Research Fields
Security core technology group
- Research representativeAtsuko Miyaji
- Osaka University/Japan Advanced Institute of Science and TechnologyAkinori Kawachi / Tung Chou / Chen-Mou Cheng / Nakasho Kazuhisa / Yuki Takano
Research
The goal is to establish a secure and fair circulation platform for big data, enabling big data collection, analysis, utilization and feedback to data owner. Our research mainly focus on 4 specific fields as follows:
1.Secure Data Management
To design and implement highly reliable secure data secure management system. Within the system, data management server is able to perform dynamic data processing such data adding, data updating and deleting, furthermore, capable of executing the data recovering while keeping cost lower for data owner.
2.Cyber-attack Tolerance
This research topic considers the bidirectional cyber-attacks, both internal and external. We undertake to construct a client-oriented framework for detecting the SSL/TLS attack at a early stage in a simple way. We undertake to build a robust system to enhance the security against malicious server and external server attack with consideration of crystallographic primitives in use, while providing the functionality for hiding memory access pattern.
3.Secure Data Mining
It aim to build a flexible and extendable privacy-preserving data mining framework according to required privacy setting of data owner in a secured way, data user can perform privacy-preserving computation for their target functionality within the framework.
4.Secure Data Utility
Toward developing the core technologies for secure and fair P4V of big data analysis result, feedback and utilization, we make effort to implement the anonymous traceability for data owner from data analysis results and provide correctness proof and verification of trace without violating the anonymity.
Secure data management group
- Main research collaboratorShinsaku Kiyomoto
- KDDI R&D laboratories Inc.Toru Nakamura / Basu Anirban / Mohammad Shahriar Rahman / Mitsumoto Tomoaki / Yuto Nakano
Research
1.Data Protection in Cyber Security
Memory Protection techniques has been considered in order to prevent leakage of sensitive data on memory by bugs or vulnerabilities on software implementation. We are implementing the technologies on the SSL protocol and evaluating security against several types of attacks. Tracing techniques for anonymized datasets has been considered in this project.
2.Secure Data Management
Quantitative risk analysis methods on anonymized datasets have been considered in order to support user's decision. We also consider user support functions for configuration of user's privacy policy.
3.Secure Data Trading and Mining
Fair and secure data trading between data holders and service providers is a key factor in this project. We are considering some protocols for data value checking, data price negotiation, and data mining.
Living safety testbed group
- Main research collaboratorYoshifumi Nishida
- National Institute of Advanced Industrial Science and TechnologyYoichi Motomura / Koji Kitamura / Tatsuhiro Yamanaka / Tachio Takano
Research
1.Secure Data Mining Technology
In this study, we develop a secure data mining technology that allows integrated utilization of multi-organizational data. More specifically, we develop dynamic privacy policy control technology, causal modeling technology by using basic technologies such as living ontology, machine learning, dimension manipulation technology and graph structure analysis technology.
2.Living Safety Testbed
In this study, we apply the above technologies to the field of living safety. To this end, we develops a database platform consisting of first-aid data, injury data, insurance data, and so forth, and we evaluate the effectiveness of the developed technologies and the feasibility of industrial application.
Health testbed group
- Main research collaborator Katsuya Tanaka
- University of Tokyo Ryuichi Yamamoto/ Kazuhiko Ohe / Soichi Koike/ Mayu Yoshida
Research
1.Establish secure ID linkage technique
Establish and apply IHE PIX/PDQ based secure ID linkage technique, and evaluate the efficiency and feasibility.
2.Test bed for health big data
Making databases of dummy data of clinical data gathered with SS-MIX2 format, insurance claim data, and precise data set for DPC/PDPS (Diagnosis Procedure Combination / Per-Diem Payment System) for evaluating secure data management tools and data mining tools developed by security techniques study group with random swapping and other methods and evaluate the effectiveness and feasibility of these tools. Establish the data model for various sensor data useful for healthcare, and evaluating secure collecting tools and privacy preserving data mining tools.
