Constructing conditional PKEET with verification mechanism for data privacy protection in intelligent systems

Intelligent systems provide a wide range of smart services such as data analytics, mining and recommendation. Since the efficacy of intelligent algorithms are assured by the large volume of data which contain user-related sensitive information, how to make the best use of these data without leaking the private information becomes a primary challenge for intelligent systems. Ciphertext testing allows a system to test if the ciphertexts encrypted by different public keys correspond to the same plaintext and are suitable for providing privacy preserving calculation in intelligent systems, but conventional schemes lack the verification mechanism. To better solve the problem, in this paper, we present a new paradigm named conditional public key encryption with equality test (CPKEET). This paradigm allows a user to conduct ciphertext testing only if he holds the valid credentials generated by a designated issuer server. This property deals with the verification mechanism for data privacy protection in intelligent algorithms. We give the generic construction of this paradigm and instantiate a CPKEET scheme by proposing a round optimal attribute-based encryption with conditional equality test (RO-ABE-CET). We define the concrete algorithms and a detailed evaluation of RO-ABE-CET. The high security level and along with its superior efficacy makes it appropriate to be applied for data privacy protection in intelligent systems.