Transfer Learning
In plain terms
Reusing a pre-trained model for a new related task.
Going deeper
Transfer learning adapts a model trained on one large task to a new task with less data, leveraging learned features. It transfers prior knowledge.
Under the hood
Fine-tuning or feature extraction from foundation models slashes data and compute needs.
Examples
Fine-tuning LLMs, image models from ImageNet, low-data tasks, foundation models.
Think of it like...
Like an experienced chef quickly mastering a new cuisine using existing skills.