%43%4f
-
KAIST Develops Electrode Technology Achieving 86% Efficiency for Converting CO₂ into Plastic Precursors
<(From Left) Dr. Jonghyeok Park, Ph.D candidate Yunkyoung Han, Professor Hyunjoon Song, Dr. Sungjoo Kim>
KAIST Develops Electrode Technology Achieving 86% Efficiency for Converting CO₂ into Plastic Precursors
In the process of converting carbon dioxide into useful chemicals such as ethylene—a key precursor for plastics—a major challenge has been the flooding of electrodes, where electrolyte penetrates the electrode structure and reduces performance. KAIST researchers have developed a new electrode design that blocks water while maintaining efficient electrical conduction and catalytic reactions, thereby improving both efficiency and stability.
KAIST (President Kwang Hyung Lee) announced on the 6th of April that a research team led by Professor Hyunjoon Song from the Department of Chemistry has developed a novel electrode structure utilizing silver nanowire networks—ultrafine silver wires arranged like a spiderweb—to significantly enhance the efficiency of electrochemical CO₂ conversion to useful chemical products.
In electrochemical CO₂ conversion processes, a long-standing issue has been flooding, where the electrode becomes saturated with electrolyte, reducing the space available for CO₂ to react. While hydrophobic materials can prevent water intrusion, they typically suffer from low electrical conductivity, requiring additional components and complicating the system.
To overcome this, the research team designed a three-layer electrode architecture that simultaneously repels water and enables efficient charge transport. The structure consists of a hydrophobic substrate, a catalyst layer, and an overlaid silver nanowire (Ag NW) network, which acts as an efficient current collector while preventing electrolyte flooding.
< Schematic diagram of a porous polymer–copper oxide catalyst silver nanowire network electrode structure >
A key finding of this study is that the silver nanowires do more than just conduct electricity—they actively participate in the chemical reaction. During CO₂ reduction, the silver nanowires generate carbon monoxide (CO), which is then transferred to adjacent copper-based catalysts, where further reactions occur. This creates a tandem catalytic system, in which two catalysts cooperate sequentially, significantly enhancing the production of multi-carbon compounds such as ethylene.
The electrode demonstrated outstanding performance. It achieved 79% selectivity toward C₂₊ products in alkaline electrolytes and 86% selectivity in neutral electrolytes, representing a world-leading level. It also maintained stable operation for more than 50 hours without performance degradation. These results indicate that most of the converted products are the desired chemicals, while also overcoming the durability limitations of conventional systems.
< Conceptual diagram of a CO₂ electrolysis electrode utilizing a stacked silver nanowire structure (AI-generated image) >
Professor Hyunjoon Song stated, “This study is significant in showing that silver nanowires not only serve as electrical conductors but also directly participate in chemical reactions,” adding, “This approach provides a new design strategy that can be extended to converting CO₂ into a wide range of valuable products such as ethanol and fuels.”
This research, led by Jonghyeok Park (KAIST, first author), was published on March 24, 2026, in the international journal Advanced Science.
※ Paper title: “Overlaid Conductive Silver Nanowire Networks on Gas Diffusion Electrodes for High-Performance Electrochemical CO₂-to-C₂₊ Conversion,” DOI: https://doi.org/10.1002/advs.75003
2026.04.07 View 173 -
Undergraduate Rover Team (MR2) Advances to Finals of 'URC 2026', the World’s Largest Mars Rover Competition
<Photo: KAIST Undergraduate Club MR2 Team Members>
Undergraduate students from KAIST are set to take on the world stage with an exploration rover—a robotic vehicle designed to explore in place of humans—that they built themselves. The team has secured a spot in the finals of the world’s largest Mars rover competition, marking a first-ever achievement for KAIST.
KAIST announced on the 3rd that 'MR2' (Advised by Professor Yong-Hwa Park, Department of Mechanical Engineering), a rover team from the undergraduate robotics club MR (Microrobot Research), has earned a seed in the finals of the '2026 University Rover Challenge (URC)', the premier international Mars rover competition for university students.
The URC is organized by The Mars Society and takes place at the Mars Desert Research Station (MDRS) in Utah, USA, an environment that closely mimics the Martian surface. Participating teams compete in four key missions using rovers they developed: ▲Science Mission, ▲Delivery Mission, ▲Equipment Servicing Mission, and ▲Autonomous Navigation Mission.
This year’s competition saw 116 university teams from 18 countries engage in a fierce preliminary round. Team MR2 secured its place in the top 38 finalists by scoring 95.38 out of 100. This milestone is particularly significant as it is the first time a KAIST team has ever reached the URC finals, proving the excellence of KAIST undergraduates in robot design and control on a global scale.
The next-generation exploration rover 'GAP-1000', independently developed by MR2, is a modular rover designed for stable operation in extreme environments. It features a 6-DOF (Degrees of Freedom) robotic arm capable of precisely controlling objects over 5kg, allowing it to perform complex equipment manipulation tasks.
<Photo: Operation of GAP-1000's Manipulator and Science Module Integration>
The rover also boasts strong autonomous driving capabilities. By combining RTK-GNSS (precision satellite positioning), IMU (Inertial Measurement Units) for motion sensing, and odometry based on wheel rotation, it can autonomously navigate optimal paths through complex terrain. Additionally, a drone relay system has been integrated to ensure stable exploration even in areas with communication dead zones.
For the science mission, the rover can collect soil from 10cm underground, remove impurities via centrifugation, and analyze traces of life using protein detection reagents such as Biuret and Bradford. This is paired with spectroscopic analysis technology that identifies material composition by analyzing light wavelengths, creating an integrated system for real-time life detection.
"We experienced a lot of trial and error while managing everything from design to production ourselves, but I am thrilled that we achieved KAIST’s first-ever advancement to the finals," said Myung-woo Jung (Department of Mechanical Engineering), the team leader of MR2. "We will prepare thoroughly in the remaining time to achieve a great result on-site."
<Photo: Scenery of MDRS in Utah, USA, where the competition will be held (Photo Credit: The Mars Society)>
Advising Professor Yong-Hwa Park noted, "It is impressive that the students independently implemented a rover for extreme environments. This competition will serve as an opportunity to showcase KAIST’s technological prowess to the world."
KAIST President Kwang-Hyung Lee added, "It is a very meaningful achievement for our undergraduates to reach the finals of the world’s largest competition with a rover they designed and built themselves. I hope this experience serves as a catalyst for our students to challenge themselves and grow on the global stage."
Team MR2 consists of 13 undergraduate students from various majors, including Mechanical Engineering, Electrical Engineering, and Industrial Design. Having completed long-distance operation tests in outdoor environments, they are currently conducting final checks for the finals. The main competition will be held from May 27th to 30th at the MDRS in Utah, USA.
※ Related Links
MR2 Official Website: https://urc-kaist.github.io/
MR2 Instagram: https://www.instagram.com/urc_mr2/
MR2 YouTube: https://www.youtube.com/@MR2KAISTRoverTeam
2026.04.03 View 262 -
KAIST Professor Jinjoon Lee’s 10-Meter Hanji Scroll PhD Thesis from Oxford Enters the Permanent Collection of the World’s Oldest Museum, First Work by a Contemporary Korean Artist
<A ten-metre scroll doctoral thesis reinterpreting the 15th-century Joseon landscape painting scroll tradition, Empty Garden, exhibited at the University Church of St Mary the Virgin, Oxford, founded in the 15th century. 2020>
- Media artist and KAIST professor Jinjoon Lee's doctoral thesis 'Empty Garden' officially acquired by the Ashmolean Museum, UK, for permanent collection
- Korean artistic and academic achievement recognized as public cultural heritage at a museum predating the Louvre by 110 years — the 'heart of Western intellectual history'
- Blending Eastern aesthetics of 'wandering' (거닐기) and 'emptiness' with data technology in the AI era — awarded Oxford's unanimous 'No Corrections' in just 2.5 years in 2020
- First work by a contemporary Korean artist to enter the Ashmolean's permanent collection — officially confirmed by the museum's curator
- Korean artistic and academic achievement officially recognised as intellectual cultural heritage — permanently preserved, researched, and exhibited within the Western public knowledge system
A doctoral thesis is often imagined as a dense, bound volume. Yet a 10-meter-long hanji scroll- traditional Korean mulberry paper prized for its durability across centuries- is now drawing global attention from the art world and academia alike.
KAIST (President Kwang Hyung Lee) announced on the 26th that Empty Garden – A Liminoid Journey to Nowhere in Somewhere (2020), a doctoral thesis by media artist and KAIST Graduate School of Culture Technology Professor Jinjoon Lee, has been officially acquired by the Ashmolean Museum, University of Oxford, for its permanent collection and exhibition — through formal purchase, not donation.
Founded in 1683, the Ashmolean Museum is the world's first university museum, operated by the University of Oxford with over 340 years of history. It predates the Louvre (1793) by 110 years and the British Museum (1759) by 76 years, and is regarded as the starting point of European Enlightenment scholarship. Its collections include masterworks by Raphael, Michelangelo, Leonardo da Vinci, and Turner, alongside ancient artefacts and East Asian ceramics and paintings — over one million objects in total.
The Ashmolean is not simply an exhibition venue but an academic institution integrating collection, research, and education. Unlike Tate Modern, which engages with the contemporary art market, or the British Museum, which displays national heritage, the Ashmolean's core mission is scholarly preservation and research. The acquisition of Professor Lee's doctoral thesis here signifies that Korean aesthetics and philosophical thought have entered the public record of European intellectual history.
Professor Lee's PhD thesis Empty Garden reinterprets the concept of uiwon (意園) — an imaginary garden cultivated in the mind by Joseon-era scholars — through contemporary data and media language, proposing 'data gardening' as a methodology for tending to the philosophy of emptiness. It is a work that continues to ask fundamental questions about human sensation, memory, and existence, even within an environment dominated by AI and data.
The 10-meter hanji scroll format is itself a central feature of the thesis. As readers engage with the text, they are naturally led to move through space — physically enacting the East Asian garden tradition of 'wandering' (거닐기). The work is designed not merely to be read but to be experienced through movement and contemplation. The thesis was produced as nine hanji scrolls in total; one of these has been acquired by the Ashmolean for its permanent collection.
This thesis received unanimous 'No Corrections' approval at its DPhil in Fine Art examination at the University of Oxford in 2020, recognising its academic rigour and originality — an achievement completed in just two and a half years, where the process typically takes over four. It is an extremely rare distinction even within Oxford's 900-year history, and drew significant attention at the time.
Oxford doctoral theses are typically archived at the Bodleian Library as academic records. This acquisition is entirely separate from that process: the museum conducted an independent five-year review following the award of the degree, assessed the work on its artistic and scholarly merits, and made a formal purchase. The inclusion of a living artist's doctoral thesis in the permanent collection of the world's oldest university museum through purchase — not donation — is exceptionally rare.
Professor Shelagh Vainker, Alice King Curator of Chinese and Korean Art at the Ashmolean Museum, University of Oxford, stated:
"I am delighted that the Ashmolean Museum has been able to acquire Dr Jinjoon Lee's Empty Garden for our permanent collection. The long, contemplative scroll breaks new ground in so many ways: in the materials and techniques employed, in the breadth and depth of cultural and intellectual knowledge embedded in it, and in the complexity of the presentation of different spaces — all providing the viewer with multiple perspectives and experiences. Empty Garden is the first piece by a contemporary Korean artist to enter the collection; when not on display it will be available for viewing by appointment."
— Shelagh Vainker, Alice King Curator of Chinese and Korean Art, Ashmolean Museum, University of Oxford
<Dr Shelagh Vainker, Professor at the University of Oxford and Alice King Curator of Chinese and Korean Art at the Ashmolean Museum, reviewing the doctoral thesis Empty Garden in the Eastern Art Study Room, Ashmolean Museum. 2026>
Professor Lee noted that during his doctoral research at Oxford, a serious leg injury left him using a wheelchair for an extended period, during which he reflected deeply on the relationship between movement, stillness, and thought. He stated: "In the age of AI, art cannot remain confined to immaterial images on screens. Data and images can only acquire depth through material forms capable of enduring time and preservation."
He further expressed his hope that Empty Garden, now housed within the public collection of Western intellectual history, would "serve as a continuing reference point connecting East Asian thought — including that of Korea — with new sensory frameworks for the age of artificial intelligence."
The first practicing artist to be appointed as a tenure-track professor at KAIST, Professor Lee currently holds concurrent positions as Visiting Fellow at Exeter College, University of Oxford, Visiting Senior Researcher at Tokyo University of the Arts, and Adjunct Professor at New York University, continuing interdisciplinary research across art, technology, and the humanities. Most recently, his work has drawn international attention from arts community, including Good Morning, Mr. G-Dragon, a space art project based on the iris data of K-pop artist G-Dragon, and Cine Forest: Awakening Bloom, an AI-based media symphony at Bundang Central Park in S. Korea.
<Jinjoon Lee, artist's studio, Seoul. 2025>
This acquisition is an exceptionally rare case of a doctoral thesis entering the permanent collection of the world's oldest university museum through formal purchase, and a historic event in which a work by a contemporary Korean artist has entered the Ashmolean's collection for the first time. Korean research that poses new questions about the role of art and the humanities in the post-AI era has now found a permanent place within the public record of Western intellectual history.
2026.03.26 View 505 -
KAIST Reveals the Formation Mechanism of Skyrmions Inside Magnets… A Clue to Solving AI Power Consumption
<(From Left) Prof.Se Kwon Kim, Dr. Gyungchoon Go>
“Skyrmions,” in which electron spins inside a magnet are arranged like vortices, are a key structure in next-generation spintronics technology. KAIST researchers have shown that skyrmions can form using only the fundamental physical interactions within magnets, without requiring special physical conditions. This finding expands the possibility of realizing skyrmions in a wide range of magnetic materials and suggests new potential for developing next-generation ultra-low-power information devices with data storage densities tens to hundreds of times higher than current technologies.
KAIST (President Kwang Hyung Lee) announced on the 19th of March that a research team led by Professor Se Kwon Kim from the Department of Physics has proposed a new theoretical framework showing that vortex-like magnetic structures can naturally emerge solely through magnetoelastic coupling—the interaction between magnetism and lattice structure.
The team demonstrated that the interaction between spins (the intrinsic magnetic property of electrons) and lattice deformation (the slight distortion of atomic arrangements) alone can lead to the spontaneous formation of vortex-like magnetic structures.
In particular, skyrmions—vortex-like spin structures found inside magnetic materials—are extremely small and highly stable, making them promising candidates for ultra-high-density, low-power information devices. However, until now, forming such structures was believed to require specific physical conditions such as crystal asymmetry or strong spin–orbit coupling.
The researchers theoretically showed that even without such special conditions, magnetoelastic coupling, which naturally occurs in most magnetic materials, is sufficient to generate a structure in which skyrmions and antiskyrmions are alternately arranged.
Magnetoelastic coupling refers to the phenomenon in which magnetism (spin) and lattice deformation influence each other, and it is a fundamental physical property present in nearly all magnetic materials. The team showed that when this coupling becomes sufficiently strong, the original ground state—where magnetization is uniformly aligned—becomes unstable and transitions into a new vortex-like ordered state.
In this process, they proposed a new mechanism in which spin tilting and lattice distortion occur simultaneously, forming a chiral spin texture composed of alternating skyrmions and antiskyrmions.
Professor Se Kwon Kim explained, “This study demonstrates that skyrmion-like magnetic structures can form even without specific or exotic interactions. It is particularly meaningful in that it suggests the possibility of realizing such structures in two-dimensional magnetic materials, where research is currently very active.”
This study was led by Gyungchoon Go, who participated as the first author. The research was published on February 11 in the internationally renowned journal Physical Review Letters, recognizing its significance in the field of physics.
※ Paper title: “Magnetoelastic Coupling-Driven Chiral Spin Textures: A Skyrmion-Antiskyrmion-like Array,” DOI:https://doi.org/10.1103/5csz-pw7x
※ Main Authors: Gyungchoon Go (first author), Se Kwon Kim (corresponding author)
This research was supported by the Samsung Science and Technology Foundation, the Brain Pool Plus Program by the National Research Foundation of Korea, and the Sejong Science Fellowship.
2026.03.19 View 578 -
World’s First SoulMate AI Semiconductor: A Personalized Digital Soulmate Developed
< (From left) KAIST Professor Hoi-Jun Yoo and PhD candidate Seongyon Hong >
While Large Language Models (LLMs) like ChatGPT are adept at answering countless questions, they often remain unaware of a user's minor habits or previous conversational contexts. This is why AI, despite being deeply integrated into our daily lives, can still feel like a "stranger." Overcoming these limitations, researchers at KAIST have developed the world’s first AI semiconductor, dubbed "SoulMate," which learns and adapts to a user’s speech style, preferences, and emotions in real-time—becoming a true "digital soulmate."
KAIST announced on March 17th that a research team led by Professor Hoi-Jun Yoo from the Graduate School of AI Semiconductors has developed SoulMate, a personalized LLM accelerator that evolves according to the specific characteristics of the user.This technology is being hailed as a core semiconductor breakthrough that will accelerate the era of "Hyper-Personalized AI"—moving beyond "AI for everyone" to an AI that learns and responds to an individual's unique conversational style and preferences.
The core of SoulMate lies in On-Device AI technology, which processes data directly on the device without going through external servers (the cloud). The team directly implemented Retrieval-Augmented Generation (RAG), which generates customized answers based on remembered conversations, and Low-Rank Adaptation (LoRA), which immediately reflects and learns from user feedback, within the semiconductor itself.
< SoulMate AI Semiconductor Chip >
Through this, SoulMate has realized a real-time personalized AI system that responds to the user at a staggering speed of 0.2 seconds (216.4 ms) while simultaneously performing learning tasks.
< SoulMate Application Demo >
Furthermore, the team applied a Mixed-Rank architecture that optimizes processing methods based on the importance of information, drastically reducing power consumption. The semiconductor operates at an ultra-low power of just 9.8 milliwatts (mW)—approximately 1/500th of a typical smartphone processor's power consumption—allowing it to handle complex learning and inference simultaneously on mobile devices without battery concerns.
In particular, SoulMate features a "Security-Complete AI" structure where all personal data is processed internally within the device rather than being transmitted to external servers, fundamentally blocking any risk of personal information leaks. The research team expects this technology to pair with next-generation platforms such as smartphones, wearables, and personal AI devices to open a true era of personalized AI services.
< SoulMate Demo Screen >
"This research mimics the process of people building friendships, providing the technical foundation for AI to evolve into a true companion for the user," said Professor Hoi-Jun Yoo. "Future AI will move beyond being a mere tool to become a 'Best Friend' that understands me best anytime, anywhere, while perfectly protecting personal privacy."
The study, with PhD student Seongyon Hong as the first author, was selected as a "Highlight Paper" at the International Solid-State Circuits Conference (ISSCC) held in San Francisco this past February, garnering significant attention from the global academic community.
Paper Title: SoulMate: A 9.8mW Mobile Intelligence System-on-Chip with Mixed-Rank Architecture for On-Device LLM Personalization Authors: Seongyon Hong, Jiwon Choi, Jeonggyu So, Nayeong Lee, Wooyoung Jo, Zhamaliddin Kalzhan Link: https://ieeexplore.ieee.org/document/11409048
At the conference, the research team successfully demonstrated how the AI's response style changes in real-time according to user reactions using the actual semiconductor chip, proving the excellence of Korean AI semiconductor technology. The SoulMate AI semiconductor is planned for commercialization around 2027 through the faculty-led startup "OnNeuro AI."
< SoulMate Demonstration Photo >
This research was conducted with support from the Information and Communication Broadcast Innovation Talent Cultivation Program of the Ministry of Science and ICT and the Institute of Information & Communications Technology Planning & Evaluation (IITP).
2026.03.17 View 894 -
KAIST Solves the 500-Year-Old ‘Pain’ Behind Michelangelo’s painting of The Creation of Adam
<(From Left) Ph.D candidate Minwoo Choi, Ph.D candidate Hyejoon Jun, Professor Hyoungsoo Kim>
More than 500 years ago, Michelangelo spent four years painting The Creation of Adam on the ceiling of the Sistine Chapel, struggling with paint dripping onto his face. He described the process as “closer to torture than painting.” Now, researchers at KAIST have developed a technology that can effectively “hold up falling paint.” Beyond ceiling paintings, this principle could help solve the problem of liquid films collapsing on inclined surfaces, with potential applications in precision coating, electronic circuit printing, 3D printing, and fluid control in space environments.
KAIST (President Kwang Hyung Lee) announced on the 12th of March that a research team led by Professor Hyoungsoo Kim of the Department of Mechanical Engineering has reinterpreted the fundamental cause of downward flow under gravity—known as gravitational instability—from the perspective of interfacial fluid mechanics* and proposed a method to control it by mixing a small amount of volatile liquid into a suspended liquid film.*Interfacial fluid mechanics: the study of the balance of microscopic forces acting at the surface of liquids.
Why was it so difficult for Michelangelo to paint on the ceiling? When paint is applied to a ceiling, a thin liquid film forms. However, this film gradually becomes unstable due to gravity and eventually drips down. This phenomenon is common in everyday life.
For example, when steam condenses on a bathroom ceiling, it first forms a thin layer of water that eventually gathers into droplets and falls. Similarly, droplets that appear on the ceiling of a refrigerator initially form a thin layer but gradually grow and begin to drip downward. This type of instability, where liquid accumulated on an upper surface collapses under gravity, is known as Rayleigh–Taylor instability. Until now, it has generally been considered unavoidable in the presence of gravity.
The research team proposed mixing a small amount of volatile liquid into the suspended liquid film. As the volatile component evaporates, it changes the concentration distribution along the liquid surface, creating differences in surface tension. Surface tension is the force that pulls a liquid surface inward, which is why water droplets maintain a rounded shape.
When differences in surface tension arise, the region with stronger tension pulls liquid toward itself from regions with weaker tension. This creates a surface flow known as the Marangoni effect. Through experiments and theoretical analysis, the researchers demonstrated that this surface flow can effectively hold the liquid in place and suppress the gravitational instability that would otherwise cause it to fall.
A familiar example can illustrate this effect. If pepper powder is sprinkled evenly on the surface of water, it remains floating. However, if a drop of detergent is placed in the center, the pepper suddenly moves outward toward the edges. This happens because the detergent reduces the surface tension where it touches the water, allowing the surrounding regions with stronger surface tension to pull the liquid outward. As the surface flow develops, the pepper particles move along with it.
In this study, evaporation of the volatile liquid created a similar surface tension difference. But instead of pushing particles outward like in the pepper example, the flow pulled the liquid upward, counteracting the force that would otherwise cause it to drip downward.
As a result, under certain conditions the liquid film remained intact despite gravity. In some cases, the researchers even observed a new behavior in which droplets did not fall but the liquid film oscillated periodically. This demonstrates that gravitational instability can be actively controlled using only natural processes—such as liquid composition and evaporation—without any external energy input.
This principle could enable thinner and more uniform liquid films in precision coating, printing, and layer-by-layer manufacturing processes, allowing stable coating even on tilted surfaces. It may also extend to technologies such as 3D printing and fluid control in specialized environments like space. In essence, the physical limitation that Michelangelo struggled with 500 years ago may now inspire future industrial technologies.
The Creation of Adam (AI-generated image)>
Professor Hyoungsoo Kim stated, “Rayleigh–Taylor instability has long been considered unavoidable as long as gravity exists. This research is meaningful because it shows that gravitational instability can be actively controlled without external energy by utilizing natural processes such as liquid composition and evaporation.” He added, “This principle could extend beyond coating, printing, and layering processes to fluid control technologies in space environments.”
This study was led by Minwoo Choi, an integrated master’s–PhD student in Mechanical Engineering, as the first author. The discovery, recognized as a new finding in the control of hydrodynamic instability, was published online on January 29 in the international journal Advanced Science (Wiley) and was selected as a Frontispiece article.
※ Paper title: “Evaporation-Driven Solutal Marangoni Control of Rayleigh–Taylor Instability in Inverted Films,” Authors: First author Minwoo Choi (KAIST), co-author Hyejoon Jun (KAIST), corresponding author Hyoungsoo Kim (KAIST), DOI: https://doi.org/10.1002/advs.202520343
This research was supported by the Mid-Career Researcher Program of the National Research Foundation of Korea (MSIT: 2021R1A2C2007835)
2026.03.12 View 610 -
KAIST Develops Self-Regenerating Catalyst That Restores Its Own Performance, Opening a Breakthrough for CO₂ Conversion Technology
<(From Left) Professor Dong Young Chung, Ph.D Candidate Hongmin An, Hanjoo Kim>
Technologies that convert carbon dioxide (CO₂) emitted from factories and power plants into useful chemical feedstocks are considered key to achieving carbon neutrality. However, rapid degradation of catalyst performance has long hindered commercialization. KAIST researchers have now developed a “self-regenerating” catalyst that restores its activity during operation, offering a potential solution to this challenge.
KAIST (President Kwang Hyung Lee) announced on the 11th of March that a research team led by Professor Dong Young Chung from the Department of Chemical and Biomolecular Engineering has identified the fundamental cause of catalyst degradation in electrochemical reactions that convert CO₂ into useful materials and has developed a new design strategy that allows catalysts to maintain their active state during the reaction.
<Schematic Illustration of Copper Catalyst Reconstruction>
The research team focused particularly on copper (Cu) catalysts, which are widely used in CO₂ conversion reactions. Copper catalysts are known not to simply degrade during reactions but instead undergo a process called surface reconstruction, in which their surface structure continuously changes. The study revealed that the performance and lifetime of the catalyst vary significantly depending on how this reconstruction occurs.
The researchers discovered that copper catalyst reconstruction occurs mainly through two different mechanisms. The first involves formation and reduction of oxide layers on the catalyst surface. While this temporarily increases catalytic activity, it ultimately leads to long-term degradation of catalyst performance.
The second mechanism involves partial dissolution of the catalyst metal into the electrolyte followed by redeposition onto the catalyst surface. During this process, new reactive sites—known as active sites—are continuously created on the catalyst surface.
Based on this mechanism, the team proposed a method that allows the catalyst to maintain its active state during the reaction. By introducing a trace amount of copper ions into the electrolyte, dissolution and redeposition of copper occur in a balanced cycle on the catalyst surface. This continuous cycle generates new active sites, enabling the catalyst to maintain stable performance over extended periods.
Importantly, this technology can be implemented without complex additional processes or high-voltage conditions, significantly reducing energy consumption while enabling stable production of high-value C₂ compounds such as ethylene and ethanol. C₂ compounds are molecules containing two carbon atoms and are industrially important chemicals used as feedstocks for plastics, fuels, and other materials.
This research is significant because it proposes a new design concept in which catalysts are not merely optimized at the initial stage but are engineered to maintain their optimal state throughout the reaction process. The concept is expected to be applicable not only to CO₂ conversion technologies but also to a wide range of electrochemical energy conversion systems.
Professor Dong Young Chung stated, “This research approached catalyst degradation not as an inevitable phenomenon but as a controllable process,” adding, “We proposed a new strategy that allows catalysts to continuously maintain optimal activity during the reaction.”
The study was led by Hanjoo Kim, a doctoral student at KAIST, and Hongmin An, a combined master’s-doctoral student, as co-first authors. The research was published online on February 5 in the Journal of the American Chemical Society (JACS), one of the world’s most prestigious journals in chemistry.
※ Paper title: “Dynamic Interface Engineering via Mechanistic Understanding of Copper Reconstruction in Electrochemical CO₂ Reduction Reaction” DOI: 10.1021/jacs.5c16244
This research was supported by the Global Young Connect Program for Materials and the National Strategic Materials Technology Development Program funded through the National Research Foundation of Korea.
2026.03.11 View 878 -
KAIST Team Led by Dong-won Lee Wins Grand Prize at the 2nd Global Quantum AI Competition
< (From Left) M.S candidate Dongwon Lee from School of Electrical Engineering, Ph.D candidate Jaehun Han from Graduate School of Quantum Science and Technology >
"Team Yangja-jorim," consisting of Dongwon Lee, Gyungjun Kim and Jaehun Han , has been honored with the Grand Prize at the '2026 2nd Global Quantum AI Competition.' The event was hosted and organized by NORMA, a specialized quantum computing company.
This global competition was designed to expand hands-on experience with quantum cloud services and to discover next-generation talent in the field of quantum artificial intelligence. The event spanned approximately 70 days, beginning with the preliminary opening ceremony held at Korea University’s Hana Square on December 17 last year. The final winners were announced during an awards ceremony held at NORMA's headquarters on the 27th of last month.
The competition attracted significant interest from quantum technology talent worldwide, including university students, developers, and researchers. A total of 137 teams participated in the preliminaries, with the top 10 teams advancing to the finals—a competitive ratio of approximately 13.7 to 1.
< An acquaintance attended the awards ceremony of the 2nd Global Quantum AI Competition to accept the prize on behalf of the team. >
In the final round, participants were presented with four generative problems utilizing the Quantum Circuit Born Machine (QCBM) model. To overcome the current limitations of quantum machine learning, the contestants were tasked with designing and validating Quantum-Classical Hybrid Generative AI models that integrate classical techniques. Notably, the final problem provided an opportunity to verify the proposed methods using a real Quantum Processing Unit (QPU) from Rigetti Computing, a leading global quantum computing firm.
The judging process employed a double-blind system, where the identities of both evaluators and participants remained undisclosed to ensure maximum fairness and credibility.
"Through this competition, we were able to explore the research potential of the quantum AI field more deeply," said KAIST's Team Yangja-jorim in their acceptance speech. "We hope to continue contributing to the advancement of quantum technology through consistent research and new challenges."
2026.03.10 View 894 -
AI Developed to Locate Slums Worldwide... Wins Best Paper Award at AAAI 2026
<(From Left) Sumin Lee, Sungwon Park, Prof. Jihee Kim, Prof. Meeyoung Cha, Prof. Jeasurk Yang>
"Cities don't even know where their slums (impoverished areas) are located."
In many developing nations, the most vulnerable citizens are invisible to the state simply because their homes don't appear on any official map. Today, a breakthrough using Artificial Intelligence (AI) is changing that.
A joint research team from KAIST and Chonnam National University in South Korea and MPI-SP in Germany has developed an AI technology that autonomously identifies slum areas using nothing but satellite imagery. This technology is expected to fundamentally transform urban policy-making and public resource allocation in developing countries where data is scarce and has won the Best Paper Award in the ‘AI for Social Impact’ category at the AAAI 2026 (Association for the Advancement of Artificial Intelligence), the world's premiermost prestigious AI academic conference.
Why it Matters
While previous studies struggle to recognize slums across countries due to varying architectural styles, the team introduced a "Mixture-of-Experts (MoE)" structure. In this system, multiple AI models learn different regional characteristics; when a new city is inputted, the system automatically selects the most appropriate model.
<Figure1. Overview of the Mixture-of-Experts(MoE) structure to identify slum areas>
The core of this research is "Test-Time Adaptation (TTA)" technology. Even if humans do not pre-mark slum locations in a new city, the AI reduces its own errors by comparing and verifying the prediction results of multiple models, trusting only the areas where they commonly agree. This ensures stable performance even in regions with insufficient data.
The research team applied this technology to major cities such as Kampala (Uganda) and Maputo (Mozambique) and confirmed that it distinguishes slum areas more precisely than existing state-of-the-art technologies.
This technology is expected to be utilized in various policy fields, including:
Establishing urban infrastructure expansion plans for developing countries.
Identifying areas vulnerable to disasters and infectious diseases in advance.
Selecting targets for housing environment improvement projects.
Monitoring the implementation of UN Sustainable Development Goals (SDGs).
<Figure2. Slum segmentation results in Kampala in 2015 (yellow) and 2023 (red). Over the eight-year period, the slum ratio in the city increased from 8.4% to 8.6%>
Meeyoung Cha, an AI researcher and author, stated, "This research proves that AI is no longer just a tool for analysis. It is a tool for action. Our technology can bridge the data gap to solve the world’s most pressing social challenges." Jihee Kim, an economist and author, added, "It will complement costly field surveys and help effectively allocate limited resources to the areas that need them most."
The research results were presented at AAAI 2026 in Singapore on January 25th.
Paper Title: Generalizable Slum Detection from Satellite Imagery with Mixture-of-Experts
Paper Link: https://aaai.org/about-aaai/aaai-awards/aaai-conference-paper-awards-and-recognition/
This research was supported by the National Research Foundation of Korea (NRF) through the Mid-career Researcher Support Program and the Data Science Convergence Human Resources Training Program.
2026.03.08 View 694 -
Professor Kuk-Jin Yoon’s Research Team at the Department of Mechanical Engineering Achieves Landmark Success with 10 Papers Accepted at CVPR 2026
<Professor Kuk-Jin Joon from Department of Mechanical Engineering>
Professor Kuk-Jin Yoon’s research team from our university’s Department of Mechanical Engineering has once again demonstrated its overwhelming academic prowess by having a total of 10 papers accepted as lead authors at the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2026 (CVPR 2026).
CVPR is the most influential international conference in the fields of artificial intelligence and visual intelligence. Since its inception in 1983, it has selected outstanding research through a rigorous peer-review process every year. For CVPR 2026, a total of 16,092 papers were submitted worldwide, with 4,090 accepted, resulting in a competitive acceptance rate of approximately 25.42%. Achieving 10 accepted papers as lead or corresponding authors from a single laboratory is regarded as an exceptionally rare and world-class feat.
Professor Kuk-Jin Yoon’s team conducts extensive research with the ultimate goal of achieving human-level visual intelligence. The papers accepted this year cover cutting-edge topics in computer vision, including:
Event camera-based technologies
Perception technologies for autonomous driving
AI optimization and adaptation techniques
This achievement follows the team's remarkable success at ICCV 2025 last year, where they published 12 papers as lead/corresponding authors. The results at CVPR 2026 further solidify the laboratory's position as a global hub for pioneering computer vision research. The research team plans to continue contributing to the advancement of future AI technologies by tackling challenging research that transcends the limitations of existing methods.
Meanwhile, CVPR 2026 is scheduled to be held in Denver, Colorado, USA, from June 3 to June 7.
<CVPR 2026 (Denver, USA)>
2026.03.06 View 1540 -
KAIST Develops Brain-Like AI… Thinks One More Time Even When Predictions Are Wrong
<(From left) Professor Sang Wan Lee, Myoung Hoon Ha, and Dr. Yoondo Sung>
Artificial intelligence now plays Go, paints pictures, and even converses like a human. However, there remains a decisive difference: AI requires far more electricity than the human brain to operate. Scientists have long asked the question, “How can the brain learn so intelligently using so little energy?” KAIST researchers have moved one step closer to the answer.
KAIST (President Kwang Hyung Lee) announced on the 29th that a research team led by Distinguished Professor Sang Wan Lee of the Department of Brain and Cognitive Sciences has developed a new technology that applies the learning principles of the human brain to deep learning, enabling stable training even in deep artificial intelligence models.
Our brain does not passively receive the world. Instead of merely perceiving what is happening in the present, it first predicts what will happen next and, when reality differs from that prediction, adjusts itself to reduce the difference (i.e., prediction error). This is similar to anticipating an opponent’s next move in Go and changing strategy if the prediction turns out to be wrong. This mode of information processing is known as “Predictive Coding.”
< Predictive Coding (PC) Module >
Scientists have attempted to apply this principle to AI, but encountered difficulties. As neural networks become deeper, errors tend to concentrate in specific layers or vanish altogether, repeatedly leading to performance degradation.
The research team mathematically identified the cause of this problem and proposed a new solution. The key idea is simple: instead of predicting only the final outcome, the AI is designed to also predict how its prediction errors will change in the future. The team refers to this as “Meta Prediction.” In simple terms, it is an AI that “thinks once more about its mistakes.” When this method was applied, learning proceeded stably in deep neural networks without halting.
<Analysis of Instability in Predictive Coding Model Errors>
The experimental results were also impressive. In 29 out of 30 experiments, the proposed method achieved higher accuracy than the current standard AI training method, backpropagation. Backpropagation is the representative learning method in which AI “goes backward by the amount of error and corrects it.”
Conventional AI training methods (backpropagation) require tightly interconnected layers, meaning the entire network must be computed and updated simultaneously. In contrast, this new approach demonstrates that, like the brain, large AI models can be effectively trained even when learning occurs in a distributed and partially independent manner.
<Performance Comparison of Predictive Coding Models>
This technology is expected to expand into various fields where power efficiency is critical, including neuromorphic computing, robot AI that must adapt to changing environments, and edge AI operating within devices.
Distinguished Professor Sang Wan Lee stated, “The key to this research is not simply imitating the structure of the brain, but enabling AI to follow the brain’s learning principles themselves,” adding, “We have opened the possibility of artificial intelligence that learns efficiently like the brain.”
This study was conducted with Dr. Myoung Hoon Ha as the first author and Professor Sang Wan Lee as the corresponding author. The paper was accepted to the International Conference on Learning Representations (ICLR 2026) and was published online on January 26.
※ Paper title: “Stable and Scalable Deep Predictive Coding Networks with Meta Prediction Errors”Original paper: https://openreview.net/forum?id=kE5jJUHl9i¬eId=e6T5T9cYqO
This research was supported by the Ministry of Science and ICT and the Institute of Information & Communications Technology Planning & Evaluation (IITP) through the Digital Global Research Support Program (joint research with Microsoft Research), the Samsung Electronics SAIT NPRC Program, and the SW Star Lab Program.
2026.03.06 View 1259 -
2026 KAIST Commencement: Shining Their Own Light on Their Respective Stages
KAIST (President Kwang Hyung Lee) announced that it held its 2026 Commencement Ceremony at 2 p.m. on February 20th at the Sports Complex on its Main Campus in Daejeon.
At this year’s ceremony, a total of 3,334 graduates received degrees, including 817 doctoral, 1,792 master’s, and 725 bachelor’s degrees. Since its founding in 1971, KAIST has now produced a total of 84,490 highly qualified science and technology professionals, including 18,130 Ph.D. recipients, 43,358 master’s graduates, and 23,002 bachelor’s graduates.
KAIST selected three representative graduates who embody the university’s vision of talent. They are Seunghyun Ryu (Department of Bio and Brain Engineering), the doctoral representative known as the “pianist neuroscientist” for his interdisciplinary research bridging brain science and piano performance; Jeanne Choi (School of Computing), the master’s representative who has pursued warm and inclusive technologies for socially vulnerable groups under the themes of accessibility and inclusion; and Mert Yakup Baykan (Department of Aerospace Engineering), the bachelor’s representative from Cyprus holding Turkish nationality, who became the first international recipient of the KAIST Presidential Scholarship.
Seunghyun Ryu, selected as both the doctoral representative and one of the notable graduates, spent 14 years at KAIST completing his undergraduate through doctoral studies while balancing research and music. He organized and managed performances through the campus piano club “PIAST,” expanding artistic activities within the campus community. His research explored the inverse relationship between Alzheimer’s disease and cancer, revealing how disease-related proteins and anticancer drugs act in neurons and offering new perspectives on inter-disease connections.
Jeanne Choi, the master’s representative and another notable graduate, presented research at AAATE 2023 in Paris, analyzing the experiences of visually impaired users engaging with the metaverse and artificial intelligence. Accompanying a visually impaired professor during the conference, Choi gained firsthand insight into mobility and safety challenges, which further expanded the scope of her research. Choi has since continued field-based research, including serving as a teaching assistant at AI and coding camps for visually impaired youth, and plans to pursue a doctoral degree while continuing research for socially vulnerable communities.
Bachelor’s representative Mert Yakup Baykan actively participated in research during his undergraduate studies, publishing four SCI-indexed papers and delivering five conference presentations. He was also selected as a visiting student researcher at King Abdullah University of Science and Technology (KAUST) in Saudi Arabia, gaining international collaborative research experience. As the first international KAIST Presidential Scholar, he plans to pursue a Ph.D. at Stanford University and grow into a leading researcher in space propulsion and combustion.
Awards for outstanding graduates were also presented. Seohyeon Kang (B.S., Brain and Cognitive Sciences) received the Minister of Science and ICT Award (Deputy Prime Minister’s Award). The Chairman of the Board Award was presented to Thai international student Punn Lertjaturaphat (B.S., Industrial Design). The President’s Award went to Kyeongmin Yeo (B.S., School of Computing), while the Alumni Association President’s Award and the KAIST Development Foundation Chairman’s Award were presented to Wonwoo Yoo (B.S., Aerospace Engineering) and Sungbeen Park (B.S., Nuclear and Quantum Engineering), respectively. Hyuk-chae Koo, 1st Vice Minister of Science and ICT, presented the awards on behalf of the Deputy Prime Minister and Minister of Science and ICT.
Seohyeon Kang developed a technology to measure key proteins related to Parkinson’s disease without surgery or tissue damage, opening new possibilities in brain disease research, and was recognized as a model graduate who combined academic excellence with community service. Punn Lertjaturaphat gained recognition at prestigious international conferences such as ACM CHI and co-founded a startup addressing rural elderly care issues, demonstrating creativity in solving social problems through technology and design.
Kyeongmin Yeo published six research papers at leading AI conferences including NeurIPS, ICLR, and CVPR, proposing new theoretical approaches to image generation and demonstrating outstanding academic achievement as a young researcher.
Wonwoo Yoo led the overseas volunteer corps and served as student representative, combining leadership with academic excellence, including winning a grand prize in a rocket launch competition. Sungbeen Park proposed a next-generation beta battery concept, linking it to patents and entrepreneurship, while contributing to public communication and outreach in nuclear technology as student council president and university ambassador.
Commencement addresses were delivered by Dongjae Kang (B.S., Industrial and Systems Engineering) and Gul Osman (Ph.D., Mechanical Engineering), an international student from Türkiye. Kang reflected on how he learned science not merely as an avenue for problem-solving but as a process for exploring the deeper meaning behind phenomena, pledging to remain attentive to unseen challenges faced by others. Osman shared his journey of nurturing his passion for science while working in a factory under difficult economic circumstances, emphasizing that opportunities open to those who persist without giving up. He began his academic journey in Korea through the Korean Government Scholarship Program.
This year, KAIST also spotlighted three notable graduates who forged their own paths encompassing research, the arts, and social value: Seunghyun Ryu, Jeanne Choi, and Daehui Kim (B.S., Civil and Environmental Engineering). Kim led campus environmental organizations and community-based environmental campaigns, earning an Environmental Contribution Award. He plans to pursue a master’s degree focusing on carbon dioxide geological storage research. He also performs as the vocalist of the KAIST metal band “INFINITE,” continuing to balance music and research.
During the ceremony, an Honorary Doctorate in Business Administration was conferred upon uey-Yu Wang, Executive Management Committee Member of Formosa Group and Chairman of Formosa Biomedical Technology Corporation.
President Kwang Hyung Lee encouraged the graduates, saying, “Cherish your dreams, seize opportunities, do not fear failure, and continue to challenge yourselves. I hope you will shine in your own way on your own stage and contribute to society as proud members of the KAIST community.”
2026.02.21 View 2256