8 – Local Minima

So let’s recall a gradient descent does. What it does is it looks at the direction where you descend the most and then it takes a step in that direction. But in Mt. Everest, everything was nice and pretty since that was going to help us go down the mountain. But now, what if we try to do it here in this complicated mountain range. The Himalayans lowest point that we want to go is around here, but if we do gradient descent, we get all the way here. And once we’re here, we look around ourselves and there’s no direction where we can descend more since we’re at a local minimum. We’re stuck. In here, gradient descent itself doesn’t help us. We need something else.

%d 블로거가 이것을 좋아합니다: