Check all the correct statements about Double Descent $($DD$)$

Check all the correct statements about Double Descent $($DD$)$

The extremum point for DD is somewhere the number of weights approximately equals to the amount of training data

DD can be easily checked on a toy datasets

DD only valid when both amount for training data and model weights are huge

DD is fairly new concept while bias$-$variance trade$-$off is a classical one

DD extends bias$-$variance trade$-$off curve to the area of an enormous complexity

DD is an improved version of Gradient Descent when we start to decrease the loss function after some number of iterations