In a number of applications, data may be anonymized, obfuscated, or highly noisy. In such cases, it is difficult to use domain knowledge or low-dimensional visualizations to engineer the features for tasks such as machine learning, instead, we explore dimensionality reduction (DR) as a...