Transfer learning is the ability to apply what you have learned in one domain or context to another domain or context that is different but related. For example, if you have learned how to play chess, you can transfer some of your strategic thinking skills to other board games or even real-life scenarios.
Model generalization is the ability to apply what you have learned in a specific case or example to a broader range of cases or examples that share some common features. For example, if you have learned how to calculate the area of a rectangle, you can generalize that formula to other shapes that have length and width.
Transfer learning is good way for model reuse. Model generalization is way trying to change a modelblackbox into a whitebox, kind of. The latter is harder and more challengable.
迁移学习（Transfer learning）侧重于模型跨场景重用。模型泛化（Model generalization）是试图将模型黑盒转变为一定程度上的白盒。后者更具挑战性。
