Much machine learning, essentially mapping in-the-wild data to training sets, uses de facto human goals. Active and integrated human-in-the-loop training deser

Much machine learning, essentially mapping in-the-wild data to training sets, uses de facto human goals. Active and integrated human-in-the-loop training deserves tool development towards an orientation for domain experts over ML/datascience experts. While further algorithmic development creates opportunities for scale, this is more correctly a UX problem.