How to deal with errors or inconsistencies in the training data?

There are inconsistant wrong labels and consistant errors in training data. For the former I tried MC-dropout and data Shapley. For the later I wonder if manual data curation is a requisite?

Topic mlops training data-cleaning

Category Data Science

About

Geeks Mental is a community that publishes articles and tutorials about Web, Android, Data Science, new techniques and Linux security.