Cross-Validation Is Useless: Reflecting On Data Science Topics

Cross - Validation is a well-established technique in data science in order to realistically report the performance of machine learning models on new, unseen data. In this video I offer perspectives on why cross-validation might be considered useless and skipped when working with big data, as well as why data scientist rely on it no matter the financial and time expenses. Take it as an exercise for your critical thinking in the days of "AI experts". Contents: 00:00 - Motion 00:53 - What is cross validation used for? 03:24 - Arguments in favor of cross validation 06:19 - Arguments against cross validation 08:55 - Let me know your opinion on this video!