The term memory-enhanced neural networks originates from the fields of artificial intelligence, big data and smart data as well as Industry and Factory 4.0. This technology combines the performance of classic neural networks with additional memory capabilities - similar to a computer that not only calculates but also remembers things.
Memory-enhanced neural networks enable machines to retain and utilise information over longer periods of time. This enables them to make better decisions and solve more complex tasks. A normal neural network quickly forgets what it has previously learnt. With a "memory", however, the network can recognise connections that exist over many steps.
An illustrative example: Imagine quality control in a factory. A simple neural network may recognise faulty products based on images. However, a memory-enhanced network also remembers which errors have occurred more frequently in the past and can deduce from this whether a machine should be serviced before major problems occur.
This makes processes more efficient, minimises downtime and saves costs. Memory-enhanced neural networks are therefore an important building block for intelligent, adaptive systems in the digital industry.
Deutsch 
English (UK)