| 研究実績の概要 |
During this fiscal year, we have made significant progress in our research on differential privacy and isogeny-based cryptography:
(1) Our analysis of the robustness of combinatorial algorithms under local differential privacy has continued, with a focus on spectral clustering, a widely used clustering method. We demonstrate that, provided the required privacy level is not excessively high, the clustering results remain unaltered by local differential privacy.
(2) Our research on isogeny-based cryptography aims to optimize the system for parallel environments such as multi-core and SIMD processors. We concentrate on two sub-operations within the system: addition and modular reduction. These operations are often overlooked in other research because they are considered fast and optimization is not expected to significantly improve the overall system. However, we show that optimizing these operations in SIMD architecture results in a substantial reduction in latency - a 30% improvement. This work has been accepted at the prestigious CHES 2023 conference.
(3) As part of the "Creation and Organization of Innovative Algorithmic Foundations for Social Advancement" (AFSA) project, we have developed a combinatorial optimization algorithm to determine the optimal use of bus routes for air quality monitoring. We provide an approximation algorithm and fixed-parameter tractability for this situation. This work has been accepted and presented at VTC-2022 Fall.
|
