Thanks! I'm curious to know how Lyst would handle outlier datapoints. ANN seems to be a variation of regression techniques (achieving nearest result) instead of creating a new mutual exclusion (unsupervised learning)