Transfer learning via modalities is particularly at home in the fields of artificial intelligence and Industry 4.0. When machines or computers learn, there are often different types of data, so-called modalities - for example images, texts or sounds. Transfer learning via modalities is about transferring knowledge from one type of data to another.
This means that an AI system that has learnt a lot of information from images can use this knowledge if it suddenly has to work with text or sounds, even though there is only a small amount of data available. This makes the learning machine more flexible and enables it to master new tasks more quickly.
An illustrative example: a robot in a factory previously only recognised dangerous situations using a camera. Through transfer learning via modalities, it can use the patterns it has learnt to also recognise danger through noise detection or vibration sensors. This saves time and costs in development and makes the use of AI much more versatile.
Transfer learning via modalities therefore offers enormous advantages when different data sources come together in modern companies. It supports innovation and ensures that AI can be used even more intelligently.
Deutsch 
English (UK)