Date of Publication :7th April 2016

Abstract: Back-propagation Neural Network is one of the important machines learning technique for classification Accuracy of any machine learning technique is improved with volume of datasets. Recent developments in computer networks provide such environment that multiple users connect with each other. This scenario leads in development of machine learning technique with collaborative or joint participation of these multiple users. In collaborative learning multiple parties participate jointly in learning where data is shared by users may contain sensitive information such as corporate data, health care data and personal data. There are chances of data leakage by any corrupt party or intruders. Users are concern about privacy of their datasets due to this main hurdle in collaborative machine learning technique many users are not interested to participate. In this project we are giving solution to privacy of individual user’s data by converting data in the form of cipher texts where users can use these cipher texts directly in the learning process. A machine learning system working on cipher texts needs various computations. Recent development in cloud computing provides good computing environment where we are utilizing these computations. In this way a system will be developed with neural network machine learning technique working on cipher text and utilizing cloud computing services.

Reference :