Malicious Agent Detection In Multi Party Data Access Structure
A data distributor has given sensitive data to a set of supposedly trusted agents (third parties). Some amount of data is leaked and found in an unauthorized machines (e.g., on the internet or computer). The distributor has to find-out from whom that data get leaked. The authors propose data allocation strategies (across the agents) that increase the chances of detecting leakages. These techniques do not need to modify data of the released data (e.g., watermarks). In some cases they can also embed real but fake data records to further increase their chances of detecting leakage and identifying the guilty agent.