Closed bmcase closed 3 months ago
Attention: Patch coverage is 99.37888%
with 1 line
in your changes missing coverage. Please review.
Project coverage is 91.28%. Comparing base (
b50beae
) to head (ae972b9
). Report is 2 commits behind head on main.:exclamation: Current head ae972b9 differs from pull request most recent head e00f2b1
Please upload reports for the commit e00f2b1 to get more accurate results.
Files | Patch % | Lines |
---|---|---|
ipa-core/src/protocol/ipa_prf/dp/mod.rs | 99.37% | 1 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
@benjaminsavage converted to binary search for finding the smallest num_beroulli and added a couple more tests for that.
Small PR to implement Thm 1 of this paper https://arxiv.org/pdf/1805.10559 which allows us to go from an epsilon, delta and other parameters to the minimum number of bernoulli samples we need to achieve that level of DP. See this doc for more discussion. https://docs.google.com/document/d/116qQIgVoz3uV8axbmVWfif6dTjBVww5THecPqC0kERw/edit Much of this PR is a translation into Rust of this earlier python script https://github.com/private-attribution/ipa/compare/main...bmcase:raw-ipa:binomial_variable_p