Gradient Descent is an important concept in the fields of Artificial Intelligence, Big Data, and Digital Transformation. It refers to a mathematical procedure that helps solve problems step by step to find the best possible solution. Gradient Descent plays a central role, particularly in the development of self-learning computer models, known as algorithms.
Imagine you are on a hill in fog and want to reach the lowest point in the valley, but you can only see a few metres ahead. You feel your way forward slowly, each time taking the path downwards that you can currently see. Gradually, you thus reach the lowest point. This is how gradient descent works: the algorithm systematically searches for the solution with the lowest error to arrive at the optimal result.
In practice, for example, Gradient Descent is used to teach Artificial Intelligence to recognise images or understand texts. The method is understandable for computers, but also efficient because it quickly analyses large amounts of data and suggests improvements. Gradient Descent thus supports companies in automating processes and in decision-making.
Deutsch 
English (UK)