What is transfer learning?
Say you were a statistician, and now you are learning about data science. You would use your past knowledge to build upon your skillset as a data scientist.
Transfer learning has the same idea. This deep learning technique enables developers to harness deep learning models used for one task and apply them to another domain. The pre-trained models are used as starting points, especially on computer vision and natural language processing tasks.
For example, say you are creating an application to identify wolves in the wild, but there aren’t any publicly available models that are good enough for your needs. With transfer learning, you can begin with a state of the art CNN model used for image recognition of animals, and you can fine-tune it to recognize images of wolves.