I have a general question for you all. In ML, we have \alpha a parameter that sets the learning rate. In hypothesis testing, we also have \alpha for significance value.

Are these two alphas related? Does the alpha in learning rate somehow tied to statistical understanding? Or are they completely unrelated?

Hi @markredito,

Although the symbols used for learning rate and significance value happen to be the same, they are completely unrelated. In mathematics, some symbols may be overloaded, or be used differently in different contexts.

hi @markredito they are not related

The learning rate is the step you take to train your model and evaluate the performance, the alpha as significance level is the probability to find a significant value when is not actual significant.

So, those two are not related, they are meant for different things and evaluate in different ways.

1 Like

Thank you! Statistics has helped me understand ML models better, so I got curious and wanted to know if the concepts were built on top of each other. But yeah, I see a pattern that math in general share symbols to communicate different concepts.

1 Like

Thank you. That’s really clarifying

1 Like