Title:Privacy-aware Data Trading

Abstract:The growing threat of personal data breach in data trading pinpoints an urgent need to develop countermeasures for preserving individual privacy. The state-of-the-art work either endows the data collector with the responsibility of data privacy or reports only a privacy-preserving version of the data. The basic assumption of the former approach that the data collector is trustworthy does not always hold true in reality, whereas the latter approach reduces the value of data. In this paper, we investigate the privacy leakage issue from the root source. Specifically, we take a fresh look to reverse the inferior position of the data provider by making her dominate the game with the collector to solve the dilemma in data trading. To that aim, we propose the noisy-sequentially zero-determinant (NSZD) strategies by tailoring the classical zero-determinant strategies, originally designed for the simultaneous-move game, to adapt to the noisy sequential game. NSZD strategies can empower the data provider to unilaterally set the expected payoff of the data collector or enforce a positive relationship between her and the data collector's expected payoffs. Both strategies can stimulate a rational data collector to behave honestly, boosting a healthy data trading market. Numerical simulations are used to examine the impacts of key parameters and the feasible region where the data provider can be an NSZD player. Finally, we prove that the data collector cannot employ NSZD to further dominate the data market for deteriorating privacy leakage.