How can we know the 'human-level error' in actual case?

[for_east]

Hi, I’m learning a deep learning through these courses, and getting a lot of useful knowledges.
Thank you for providing this cool lectures.

I have a question about the ‘human-level error’.
In a real field dataset, how can I get the human-level error?
For example, if the model’s goal is predict certain value, I mean if it is a regression model, how can I get the human-level error?

Thank you for reading it,

[Elemento]

Hey @for_east,
Welcome to the community. There have been a great many discussions in the Deep Learning Specialization in the past, regarding what to do in the absence of any human-level performance. Here are some posts for your reference:

• Bayes error, human-level performance and overfitting (structured data)
• Human Level Performance, how to set it?
• What to do when there is no human-level performance baseline?
• How to measure human level performance against human made labels?

You will find that these posts share a great deal of knowledge regarding your query, so let me know, if these help.

Cheers,
Elemento

[jenitta.jebaraj]

Hi,
Human level error can be found by making the human to perform the task. So in your example, if the field dataset is given to a human for prediction, then with what error level the human is predicting all the data in the dataset, gives the human error level.