A system learns from examples, data patterns, and regularities. What makes machine learning different, however, is the way in which the system can abstract this learned knowledge independently and apply it to unrelated problems.
In one example involving preventative quality assurance, machine learning methods are used to conduct automated analyses of data from quality checks and returns. This makes it possible to identify interdependencies between product characteristics and product quality and derive optimizations for the product development process. An assistance system can point out the costs of potential errors to product developers as early as the design phase, thus providing further direct support.