Training data ecosystems are particularly at home in the areas of artificial intelligence, big data, smart data and digital transformation. The term describes the interplay of all players, technologies and processes involved in the collection, management and utilisation of training data. Training data are the data sets with which artificial intelligence "learns" to deliver better results.
Imagine a company wants to develop an AI that can distinguish cats from dogs in photos. To do this, it needs as many diverse and high-quality photos of both animals as possible. The training data ecosystem now includes the sources of these images (e.g. online galleries or private collections), the people who correctly label the images and the IT systems that securely store and provide this data.
A well-functioning training data ecosystem ensures that artificial intelligence learns from reliable, up-to-date and representative data. Without properly structured and diverse training data, many AI applications could work incorrectly or even dangerously. This is why the development of powerful training data ecosystems is a key issue in the digital transformation and further development of AI.
Deutsch 
English (UK)