Neural Networks - The Success Story in Image Modelling

In image modelling, in neural networks, you want features that are adaptive to the task that you're handling. This is a field that's really opening odestod and it applies to chemistry. It applies to things like social network grafts to gengine interaction networks or prochin interaction networks. The really fundamental representation of the data is of relationships, and not necessarily as a grid. That revolution hasn't yet fully come to pass for graff valued inputs, but it's starting.