An Enhanced Data Distribution With Protection of Agent Colluding Attacks
In a data distribution scenario the sensitive data given to agents can be leaked in some cases and can be found in unauthorized places. The authors consider the addition of fake objects to the distributed set which do not correspond to real entities but appear realistic to the agents. The distributor must assess the likelihood that the leaked data came from one or more agents, as opposed to having been independently gathered by other means. They also present data allocation strategies and algorithms for distributing objects to agents, in a way that improves their chances of identifying a leaker.