This folder provides Jupyter notebooks that demonstrate the usage of Responsible AI tools with Cognitive Services. Combined with Responsible Use of AI Overview, these notebooks empower developers and AI system stakeholders to use Cognitive Services more responsibly.

This folder contains an example notebook of assessing a facial validation system for intersectional performance disparity using Fairlearn and the FairnessDashboard. The notebook is run using a provided dataset. If you wish to run with assessment using your own data, you can run the run_face_verify.py with your dataset to get matching_scores from the system.

This folder contains an example notebook of assessing a fspeech-to-text system for fairness-related performance disparity using Fairlearn. The notebook is run using a provided dataset. If you wish to run with assessment using your own data, you can run the run_stt.py with your dataset to get automated transcriptions from the system. Ideally, these automated transcriptions would be evaluated against a human transcriber.