Parameter sharing is a term from the fields of artificial intelligence, big data, and smart data, as well as industry and Industry 4.0. It refers to a method where several parts of a system use the same settings, instead of storing separate parameters for each unit. This leads to less memory requirement and faster calculation.
A vivid example of this can be found in artificial neural networks, which are used, for example, in image analysis. Here, many image areas can be examined according to the same pattern. The learning system therefore repeatedly „shares“ the same parameters in order to recognise certain shapes or patterns. This makes the neural network considerably more efficient, requires fewer resources and works faster – a great advantage, especially when enormous amounts of data need to be analysed.
Parameter sharing thus helps to make artificial intelligence and automation cheaper and more scalable. Costs and energy can be saved, particularly in manufacturing halls or in real-time analyses of large image data. For companies, this means: Parameter sharing keeps high-tech solutions affordable and flexibly deployable.
Deutsch 
English (UK)