Transfer learning is when an AI model applies knowledge learned during training on one task or domain to perform better on a different one. A model pre-trained on a broad corpus of text and then fine-tuned on domain-specific documents transfers its general language understanding to a specialized context.
This is more efficient than training a specialized model from scratch and typically produces better results — because the pre-trained model brings a strong foundation of general language capability that domain-specific training builds on rather than replacing. It is the basis for how most enterprise AI applications are built today.
