The term "offline reinforcement learning" belongs to the categories of artificial intelligence, automation and Industry and Factory 4.0. It refers to a special method by which artificial intelligence (AI) can learn to make decisions - not by trying things out live in the real world, but by learning from data that has already been collected.
Imagine a robot learning how to efficiently sort parcels in a warehouse. With conventional reinforcement learning, the robot would repeatedly try out how it has to move and learn directly from success or failure. However, this would be risky, expensive and time-consuming.
In offline reinforcement learning, the robot uses recorded data instead, for example the movements and decisions of experienced warehouse workers. Based on this data, the AI attempts to determine the best behaviour without making new mistakes in real operation. This makes training safer and cheaper.
Offline reinforcement learning is particularly practical when experiments would be expensive or dangerous - such as with autonomous cars or when dealing with machines in industry. This allows companies to utilise the benefits of AI without taking risks in their day-to-day business.
Deutsch 
English (UK)