Energy-Based Models are a term from the fields of Artificial Intelligence, Big Data and Smart Data, and automation. They describe a method by which computers can recognise and understand complex patterns and relationships. What's special about Energy-Based Models is that they favour or reject certain solutions with the help of a so-called „energy“ function. The models evaluate various possibilities and select the one that requires the least „energy“ – this is usually the best or most plausible solution.
Imagine an energy-based model is to recognise whether a cat or a dog is shown in pictures. The model calculates an „energy“ for different assignments. The picture is then given the label („cat“ or „dog“) that costs the least energy – meaning it fits best.
Energy-based models primarily help with sorting data, making predictions, or discovering patterns. They are flexible in their application, for example, in quality assurance in factories, for image recognition, or for suggesting suitable products to customers in online retail. This makes them a helpful tool for many companies to draw intelligent conclusions from large amounts of data.
Deutsch 
English (UK)