Machine Learning: Uncontrolled Backpropagation

I am having problems with some concepts of machine learning through neural networks. One of them is backpropagagation . In the weight update equation

delta_w = a*(t - y)*g'(h)*x

t- This is the “target conclusion” that will be your class label or something in the case of supervised learning. But what would be the “target exit” for unsupervised learning?

Can someone offer an example of how you will use BP in unsupervised learning, especially for clustering classification?

Thanks in advance.