As humans, we don’t learn tasks in isolation. As we move our knowledge path forward, we build on previously learned and related challenges, which accelerates the learning process.
If you already know a particular musical, cooking, or dance style, you will not forget that knowledge and start from scratch when learning a different style; you will reuse the knowledge you already have and finetune it to absorb the new features. This is an incredible human capability, which can also be applied to machines through the use of Transfer Learning (TL).
This is a companion discussion topic for the original entry at https://www.pinecone.io/learn/transfer-learning/