Does any have any best practices or guidance on the best way to create user interfaces for operation users to create the training/input data sets for ML model development?
Ideally, I will be working with tens (or hundreds) of thousands of trained data samples, and want a way to have our internal experts assign the correct values to the training data.
I know that, in the past, I have read a blog about using Amazon Mechanical Turk to 'crowdsource' training data, however I need to create interfaces for internal use the same way. The people that have used MT, did you create the UI for them to assign to the samples?
Is the interface as simple as just paging through the list of most needed training points, and create web-forms to post the the db?
[link][2 comments]
