AI Ethics.pptx
Download as PPTX, PDF0 likes109 views
This document contains information about Mauro Bennici, an AI Architect and AI Ethicist. It lists his credentials and experience in fields including AI, data science, and ethics. It also provides links to his website and LinkedIn profile. The document then shares resources on responsible AI, explainable AI, and metrics for evaluating fairness in AI systems. It emphasizes the importance of using multiple metrics and examining individual results rather than averages alone.
AI Ethics.pptx
- 1. Mauro Bennici May 10, 2023 AI ethics AI Architect & AI Ethicist https://www.intelligenzaetica.it/ https://www.linkedin.com/in/maurobennici/
- 2. Mauro Bennici AI Architect, Kaggle and TechStars mentor, .NET Foundation member, Data Scientist, Professional Scrum Master (PSM I), Microsoft Certified Trainer (MCT), Azure Cloud Solutions Associate (MCSA). Member of the SME Digital Alliance Working Group for Ethics and Artificial Intelligence. AI Architect & AI Ethicist https://www.intelligenzaetica.it/ https://www.linkedin.com/in/maurobennici/
- 4. Dataset Principali: ● Inconsci ● Culturali ● Storici https://datasetsearch.research.google.com/
- 5. https://www.kaggle.com/competitions/quora-insincere-questions-classification ° Public Datasets ° APIs contracts
- 9. Title Text https://cloud.google.com/responsible-ai
- 12. Numbers Measuring outcomes for: ● Trust ● Security https://pair.withgoogle.com/explorables/measuring-fairness/
- 13. Bias Principles: ● Unconscious ● Cultural ● Historians https://pair.withgoogle.com/explorables/hidden-bias/
- 14. Microsoft ● Products ● Azure ● Development ● Tools https://github.com/tensorflow/fairness-indicators
- 16. Tensorflow Fairnerss Advantages: ● UI ● Integration with TF ● Examples https://colab.research.google.com /github/tensorflow/fairness- indicators/blob/master/g3doc/tutor ials/Fairness_Indicators_Pandas_ Case_Study.ipynb https://github.com/tensorflow/fairness-indicators
- 17. Fairlearn Advantages: ● Socio-Tech companies ● Communities ● Examples https://colab.research.google.com /github/tensorflow/fairness- indicators/blob/master/g3doc/tutor ials/Fairness_Indicators_Pandas_ Case_Study.ipynb https://fairlearn.org/main/auto_examples/plot_credit_loan_decisions.html#sphx-glr-auto-examples- plot-credit-loan-decisions-py
- 18. AI Fairness 360 Text https://ai-fairness-360.org/
- 19. AI Fairness 360 Example with LIME https://github.com/Trusted-AI/AIF360/blob/master/examples/demo_lime.ipynb
- 20. Metrics ● We do not use summary averages ● We use more than one metric ● Let's look at the individual results ● EX: F1-Score MACRO
- 21. Mauro Bennici Thank you! Q&A time AI Architect & AI Ethicist https://www.intelligenzaetica.it/ https://www.linkedin.com/in/maurobennici/