Machine Learning has beautiful techniques to visualize results. And one of the experiments we did is to take these incredibly high dimensional outputs and cluster them into two-dimensional space. These are technologies that you could learn now on machinery that we won’t go into detail but the image is fun to see. The East dot in this two-dimensional sphere is now one image patch. You can then appear to dot and rearranging these dots by proximity, by similarity in our feature space as embodied by the neural network. And what you find is that, different types of images cluster very nicely in the space of features extracted by the neural networks the melanomas evaded sensing class that says in a certain primitive space that is very different from various other types of carcinomas and skin conditions in this visualization over him. There’s an interesting finding that no liquid really comes because the substance of what it means to classify it as skin cancer. In my opinion, much more profound way than possibly before this experiment. If you look back on the images that are represented by these dots you can look at the original raw images and you find to the human eye. There’s also similarities or similar images are being grouped and similar places as.