Adding explanation for parameter to backward in autograd tutorial

In a workshop on pytorch, there were questions about [the parameter passed to backward](https://github.com/pytorch/tutorials/blob/master/beginner_source/blitz/autograd_tutorial.py#L170) in the autograd tutorial. It would be nice to add an explanation such as [this one](https://discuss.pytorch.org/t/what-does-tensor-backward-do-mathematically/27953/2).