Model pruning is a term used in the fields of artificial intelligence, big data and smart data as well as Industry and Factory 4.0. It describes a method of "slimming down" artificial intelligence models, such as neural networks. This involves removing parts of the model that contribute little to accuracy so that the model works faster and more efficiently.
One advantage of model pruning is that it requires less computing power and less memory. This is particularly useful if such AI models are to be used on devices that are not particularly powerful, such as smartphones, sensors or small machines in Industry 4.0.
An illustrative example: Imagine you have a large torch with many additional functions. But if you only need light, you remove everything unnecessary - the torch becomes lighter, requires less power and is easier to use. In the same way, model pruning in AI applications ensures that the important functions are retained, but the unnecessary "additional elements" are removed.
Model Pruning helps companies to implement AI solutions cost-effectively, quickly and in a resource-saving manner - an important step towards efficient digital transformation.
Deutsch 
English (UK)