Services used to collect training data for Machine Learning

Question

What services do ML/AI researchers use to get training data? (Amazon's Mechanical Turk, CrowdFlower)

What kinds of tasks do researchers post to those services? Are tasks usually a variant of classification tasks^*?

I am trying to collect some data from ML researchers, practitioners or hobbyists.

^{* User can select one of the classes (options) provided as solution to the task.
E.g. Is this product a [1] Electronics Product [2] Beauty Product [3] Health Product?}

Answer 1

Have a look at this website, it contains various databases used for machine learning algorithms: https://archive.ics.uci.edu/ml/

Also, there's a data analysis competition platform with a lot of datasets: https://www.kaggle.com/