search

Trusted-AI / AIF360

A comprehensive set of fairness metrics for datasets and machine learning models, explanations for these metrics, and algorithms to mitigate bias in datasets and models.
https://aif360.res.ibm.com/
Apache License 2.0
2.46k stars 840 forks source link

Is there any guide to use postProcessing technique on text dataset? #418

Closed bloodteller123 closed 2 years ago

bloodteller123 commented 2 years ago

I'm doing some text classifications, and the dataset has only 'text' and 'label' attributes. I was wondering how to apply postProcessing techniques on my predictions as it seems that the library doesn't accept text input.. Thank you

nrkarthikeyan commented 2 years ago

You can create a dataset with only protected attributes, labels, and scores. That is sufficient for post-processing techniques generally.

bloodteller123 commented 2 years ago

Thank you for the reply! I wonder what the scores are in the context of text classification? (I was using Bert for classification task)

nrkarthikeyan commented 2 years ago

Predicted probabilities for classification of the positive class.

On Sat, Nov 5, 2022 at 11:36 PM bloodteller123 @.***> wrote:

Thank you for the reply! I wonder what the scores are in the context of text classification? (I was using Bert for classification task)

— Reply to this email directly, view it on GitHub https://urldefense.com/v3/__https://github.com/Trusted-AI/AIF360/issues/418*issuecomment-1304707125__;Iw!!IKRxdwAv5BmarQ!fL04BWt6SFyZdt57sHQ6rGAxxSqA3izXkcszaKVfPBm4SUW5pa9JF8qi_774HX7akXm0_KvCsRmQqTgfuIaI5AA$, or unsubscribe https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/ABNX22XEFD2V4OQTESAU4GDWG4RVNANCNFSM6AAAAAARVU6A6M__;!!IKRxdwAv5BmarQ!fL04BWt6SFyZdt57sHQ6rGAxxSqA3izXkcszaKVfPBm4SUW5pa9JF8qi_774HX7akXm0_KvCsRmQqTgfmM2ZN1g$ . You are receiving this because you commented.Message ID: @.***>