The bias-variance dilemma means that it is impossible to reduce both bias and variance at the same time, and one can only achieve a balance between the two.

In the model, if you want to reduce the deviation, you will increase the complexity of the model to prevent underfitting; but at the same time, you can't make the model too complex to increase the variance and cause overfitting. Therefore, you need to find a balance point in the complexity of the model, which can be represented by the following figure: