Low-Rank Adaptation (LoRA) is a term from the fields of Artificial Intelligence, Digital Transformation, and Big Data. It describes a smart method for adapting large AI language models to specific tasks more quickly and cheaply, without the need for a complete retraining.
Imagine a company already uses a large AI model to analyse customer queries. Now, this model needs to be tailored to the specific language and questions from their own sector. With traditional methods, this would be very costly and time-consuming. This is precisely where Low-Rank Adaptation (LoRA) comes into play: instead of altering the huge model completely, only small “adapter” layers are inserted and trained. This saves resources and time, and the AI still learns to adapt better to the specific application.
A vivid example: A fashion retailer wants to optimise its existing AI, which handles general product enquiries, specifically for enquiries about sustainable fashion. With LoRA, the existing model can be fine-tuned for this new task in a flash – without time-consuming retraining and with little extra storage required. This makes LoRA a real game-changer for many areas of artificial intelligence.
Deutsch 
English (UK)