Interpretable ml of Li Hongyi's machine learning model

【 Li hongyi machine learning 2021】 Interpretability of machine learning model (Explainable ML)_ Bili, Bili _bilibili

 

We don't use neural network because it's unexplainable ？

No , We can use powerful models , And try to explain him

So linear model More powerful interpretable decision tree Is that OK ？

But in fact decision tree Sometimes the effect is not good , And when decision tree Too many times , We also don't know why we choose this

 

  The goal of interpretable machine learning model

Printer psychology experiment

When people give one more reason , Acceptance will be higher , People just need a reason , Even if the reason is to repeat the requirements