System and method for secure two-party evaluation of utility of sharing data
Embodiments described herein provide a system for improving a classifier by computing a statistic for the utility of sharing data with a second party. The system may encrypt a set of class labels based on a public key/private key pair to obtain a set of encrypted class labels. The system may send a public key and the set of encrypted class labels to a second computing device. The system may receive an encrypted value computed by the second computing device based on the public key. The system may decrypt the encrypted value based on a private key to obtain a decrypted value. The system may then send a pair of encrypted values computed based on the decrypted value to the second computing device. The system may subsequently receive an encrypted utility statistic from the second computing device, and decrypt the encrypted utility statistic to obtain a decrypted utility statistic.
1. A computer-implemented method for obtaining a utility statistic, the method comprising a first computing device:
encrypting a set of class labels based on a public key/private key pair to obtain a set of encrypted class labels;
sending a public key and the set of encrypted class labels to a second computing device;
receiving a value computed by the second computing device using a set of data stored at the second computing device, wherein the set of data is not known to the first computing device, and wherein the value is encrypted using blinding known only to the second computing device and further encrypted using the public key;
decrypting the encrypted value based on a private key to obtain a decrypted value that remains blinded to all entities except the second computing device;
sending a pair of encrypted values computed based on the decrypted value to the second computing device;
receiving an encrypted utility statistic from the second computing device; and
decrypting the encrypted utility statistic to obtain a decrypted utility statistic.
2. The method of claim 1 , wherein sending the pair of encrypted values further comprises:
computing a pair of values based on the decrypted value; and
encrypting, based on the public key, the pair of values to obtain the pair of encrypted values.
3. The method of claim 1 , further comprising:
determining that the decrypted utility statistic exceeds a predetermined threshold;
requesting a set of data from the second computing device; and
applying the set of data to improve the classifier.
4. The method of claim 1 , wherein the class labels are associated with binary attributes.
5. The method of claim 1 , further comprising:
sending a request to the second computing device to compute a utility statistic.
6. The method of claim 1 , wherein sending the public key and the set of encrypted class labels further comprises:
computing a value based on a sum of elements of a feature vector which includes the set of class labels;
encrypting the computed value; and
sending the encrypted value to the second computing device.
7. The method of claim 1 , further comprising:
computing a second encrypted utility statistic in a two-stream mode; and
sending the second encrypted utility statistic to the second computing device.
8. The method of claim 7 , further comprising:
re-encrypting the decrypted utility statistic based on a first single-use key to obtain a re-encrypted utility statistic;
sending the re-encrypted utility statistic to the second computing device;
receiving a second re-encrypted utility statistic value and a second single-use key from the second computing device;
sending the first single-use key to the second computing device;
decrypting the second re-encrypted utility statistic based on the second single-use key to obtain a second utility statistic; and
determining that the second utility statistic is equal to the decrypted utility statistic.
9. The method of claim 1 , wherein the second computing device applies a statistical privacy technique to a set of data to obtain a second set of data to be considered for sharing/combining.
10. The method of claim 1 , wherein the second computing device performs operations comprising:
computing the encrypted value based on the set of encrypted class labels and the public key.
11. The method of claim 1 , wherein the second computing device performs at least one of:
a binary operation associated with an additively homomorphic private key encryption scheme, and
a binary operation associated with a scalar multiplication homomorphic private key encryption scheme.
12. A computing system for obtaining a utility statistic, the system comprising a first computing device and further comprising:
one or more processors,
a non-transitory computer-readable medium coupled to the one or more processors having instructions stored thereon that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
encrypting a set of class labels based on a public key/private key pair to obtain a set of encrypted class labels;
sending a public key and the set of encrypted class labels to a second computing device;
receiving a value computed by the second computing device using a set of data stored at the second computing device, wherein the set of data is not known to the first computing device, and wherein the value is encrypted using blinding known only to the second computing device and further encrypted using the public key;
decrypting the encrypted value based on a private key to obtain a decrypted value that remains blinded to all entities except the second computing device;
sending a pair of encrypted values computed based on the decrypted value to the second computing device;
receiving an encrypted utility statistic from the second computing device; and
decrypting the encrypted utility statistic to obtain a decrypted utility statistic.
13. The system of claim 12 , wherein sending the pair of encrypted values further comprises:
computing a pair of values based on the decrypted value; and
encrypting, based on the public key, the pair of values to obtain the pair of encrypted values.
14. The system of claim 12 , further comprising:
determining that the decrypted utility statistic exceeds a predetermined threshold;
requesting a set of data from the second computing device; and
applying the set of data to improve the classifier.
15. The system of claim 12 , wherein sending the public key and the set of encrypted class labels further comprises:
computing a value based on a sum of elements of a feature vector which includes the set of class labels;
encrypting the computed value; and
sending the encrypted value to the second computing device.
16. The system of claim 12 , wherein the method further comprises:
computing a second encrypted utility statistic in a two-stream mode; and
sending the second encrypted utility statistic to the second computing device.
17. The system of claim 16 , wherein the method further comprises:
re-encrypting the decrypted utility statistic based on a first single-use key to obtain a re-encrypted utility statistic;
sending the re-encrypted utility statistic to the second computing device;
receiving a second re-encrypted utility statistic value and a second single-use key from the second computing device;
sending the first single-use key to the second computing device;
decrypting the second re-encrypted utility statistic based on the second single-use key to obtain a second utility statistic; and
determining that the second utility statistic is equal to the decrypted utility statistic.
18. The system of claim 12 , wherein the second computing device performs operations comprising:
computing the encrypted value based on the set of encrypted class labels and the public key.