FASCA: Framework for Automatic Scalable Acceleration of ML Pipeline | IEEE Conference Publication | IEEE Xplore