Watermarking anonymized datasets by adding decoys

1. A method, implemented by a one or more hardware processors, for watermarking anonymized datasets using decoys in a computing environment, comprising:

embedding by the one or more hardware processors one or more decoy records in an anonymized dataset such that a re-identification attack on the anonymized dataset to reverse effects of a masking operation resulting in the anonymization targets the one or more decoy records, wherein the one or more decoy records are embedded in the anonymized dataset according to a defined risk threshold representing a probability that the one or more decoy records will be subject to the re-identification attack such that user information used to create the decoy records is chosen based on the user information being known to be desirable based on the user information being subject to previous re-identification attacks.

2. The method of claim 1 , further including creating the one or more decoy records from datasets unrelated to the anonymized dataset.

3. The method of claim 1 , further including creating the one or more decoy records from a population dataset linkable with the anonymized dataset.

4. The method of claim 1 , further including defining the one or more decoy records to have risk of re-identification equal to or greater than the defined risk threshold.

5. The method of claim 1 , further including defining the one or more decoy records to be unique to each third-party recipient of the anonymized dataset.

6. The method of claim 1 , further including tracing the re-identification attack using the one or more decoy records to a third-party recipient of the anonymized dataset, wherein the one or more decoy records are unique to the third-party.

7. The method of claim 1 , further including using the one or more decoy records to identify a third-party executing the re-identification attack.

8. A system for watermarking anonymized datasets using decoys in a computing environment, comprising:

one or more computers with executable instructions that when executed cause the system to:

embed one or more decoy records in an anonymized dataset such that a re-identification attack on the anonymized dataset to reverse effects of a masking operation resulting in the anonymization targets the one or more decoy records, wherein the one or more decoy records are embedded in the anonymized dataset according to a defined risk threshold representing a probability that the one or more decoy records will be subject to the re-identification attack such that user information used to create the decoy records is chosen based on the user information being known to be desirable based on the user information being subject to previous re-identification attacks.

9. The system of claim 8 , wherein the executable instructions create the one or more decoy records from datasets unrelated to the anonymized dataset.

10. The system of claim 8 , wherein the executable instructions create the one or more decoy records from a population dataset linkable with the anonymized dataset.

11. The system of claim 8 , wherein the executable instructions define the one or more decoy records to have risk of re-identification equal to or greater than the defined risk threshold.

12. The system of claim 8 , wherein the executable instructions define the one or more decoy records to be unique to each third-party recipient of the anonymized dataset.

13. The system of claim 8 , wherein the executable instructions trace the re-identification attack using the one or more decoy records to a third-party recipient of the anonymized dataset, wherein the one or more decoy records are unique to the third-party.

14. The system of claim 8 , wherein the executable instructions use the one or more decoy records to identify a third-party executing the re-identification attack.

15. A computer program product for, by a processor, watermarking anonymized datasets using decoys in a computing environment, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:

an executable portion that embeds one or more decoy records in an anonymized dataset such that a re-identification attack on the anonymized dataset to reverse effects of a masking operation resulting in the anonymization targets the one or more decoy records, wherein the one or more decoy records are embedded in the anonymized dataset according to a defined risk threshold representing a probability that the one or more decoy records will be subject to the re-identification attack such that user information used to create the decoy records is chosen based on the user information being known to be desirable based on the user information being subject to previous re-identification attacks.

16. The computer program product of claim 15 , further including an executable portion that creates the one or more decoy records from datasets unrelated to the anonymized dataset.

17. The computer program product of claim 15 , further including an executable portion that creates the one or more decoy records from a population dataset linkable with the anonymized dataset.

18. The computer program product of claim 15 , further including an executable portion that:

defines the one or more decoy records to have risk of re-identification equal to or greater than the defined risk threshold; and

defines the one or more decoy records to be unique to each third-party recipient of the anonymized dataset.

19. The computer program product of claim 15 , further including an executable portion that traces the re-identification attack using the one or more decoy records to a third-party recipient of the anonymized dataset, wherein the one or more decoy records are unique to the third-party.

20. The computer program product of claim 15 , further including an executable portion that uses the one or more decoy records to identify a third-party executing the re-identification attack.