The term fine-tuning of foundation models comes from the fields of artificial intelligence, digital transformation, big data and smart data. Foundation models are large, pre-trained artificial intelligence models that have learnt a great deal of data, for example language or image models. Fine-tuning means that these general models are subsequently adapted to specific tasks or company data.
Imagine a foundation model as an employee who already has many skills but does not yet know your industry. Through fine-tuning, this employee then learns specific additional skills - such as special technical terms or typical processes in your company. This saves time and money because you don't have to train a model from scratch every time.
For example, an insurance company uses a foundation model to process damage reports. To ensure that the model understands technical terms and typical formulations from the insurance industry, it is fine-tuned using the company's own insurance data. In this way, the artificial intelligence delivers more reliable results, adapts better to the company's requirements and thus improves operational efficiency.
Deutsch 
English (UK)