US 12,081,644 B2
Efficient distributed privacy-preserving computations
Jonas Boehler, Karlsruhe (DE)
Assigned to SAP SE, Walldorf (DE)
Filed by SAP SE, Walldorf (DE)
Filed on Feb. 1, 2021, as Appl. No. 17/164,274.
Prior Publication US 2022/0247548 A1, Aug. 4, 2022
Int. Cl. H04L 9/00 (2022.01)
CPC H04L 9/002 (2013.01) [H04L 9/001 (2013.01); H04L 2209/08 (2013.01); H04L 2209/46 (2013.01)] 15 Claims
OG exemplary drawing
 
1. A system, comprising:
at least one data processor; and
at least one memory storing instructions which, when executed by the at least one data processor, result in operations comprising:
receiving, from each of a plurality of clients, a utility score and a partial noise value, wherein a utility function, at each of the plurality of clients, determines the utility score, and wherein a higher utility score indicates that a selection probability of a corresponding data value is higher than another data value with a lower utility score;
performing, based on the received utility scores and the partial noise values, a secure multi-party computation of a privacy-preserving statistic, the performing of the secure multi-party computation of the privacy-preserving statistic comprising determining a noisy utility score for each data value in a domain of data values and selecting a highest noisy utility score from the determined noisy utility scores, wherein selecting the highest noisy utility score from the determined noisy utility scores comprises:
initializing a first variable to zero;
for a first possible output value, determining a first noisy utility score for the first possible output value;
in response to the first noisy utility score being greater than the first variable, setting the first variable equal to the first noisy utility score;
for a second possible output value, determining a second noisy utility score for the second possible output value;
in response to the second noisy utility score being greater than the first variable, selecting the second noisy utility score as the highest noisy utility score; and
providing, based on the selected highest noisy utility score, the second possible output value from the domain of data values for the privacy-preserving statistic.