Consider the SGD weight$-$update rule for ridge regression. Assume there is a single predictor variable. Say the predictor's current weight w is positive. If on the next labeled instance $($x$,$ y$)$ that is presented during training, the prediction hat matches y exactly, w will

a$.$ increase, b$.$ decrease, c$.$ stay the same