mirror of
https://github.com/hanzo-ml/kserve.git
synced 2026-10-30 20:53:49 +00:00
c78d2987fa
Signed-off-by: GitHub <[email protected]> Co-authored-by: copilot-swe-agent[bot] <[email protected]>
AIF Model Fairness / Bias Detection
Input parameters
| Name | Description |
|---|---|
| model_name | (str) The name that the model is served under |
| predictor_host | (str) The host for the predictor. |
| feature_names | (list(str)): Names describing each dataset feature. |
| label_names | (list(str)): Names describing each label. |
| favorable_label | (float): Label value which is considered favorable (i.e. "positive"). |
| unfavorable_label | (float): Label value which is considered unfavorable (i.e. "negative"). |
| privileged_groups | (list(dict)): Privileged groups in a list of dicts where the keys are protected_attribute_names and the values are values in protected_attributes. Each dict is a single group. |
| unprivileged_groups | (list(dict)): Unprivileged groups in the same format as privileged_groups. |
Output metrics
| Name | Description |
|---|---|
| base_rate | (float): Compute the base rate, ππ(π=1)=π/(π+π) |
| consistency | (list): Individual fairness metric from [1] that measures how similar the labels are for similar instances. |
| disparate_impact | (float): ππ(π=1 | π·=unprivileged)ππ(π=1 | π·=privileged) |
| num_instances | (float): Compute the number of instances n |
| num_negatives | (float): Compute the number of negatives, π=βππ=1π[π¦π=0] |
| num_positives | (float): Compute the number of positives, π=βππ=1π[π¦π=1] |
| statistical_parity_difference | (float): ππ(π=1|π·=unprivileged)βππ(π=1|π·=privileged) |
[1] R. Zemel, Y. Wu, K. Swersky, T. Pitassi, and C. Dwork, βLearning Fair Representations,β International Conference on Machine Learning, 2013.
Build a development AIF bias detector docker image
First build your docker image by changing directory to kserve/python and replacing dockeruser with your docker username in the snippet below (running this will take some time).
docker build -t dockeruser/aifserver:latest -f aiffairness.Dockerfile .
Then push your docker image to your dockerhub repo (this will take some time)
docker push dockeruser/aifserver:latest
Once your docker image is pushed you can pull the image from dockeruser/aifserver:latest when deploying an inferenceservice by specifying the image in the yaml file.