用神经网络实现一般访问结构的多重秘密共享

Multi-secret sharing scheme for general access structure based on artificial neural network

利用人工神经网络,结合RSA密码体制,实现了一种基于一般访问结构的多重秘密共享方案。在该方案中,秘密份额是人工神经网络收敛结果,各参与者共享多个秘密只需要维护一个秘密份额。共享多个秘密只需要进行一次人工神经网络训练,从而提高了方案的效率;在秘密分发和恢复时,利用RSA密码体制保证方案的安全性和正确性。分析表明,该方案是一个安全的、实用的秘密共享方案。

Based on RSA cryptosystem and artificial neural network, a new multi-secret sharing scheme for the general access structure is proposed. In the scheme, shadows are input value of converged artificial neural network. Participant shares many secrets with other participants by holding only one shadow. In the other word, to share many secrets, artificial neural network just converge once. As a result, scheme is more efficient. In the distributed and recovery phase, the security and validity of this scheme is same as the RSA cryptosystem. Analyses show that this scheme is computationally secure and practical scheme.