Recurrent neural networks (RNNs) belong to the artificial intelligence category and also play an important role in big data, smart data and automation. RNNs are special forms of artificial neural networks that have been developed to deal with data that has a sequence - for example, text, sound or time series.
In contrast to classic artificial neural networks, RNNs can "memorise" information from previous steps and use it for later decisions. This enables them to recognise connections that are not immediately obvious. This makes RNNs particularly useful for tasks such as speech and text recognition, translations or even predicting share prices.
A striking example: when typing in a smartphone message, the device often suggests how a sentence could continue. Behind this is often a recurrent neural network that has learnt patterns from previous text input and suggests the next words based on this.
In this way, RNNs help companies to automate everyday processes, make services smarter and help to analyse large volumes of data in a meaningful way.