clamping gaussian noise multiplier

Qrlew · Mar 20, 2024 · f619931 · f619931
1 parent 8dc9795
commit f619931
Show file tree

Hide file tree

Showing 3 changed files with 8 additions and 19 deletions.
diff --git a/src/differential_privacy/aggregates.rs b/src/differential_privacy/aggregates.rs
@@ -1,12 +1,5 @@
 use crate::{
-    builder::{With, WithIterator},
-    data_type::DataTyped,
-    differential_privacy::dp_event::DpEvent,
-    differential_privacy::{dp_event, DpRelation, Error, Result},
-    expr::{aggregate::{self, Aggregate}, AggregateColumn, Expr, Column, Identifier},
-    privacy_unit_tracking::PupRelation,
-    relation::{field::Field, Map, Reduce, Relation, Variant},
-    DataType, Ready,
+    builder::{With, WithIterator}, data_type::DataTyped, differential_privacy::{dp_event::{self, DpEvent}, DpRelation, Error, Result}, display::Dot, expr::{aggregate::{self, Aggregate}, AggregateColumn, Column, Expr, Identifier}, privacy_unit_tracking::PupRelation, relation::{field::Field, Map, Reduce, Relation, Variant}, DataType, Ready
 
 };
 use std::{cmp, collections::HashMap, ops::Deref};
@@ -78,7 +71,6 @@ impl Relation {
             // Cf. Theorem A.1. in (Dwork, Roth et al. 2014)
             log::warn!("Warning, epsilon>1 the gaussian mechanism applied will not be exactly epsilon,delta-DP!")
         }
-
         let number_of_agg = bounds.len() as f64;
         let (dp_relation, dp_event) = if number_of_agg > 0. {
             let noise_multipliers = bounds

diff --git a/src/differential_privacy/dp_event.rs b/src/differential_privacy/dp_event.rs
@@ -164,7 +164,8 @@ impl From<Vec<DpEvent>> for DpEvent {
 }
 
 pub fn gaussian_noise(epsilon: f64, delta: f64, sensitivity: f64) -> f64 {
-    (2. * (1.25_f64 / delta).ln()).sqrt() * sensitivity / epsilon
+    // it can be inf so we clamp the results between 0 and f64::MAX
+    ((2. * (1.25_f64 / delta).ln()).sqrt() * sensitivity / epsilon).clamp(0, f64::MAX)
 }
 
 pub fn gaussian_tau(epsilon: f64, delta: f64, sensitivity: f64) -> f64 {

diff --git a/src/rewriting/mod.rs b/src/rewriting/mod.rs
@@ -568,7 +568,7 @@ mod tests {
                     ("CODE", DataType::integer()),
                     ("DESCRIPTION", DataType::text()),
                     ("BASE_ENCOUNTER_COST", DataType::float()),
-                    ("TOTAL_CLAIM_COST", DataType::float_min(-1.,)),
+                    ("TOTAL_CLAIM_COST", DataType::float_min(-1.)),
                     ("PAYER_COVERAGE", DataType::float()),
                     ("REASON_CODE", DataType::integer()),
                     ("REASONDESCRIPTION", DataType::integer()),
@@ -593,18 +593,14 @@ mod tests {
         let dp_parameters = DpParameters::from_epsilon_delta(1., 1e-3);
 
         let queries = [
-            // r#"
-            // SELECT
-            //     1/"TOTAL_CLAIM_COST"
-            // FROM private.axa_encounters e
-            // "#,
             r#"
             SELECT
                 "ENCOUNTERCLASS",
                 COUNT(p."Id") as patient_count,
-                SUM("TOTAL_CLAIM_COST") as total_cost
-            FROM  private.axa_patients p
-            JOIN private.axa_encounters e
+                SUM("TOTAL_CLAIM_COST") as sum_cost,
+                AVG("TOTAL_CLAIM_COST") as avg_cost
+            FROM  axa_patients p
+            JOIN axa_encounters e
             ON p."Id" = e."PATIENT"
             GROUP BY "ENCOUNTERCLASS"
             "#,