This story was originally published on HackerNoon at: https://hackernoon.com/privacy-preserving-computation-of-fairness-for-ml-systems-acknowledgement-and-references.
Discover Fairness as a Service (FaaS), an architecture and protocol ensuring algorithmic fairness without exposing the original dataset or model details.
Check more stories related to tech-stories at: https://hackernoon.com/c/tech-stories.
You can also check exclusive content about #ml-systems, #ml-fairness, #faas, #fairness-in-ai, #fairness-as-a-service, #fair-machine-learning, #fairness-computation, #cryptograms, and more.
This story was written by: @ashumerie. Learn more about this writer by checking @ashumerie's about page,
and for more stories, please visit hackernoon.com.
Fairness as a Service (FaaS) revolutionizes algorithmic fairness audits by preserving privacy without accessing original datasets or model specifics. This paper presents FaaS as a trustworthy framework employing encrypted cryptograms and Zero Knowledge Proofs. Security guarantees, a proof-of-concept implementation, and performance experiments showcase FaaS as a promising avenue for calculating and verifying fairness in AI algorithms, addressing challenges in privacy, trust, and performance.