OATH

Secure and Confidential Certificates of Online Fairness (OATH)

Olive Franzeseunverified

Approach

Cut-and-choose protocol exploiting statistical properties of group fairness; uses a ZK proof of correct inference as a subroutine

Notes

Sometimes cited under the title "OATH: Efficient and Flexible Zero-Knowledge Proofs of End-to-End ML Fairness". Composes with the inference systems in this repo rather than replacing them.

Our reading

Citation neighbourhood

Builds on
  • none recorded
Cited by, in this corpus
  • none recorded

Edges are a proxy: paper A's text mentions B anywhere (body or bibliography). See the full graph.

Discussed in

Other recorded fields
property_proven: Online group fairness w.r.t. demographic variables (sex, race, location),
  under distribution shift and provider malfeasance