Creating Machine Learning models is a non-trivial task. This is especially true when working with images, video or any other kind of unstructured data. As in most situations, a model can only ever be as good as the data it has been trained on, placing a large burden on ensuring that the training set is well representative of real-world data.

To make things more difficult, biases are easy to introduce into a data set but notoriously difficult to spot. Data visualisation can play a key role in identifying biases, but traditional data visualisation tools struggle to function with unstructured data.

For this reason, a new breed of tools specifically designed to address this problem is needed. Zegami sets a marker in the development of such tools…