Project used sampling/de-dup technique based on schema knowledge on HoloClean system. Goal was to reduce number of training points in order to decrease memory footprint and in same improve run-time.
sql database neural-network sampling deduplication conda-environment backend-development holoclean domain-knowledge-integration
-
Updated
Apr 16, 2019 - Python