Internet
Fact-checked

At WiseGEEK, we're committed to delivering accurate, trustworthy information. Our expert-authored content is rigorously fact-checked and sourced from credible authorities. Discover how we uphold the highest standards in providing you with reliable knowledge.

Learn more...

What Is Involved in Neural Network Programming?

G. Wiesen
G. Wiesen

Neural network programming is quite complicated and can utilize different programming languages and hardware to accomplish the creation of an artificial neural network (ANN). In general, however, this type of programming begins with the establishment of parameters that can be used to describe objects and then separate those objects into categories. Different types of input can then be fed into this system to allow the program to analyze the incoming parameters and output an indication of how the input should be categorized. Neural network programming typically repeats this process numerous times to allow the network to “learn” correct and incorrect answers for different input.

A neural network is a large network made up of individual pieces, referred to as neurons in the human brain, often emulated by those working on artificial intelligence (AI). Neural network programming is typically used to create artificial neural networks that emulate the functions of the human brain for problem solving and categorization of different objects. This programming can use different languages and syntaxes, depending on the preferences of the programmer and the overall purpose of the ANN being designed. Both hardware and software are utilized in neural network programming, with individual circuits often used to emulate the separate neurons found in biological neural networks.

Man holding computer
Man holding computer

Neural network programming can begin with the creation of the network and various parameters used in identifying different objects. Input is fed into the neural network and the program is allowed to analyze this input to determine various identifiers used in categorizing the input received. Someone might input different parameters about types of dogs, for example, such as large and small, tail or no tail, and furry or hairless. Neural network programming then involves the neural network analyzing the individual parameters to identify a particular type of dog that is being identified.

If the network identifies parameters including large, tail, and furry, for example, then it may conclude the input is meant to identify a German shepherd. If the same information caused the network to identify a Chihuahua, then the analysis would have been incorrect, and the neural network would “learn” from the mistake to properly identify the dog in the future. This is, of course, a simple example of how neural network programming works and the actual process typically involves hundreds or thousands of parameters and numerous checks by the network. Through this process, the network establishes a means for properly identifying the input in the future, allowing neural network programming to create AI systems that effectively learn from mistakes and adapt to new data.

Discuss this Article

Post your comments
Login:
Forgot password?
Register:
    • Man holding computer
      Man holding computer