural Lab and Matlab are popular software products used to design and simulate artificial neural networks. They provide a set of commands and a suitable environment to solve ample sets of problems. Although, it is possible to implement an ANN using hardware, most ANN simulations are performed using software.

Problem 2
Mention one advantage and one disadvantage of using hardware to implement an ANN.

The input of an ANN can be any kind of information: an image, a sound, a temperature value, etc. The output of an ANN is always dependent upon its input. That is, a specific input will produce a specific output. The purpose of an ANN is to extract, map, classify or identify some sort of information that is allegedly hidden in the input.

Problem 3
Frank Rosenblatt was a psychologist who completed the Perceptron in 1960. This was the first computer that could learn new skills by trial and error, using a type of neural network that simulates human thought processes. Rosenblatt designed the first artificial neural network in 1958. Write a page about the history of ANNs focusing in the problems that Rosenblatt faced after the publication of the book written by Marvin Minsky and Seymour Papert, and how ANNs were widely accepted by the scientific community 10 years after his death.

Problem 4
Search the Internet for the types of artificial neural networks.


Some input data may not be suitable for ANN input. In these cases, it is necessary to perform some sort of transformation (pre-processing) before applying the data to the network input. Pre-processing is one of the most important steps and can affect considerably the performance of the network. Pre-processing can be performed by different procedures. The design of the pre-processing phase requires both imagination and creativity. In the figure below, the microphone generates an electric signal. The analog to digital converted (ADC) transform this electric signal into a digital signal. At the output of the ADC there is considerable amount of data, the pre-processing block in the figure has the intention of preparing the data so that the ANN can properly perform by focusing in some features of the signal.


© Copyright 2000-2019 Wintempla selo. All Rights Reserved. Sep 05 2019. Home