Suppose you have found a learning rate $\backslash$alpha $1$ that achieves low loss when using batch gradient descent. A learning rate $\backslash$alpha $2$ for stochastic gradient descent that achieves similarly low loss would be expected to be:

Group of answer choices

Lower because more updates are made per epoch.

Higher because less updates are made per epoch.

Lower because less updates are made per epoch.

Higher because more updates are made per epoch.