@article{Sayyad_Kulkarni_2022, title={Privacy Preserving Inference Over Encrypted Data}, volume={10}, url={https://ijisae.org/index.php/IJISAE/article/view/2147}, abstractNote={<p>Machine learning and deep learning techniques provide solution to various medical applications relevant to disease detection. Many a time’s these learning algorithms and their inferences are generated in untrusted environments like cloud. Medical practitioners would like to protect their data in such cases in untrusted environments but would like to generate an inference on their data. Our work provides a solution to generate privacy preserving inference over encrypted data in such untrusted environment like public cloud. HELib based fully homomorphic encryption approach is used to provide security to the trained model and the data of the owner. Our results shows the effectiveness of using the technique to generate inference on encrypted data without comprising the accuracy of the system. Our work demonstrates on benchmark datasets like MNIST for prototyping and heart disease detection that are used by many machine learning applications for benchmarking.</p>}, number={3}, journal={International Journal of Intelligent Systems and Applications in Engineering}, author={Sayyad, Suhel and Kulkarni, Dinesh}, year={2022}, month={Oct.}, pages={129–134} }