[Courtesy of KAIST]
SEOUL -- Using an artificial intelligence chip, South Korean researchers have created a face correction system that automatically fixes or completes features such as hair, glasses and eyebrows when asked to be added or erased from a picture taken with a tablet camera.
A generative adversarial network (GAN) is a class of machine learning systems. CANs, which are made up of several deep neural networks, can be used to create photos of imaginary fashion models or portraits, landscapes, and album covers.
it is difficult to implement CANs on small mobile devices with limited computational capabilities. To implement AI technology in mobile devices, low power is required for high-speed operations to prevent battery explosion.
The state-run Korea Advanced Institute of Science & Technology (KAIST) said that a research team led by Yoo Hoi-jun, an electrical engineering professor, developed a CAN processing unit (GANPU) capable of handling multiple deep neural networks on low-power mobile devices. It is possible to self-learn CANs within a mobile device without sending data to the server.
"By expanding the area of artificial intelligence utilization in mobile devices, it can be applied to applications that can transform image styles, synthesize images and restore images," Yoo said.
Using an AI semiconductor that efficiently processes deep learning by applying variable artificial neural network technology, Yoo has already developed an emotion recognition system that automatically recognizes seven emotions, including happiness, sadness, surprise, fear and expressionless, by recognizing facial expressions through smartphone cameras.
Copyright ⓒ Aju Press All rights reserved.
Related News
-
Education Minister Must Change: Urgent Reform Needed for South Korea's Education in the AI Era
South Korea's education system stands at a critical crossroads. As artificial intelligence (AI) transforms the production and delivery of knowledge, as well as the structure of industry and labor, the education system remains trapped in outdated frameworks focused on exams, scores, and university rankings. Education, which should be racing toward the future, is hindered by the need to manage past systems. It is time to fundamentally question whether the Ministry of Education is adequately responding to the changes of the era.The college entrance exam held on November 19, 2026, encapsulated these issues. According to the Ministry of Education, 98 individuals were caught cheating by the end of the fourth session. Among these, five cases involved the use of smart devices during the exam. Advanced devices such as smart glasses, miniature cameras, and wireless earphones have emerged as threats to the fairness of examinations. Education authorities utilized disaster text alerts and wireless signal detection to identify cheating. This indicates that the supervision system functioned to some extent. However, whether the existing exam management system can maintain fairness in the face of increasingly sophisticated cheating methods is a separate question.Allegations of cheating by a perfect scorer on the Law School Admission Test (LEET) and a related death incident have also shocked society. Caution is warranted against hasty judgments regarding the truth of the allegations and the cause of death. Nonetheless, it is true that this incident has sparked discussions about excessive competition for grades and a test-centric culture in South Korea.The essence of the problem does not lie in a few instances of exam cheating. In an era where AI fundamentally alters human intellectual activity, what kind of talent is South Korea truly cultivating? Does the Ministry of Education possess the national vision and execution capabilities to answer that question?The Ministry of Education must not remain a department that merely manages exams. It must transform into a national strategic agency that designs the future of South Korea.China's university reforms are a case we must examine soberly. Since the 1990s, China has pursued national-level university innovation. Through the 211 Project, it has enhanced the level of university education and scientific research, and with the 985 Project initiated in 1998, it has focused on nurturing world-class research-oriented universities. In the 2000s, these policies were expanded to closely link universities with national scientific and technological development.Tsinghua University has expanded its traditional engineering strengths into advanced industries such as semiconductors, information communication, and AI. Peking University has maintained its tradition in the humanities and social sciences while strengthening its capabilities in basic sciences and advanced research in mathematics, physics, computer science, and life sciences. Both universities have grown into representative research-oriented institutions supporting China's scientific and technological competitiveness.This does not mean we should blindly follow China's state-led education system. The values of academic freedom, university autonomy, and a creative research environment are indispensable. However, we need to learn from the long-term strategies that connect the nation's future industries with the research and education systems of universities.South Korea also has excellent research-oriented universities, including KAIST, POSTECH, and Seoul National University. There are many researchers achieving global success in fields such as semiconductors, AI, and biotechnology. Nevertheless, when looking at the entire national education system, the cultivation of talent required by future industries and the reform of university education structures are not progressing at a sufficient pace.The concentration of outstanding talent in the medical field is one aspect of this issue. Medicine, law, and business are all essential fields for society. The problem lies not in the preference for specific professions but in the structure where those who challenge and research in the fields of science and technology, and engage in entrepreneurship, do not receive appropriate rewards and social respect. While the world fiercely competes in AI, semiconductors, robotics, and aerospace, our talent selection system still relies excessively on entrance scores and university prestige.It is time to change the focus of university education. Simply increasing the number of students in science and engineering is insufficient. We must strengthen basic sciences and advanced engineering while expanding the integration of humanities, social sciences, and business studies. We need to cultivate not only the ability to utilize AI but also the capacity to discover new problems, assess the social impacts of technology, and design creative solutions.The Ministry of Education is already promoting policies to expand AI education, nurture talent for advanced industries, and support regional universities. However, merely expanding individual projects does not equate to educational reform. We must evaluate how much universities have actually changed, how much research personnel have grown, and how closely industry and education are connected, rather than just counting the number of projects.To achieve this, a comprehensive redesign of the Ministry of Education's organization and functions is necessary.First, the Ministry of Education should strengthen its function to oversee the national future talent strategy. A cross-government talent strategy system involving relevant ministries such as the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, and the Ministry of Employment and Labor should be established. We must end the siloed administration that manages university education, research and development, industrial workforce supply, and lifelong education separately.Second, we must expand both the autonomy and accountability of universities. The Ministry of Education should move away from detailed control over universities through quotas, evaluations, and financial support projects. While ensuring autonomy for universities to establish new departments, reform integrated curricula, and promote joint research with businesses, we must hold them strictly accountable for the quality of education, research outcomes, and student growth.Third, we need to diversify the functions of universities. There is no reason for all universities to compete in the same ranking. We should allow world-class research-oriented universities, specialized universities linked to regional industries, vocational education-focused universities, and lifelong education hub universities to fulfill their respective roles. We must move away from policies that reduce regional universities to mere replicas of metropolitan universities and instead build a structure where regional industries, research institutions, and local governments grow together.Fourth, we must make AI education a fundamental capability for all citizens. In elementary, middle, and high schools, students should learn not only AI utilization skills but also the basics of mathematics and science, logical thinking, and ethical judgment. In universities, we should combine majors with AI, and provide retraining opportunities for working adults. Older adults and retirees should also have the chance to learn new technologies and participate in society. Education in the AI era does not end the moment one graduates from school.Fifth, we must innovate both technical education and human education. As AI processes knowledge and robots perform labor, the importance of human conscience, responsibility, and moral judgment becomes even greater. We must not neglect history, philosophy, literature, and the arts in the name of strengthening scientific and technological education. We need to cultivate the ability to create technology and the wisdom to use that technology for humanity and society.The qualifications of the education minister who will lead this major transformation must also change.The new minister should not merely be someone familiar with the practices of educational administration. They must understand the flow of scientific and technological advancements and have direct experience leading university innovation. Experience managing education, research, personnel, finances, international cooperation, and industry-academia collaboration as a university president would be a significant asset. They should be able to understand the demands of advanced industries while upholding academic freedom and the public nature of education.Above all, they must possess a national vision that expands education in the AI era beyond merely enhancing the competitiveness of a select elite to provide growth opportunities for all citizens.Education Minister Lee Jae-myung has accumulated extensive experience in educational administration through roles as a teacher, education activist, and superintendent of Sejong City. That experience deserves respect. However, the educational reform that South Korea currently demands goes beyond the realms of school administration and educational welfare to the level of designing the future of national science, technology, and industry. We must critically evaluate whether the current leadership of the Ministry of Education is adequately responding to these changes.The government should consider a complete renewal of educational leadership, including the replacement of the education minister. While personnel changes alone will not solve all problems, a responsible figure is needed to design and implement a new direction for national education strategy. The new minister should present a plan for reorganizing the Ministry of Education, a university innovation plan, and a nationwide AI education strategy upon taking office, and be held accountable for specific outcomes.Reform must not remain a mere declaration. The government should establish a cross-government education reform promotion system within the next 100 days and implement regulatory reforms for university academics, research, and industry-academia collaboration within a year. A long-term plan to evaluate the accessibility of AI education, the cultivation of talent in advanced fields, and the research and employment outcomes of regional universities should also be developed every three years. If performance is lacking, policies, budgets, and responsible parties must be adjusted accordingly.South Korea achieved industrialization and democratization through the power of education. Now, we must achieve a new leap into the AI era through the power of education. The exam-centered education that enabled past successes does not guarantee future success.The education minister must be a future designer preparing for the next 30 years of the nation, not merely an administrative manager of exams and schools.The government must make a decision. It should comprehensively overhaul the organization and functions of the Ministry of Education and establish new leadership that can integratively lead scientific and technological innovation, university reform, industry-academia collaboration, and lifelong education.From a ministry that manages exams to one that creates the future of education. The replacement of the education minister and the major transformation of the education system can no longer be delayed.* This article has been translated by AI.
-
Shin Beom-cheol: 'Beware of North Korea's Combat Experience... Innovation Requires Discarding Yesterday's Successes'
Shin Beom-cheol, former Deputy Minister of National Defense, emphasized on November 20 the importance of utilizing artificial intelligence (AI) experts in the military, stating, "We need to reflect on whether we are training people, integrating technology, and creating systems to support technology with a sense of urgency."During his keynote speech at the '2026 Defense Industry Forum' held at the Ntree Premier Hotel in Yongsan, Seoul, Shin stressed that to innovate military capabilities using AI, the foundation of human resources and personnel systems must be restructured.This forum, co-hosted by Aju Economy, the Korea Defense Industry Association, and Jeonbuk National University, was themed 'The Future Battlefield Transformed by AI: A New Leap for K-Defense.'Shin shared his experience from when he took office as Deputy Minister of National Defense in May 2022, pointing out the shortage of AI experts in the defense sector. He recalled asking if there were any AI majors among the senior officials at the Ministry of National Defense, but no one responded, saying, "That is our reality."He then cited the case of an officer with a Ph.D. in AI from KAIST, whom he had placed in the Ministry of National Defense. Although the officer wanted to continue working in AI after being promoted to lieutenant colonel, he was pressured to take on a battalion commander position and eventually left the military for a university position.Shin noted, "The loss of specialized personnel is a disadvantage for our military," pointing out the conflict between the need for continuous development of expertise in technology fields and the existing military personnel system. He argued that it is essential to improve systems so that experts can build their capabilities in their respective fields, rather than merely paying lip service to the introduction of advanced technologies.He highlighted the core aspects of future battlefields transformed by AI, referencing Napoleon's statement on 'time, space, and combat power.' He explained that AI would support commanders' awareness, judgment, and coordination while also directly participating in combat through unmanned combat systems, thereby changing the way wars are conducted.Shin described the role of AI as enabling operations "faster, longer, and simultaneously." He predicted that the ability to shorten decision-making time, enhance the continuity of surveillance and tracking, and operate multiple unmanned systems would determine the advantage in future battlefields.In terms of space, he explained that manned and unmanned combined combat systems would expand the operational range of units and allow for the concentration of firepower at necessary points by connecting dispersed forces through a network.He also warned against underestimating North Korea's accumulation of combat experience. Shin stated that North Korea is gaining real combat experience through the Russia-Ukraine war, particularly in drone warfare, asserting that "the difference between having war experience and not having it is like night and day."He emphasized that North Korea's capabilities should not be evaluated based on past standards, urging a sense of crisis that they could fall behind in terms of time, space, and combat power. He particularly noted that "innovation begins with discarding yesterday's successes," a mindset that should also apply to the field of AI.Additionally, he stated that despite the export achievements of K-defense, the competitiveness in the AI and drone sectors should be assessed realistically. He explained that low-cost drones may offer less profit incentive for companies compared to large weapon systems, necessitating government investment and encouragement, as well as technology development by universities and small businesses.Shin identified defense data management as a key issue. He stated that the data produced during the activities of the Ministry of National Defense, the Defense Acquisition Program Administration, and frontline units forms the basis for AI development, and that it is essential to check whether this data is being properly managed.He also proposed improving the cooperative structure that connects civilian technology with the military. He pointed out that concerns about defense industry corruption can inhibit necessary communication between the government and businesses, and emphasized the need for the military to actively visit companies and universities to find technologies that can be utilized.Shin concluded that developing defense AI cannot be achieved solely by specific agencies, stating, "We must utilize all available data through a national mobilization effort, integrating contributions from small and large enterprises, academic societies, and more."* This article has been translated by AI.
-
Semes Expands Collaboration with Universities for Semiconductor Equipment Development
Semiconductor equipment company Semes is accelerating the development of next-generation semiconductor equipment by strengthening technological collaboration with universities and research institutions. Starting this year, the company is expanding practical research projects aimed at solving technical challenges in the industry, focusing on enhancing the applicability of research outcomes to actual equipment.On November 16, Semes announced that it held the 11th Korean Society of Mechanical Engineers-Semes Open Innovation Challenge (KSOIC) at the Jeju International Convention Center, where 21 outstanding research papers were selected.Launched in 2016, this competition seeks to explore the potential for joint technology development between industry and academia, focusing on topics such as semiconductor and display process equipment, as well as ultra-precision sensors and measurement technologies.This year, the introduction of 'KSOIC 2.0' increased the emphasis on real-world technical challenges faced by companies. A total of 65 proposals were submitted, including 40 practical research projects and 25 academic research projects from 34 universities and research institutions across the country.As semiconductor equipment becomes increasingly sophisticated, the importance of precision control and measurement technologies grows, highlighting the need for collaboration between academia and industry to connect fundamental research in mechanical engineering with actual equipment development.The grand prize was awarded to a team led by Professor Kim San-ha from KAIST for their proposal on an ultra-low-damage electric discharge micro-porous capillary gripper for pick-and-place of irregular packages. The gold prize was awarded to teams led by Professor Cho Young-tae from Changwon National University and a student team from UNIST.Over the past decade, Semes has selected more than 112 award-winning works, of which 15 have been pursued as joint research projects with academia. This has enabled the company to secure core technologies in precision measurement, sensors, and ultra-fine control for semiconductor equipment.Semes CEO Shim Sang-pil stated, "Starting this year, we have selected high-impact technical challenges through an external cooperation committee, and we will leverage academic research capabilities to strengthen the development of advanced semiconductor process equipment technologies."* This article has been translated by AI.