VOL.57
October 2019
News — ETRI Issue
Headline News
ETRI unveils latest AI technologies at IFA 2019 in Germany
Korean researchers recently revealed the results of AI research at Internationale Funkausstellung (IFA), the world's largest home appliance exhibition held in Berlin, Germany. ETRI exhibited the recent outcomes of its R&D and technology marketing at IFA, which began on September 6.
It introduced four technologies including a smart factory operation control system, the Internet of media technology, a deep learning high-speed processing system, and a new AI motion analysis technology. Furthermore, Dr. Kim Hyun of the Smart ICT Convergence Research Department gave a lecture on the theme of the ICT-enabled Smart Factory to the world through the IFA NEXT Innovation Engine.
ETRI develops sensor using sound to detect intrusion and fire
A Korean research team has developed a sensor that uses sound to detect unauthorized intrusions and fire. The product is expected to enhance facility security as it can be installed at a low cost without a blind spot and contribute significantly to the localization of parts and materials. ETRI disclosed that it had formed a partnership with the research company Secuworks to develop the world’s first smart safety sensor capable of detecting intrusion, movement, and even fire based on changes in the sound field. The new technology is particularly significant in that it mobilizes the ear, whereas conventional security technologies focus largely on the eyes.
ETRI’s AI deep learning to learn more quickly
ETRI recently announced that it has developed a high-speed processing technology optimized for deep learning distributed technology. The application of this technology enables an AI model to reduce learning from one week to just one or two days in the same environment. In other words, it increases learning efficiency by up to four times. Deep learning is a technology that enables computers to think and learn like humans. However, it still takes a computer a long time to learn a large volume of data or models such as videos, images, and sounds. In essence, the distributed learning technology can reduce the overall learning time by using multiple computers to distribute the learning load and process a large volume of data.
ETRI to build global research cooperation system for quantum computing
ETRI held the international workshop for quantum computing security technology, which was attended by more than 200 people, at the ICC Hotel in Doryong-dong, Daejeon on September 19. The workshop featured special lectures by Dr. Lily Chen of the US NIST, which is leading a quantum computing standardization project, and Professor O’Neill of Queens University, which is leading the European Information Security Project in the United Kingdom.
ETRI plans to develop and create the same environment as quantum computing by running a quantum algorithm specific to encryption analysis and building platform to measure the security strength of encryption in advance.
ETRI develops integrated cloud-storage interface technology
A Korean research team has developed a technology that integrates the in-house storage space and the outside cloud storage spaces provided by IT companies. The new technology should assist greatly with the development of the cloud infrastructure by enabling safe and easy storage and management of the ever increasing quantities of big data. ETRI announced that it has developed the Cloud storage integrated Solution (CiS) to safely manage storage spaces where large volumes of data are scattered in a platform.