Scanflow: A multi-graph framework for Machine Learning workflow management, supervision, and debugging

作者：

Highlights：

• Machine Learning is more than just training, the whole view must be considered.

• Security vulnerabilities, concept/data drift and lack of explainability are issues.

• Model needs to be supervised and debugged to guarantee its validity and robustness.

• A hybrid symbolic (human knowledge) and learning (abnormal behavior) system is needed.

摘要

•Machine Learning is more than just training, the whole view must be considered.•Security vulnerabilities, concept/data drift and lack of explainability are issues.•Model needs to be supervised and debugged to guarantee its validity and robustness.•A hybrid symbolic (human knowledge) and learning (abnormal behavior) system is needed.

论文关键词：Machine Learning,Symbolic knowledge,Graph,Robustness,Containerization,Concept drift

论文评审过程：Received 19 October 2021, Revised 8 April 2022, Accepted 9 April 2022, Available online 19 April 2022, Version of Record 5 May 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117232