Measuring Embedding Drift. Approaches for measuring…

Data drift in unstructured data like images is complicated to measure. The metrics typically used for drift in structured data -- such as population stability index (PSI), Kullback-Leibler divergence (KL divergence), and Jensen-Shannon divergence (JS divergence) -- allow for statistical analysis on structured labels, but do not extend to unstructured data. The general challenge with measuring unstructured data drift is that you need to understand the change in relationships inside the unstructured data itself. In short, you need to understand the data in a deeper way before you can understand drift. The goal of unstructured drift is to detect whether two unstructured datasets are different -- and, if so, to give workflows to understand why the datasets are different.