Meta-learning falls into the Artificial Intelligence category and also plays an important role in the fields of Big Data and Smart Data as well as automation.
Meta-learning translates to „learning to learn“. In the world of Artificial Intelligence, meta-learning describes methods where machines not only learn individual tasks but also how to learn new tasks more quickly and effectively. Instead of starting from scratch for every new situation, algorithms using meta-learning utilise their previous experiences and transfer this knowledge to new challenges.
An everyday example: Imagine a robot learning to grasp different cups. With meta-learning, it recognises certain basic patterns when grasping – such as how round a cup is or how heavy it is. Then, when a completely new cup comes into play, the robot can use its „knowledge about learning“ to figure out how best to grasp this new cup much more quickly.
Meta-learning saves time and resources because machines can independently adapt to ever-new tasks. This is an important breakthrough for companies, especially in the rapidly changing world of big data and automation.
Deutsch 
English (UK)