If the test-dataset also have some noise data, do you think it is hard to get the real performance?

Question

If the test-dataset also have some noise data, do you think it is hard to get the real performance?

guotong1988 opened this issue 3 years ago · 2 comments

guotong1988 commented 3 years ago

This question is based on https://github.com/ZJU-DAILY/SGDL/issues/3.
Thank you very much!

Answer 1 · 2022-04-25T09:17:02.000Z

If test set is also noisy, its data distribution is consistent with the training set, which is not our major concern. It can be solved by other normal machine learning solutions. We are happy to answer questions related to our paper. Thanks!

Answer 2 · 2022-04-25T09:35:27.000Z

Thank you again!