TFX: A TensorFlow-Based Production-Scale Machine Learning Platform, Denis Baylor, Jian Li, Andrei Lopatenko, Aurélien Plab, Mikhail Berezovskiy, Daniel Golovin, Robby Neiger, Andrew M. Miller, Stephen Kidd, Michael R. Jones, Alex Sergeev, Max G. E. Bauman, Ted H. Lee, Alexey Smirnov, David S. Berg, Todd Phillips, David F. Blank, Christine Cheng, Josh Smith, D. Sculley, Michael A. Walker, Clemens Mewald, 2017Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (ACM)DOI: 10.1145/3097983.3098021 - 介绍了TensorFlow Extended (TFX),这是一个全面的ML平台,其中特征转换管道是管理数据预处理和确保训练与服务之间一致性的核心组成部分。
Hidden Technical Debt in Machine Learning Systems, D. Sculley, Gary Holt, Daniel Golovin, Eugene Davydov, Todd Phillips, Dietmar Ebner, Vinay Chaudhary, Michael Young, Jean-François Crespo, Dan Dennison, 2015Advances in Neural Information Processing Systems 28, Vol. 28 (NeurIPS) - 一篇基础性论文,讨论了实际ML系统中常见的缺陷和技术债务来源,包括确保训练和服务环境中数据转换一致性的挑战。