The National Center for High-performance Computing (NCHC) of the National Applied Research Laboratories (NARLabs) hosts Taiwan's inaugural PETs (Privacy Enhancing Technologies) Prospective Information Security Technology Seminar to accelerate the development and introduction of strengthened privacy technologies

With the popularization of cloud services, 5G, IoT, and other networking technologies, data flow will be within easy reach of members of society in the future. However, given that the use of big data is all the rage, how to strike a balance between public interest and strengthened protection of the privacy of personal data has become a challenge that scientists need to work hard to overcome and a goal for which the Ministry of Science and Technology has promoted important scientific research.
 

To ensure that the data in the cloud has both high availability and reliable privacy protection, the National Center for High-performance Computing (NCHC) of the National Applied Research Laboratories (NARLabs) of the Ministry of Science and Technology, in addition to using its national high-speed computing facilities to accelerate the development and introduction of a new generation of technology to protect the privacy of personal information in Taiwan, held Taiwan's inaugural Privacy Enhancing Technologies/PETs Prospective Information Security Technology Seminar today (October 15) in which in-depth discussions took place among domestic and foreign experts in information on encryption and privacy protection, including the Academia Sinica, National Taiwan University, National Yang Ming Chiao Tung University, IBM, Duality Technologies, Nvidia, etc., hoping to accelerate the realization of Taiwan's vision of becoming a "digital country, smart island."
 

The National Center for High-performance Computing (NCHC) of the National Applied Research Laboratories (NARLabs) has implemented the forward-looking digital infrastructure design to build a Taiwania supercomputers cluster, providing world-class supercomputers and AI high-speed computing services. To strengthen the protection of the future digital sovereignty of this country, NCHC also promotes the domestic development of PETs-related technologies, such as homomorphic encryption and federal learning and other related technologies in the country.
 
For example, in homomorphic encryption, encrypted data is uploaded to the cloud to be computed and operated on directly. Because such work consumes tremendous amounts of computing resources, it used to take millions of times more computing time just to get the result. However, computing power has increased in recent years and algorithms have enjoyed breakthroughs. This year, NCHC has tested its Taiwania supercomputers and found that computing has been accelerated to levels that are close to practical applications, proving that it has become possible for this method to become practical. 
 

The NCHC also for the first time publicly showed its own homomorphic encryption based on the concept of quantum computing. Wrong choices grew exponentially during encryption, so it can withstand future attacks from quantum computers. It uses a large number of quantum gate conversions, so regardless of how long the calculation process is, errors do not accumulate, while traditional homomorphic encryption technology inherently accumulates errors quickly. Therefore, our approach affords the advantages of accuracy and speed. Furthermore, the combination of the huge amounts of quantum gates in the calculation process makes it impossible for onlookers to discern the original calculation mode. Computing on the cloud offers the effect of shielding. This quantum-based calculation method has completely solved the dilemma of the existing technology, making it possible to realize homomorphic encryption, which not only helps expand new research directions but also more likely create a new industrial model.
 

In addition, federated learning has also contributed to the protection of privacy. In a decentralized manner, in federal learning, data is shared without leaving the organization of its owner. With AI training, the training results of all data can be shared. The National Center for High-performance Computing (NCHC) of the National Applied Research Laboratories (NARLabs) has successfully developed a solution that integrates Nvidia Clara, designed the browser user interface, and lowered the threshold for using federal learning. 
 

NCHC Director General Shepherd Shi indicated that with NCHC's supercomputers providing a verification simulation environment, more forward-looking privacy-enhancing technologies can be experimented, including new encryption technologies that can be adapted to the post-quantum era. This solution benefits the public interest that takes into account development and the protection of personal privacy. This is the major benefit manifested by the digital infrastructure of the nation.