-
Secret to Drug Addiction Relapse Found: Brain's Addiction Circuit Identified
<(From Left) Dr. Minju Jeong,(UCSD), Prof. Byung Kook Lim (UCSD), Prof. Se-Bum Paik (KAIST)>
Drug addiction carries an extremely high risk of relapse, as cravings can be reignited by minor stimuli even long after one has stopped using. Previously, this phenomenon was attributed to a decline in the function of the prefrontal cortex (PFC), which regulates impulses. However, a joint international research team has recently revealed that the cause of addiction relapse is not a simple decline in brain function, but rather an imbalance in specific neural circuits.
KAIST announced on March 9th that a research team led by Prof. Se-Bum Paik from the Department of Brain and Cognitive Sciences and Prof. Byung Kook Lim from the University of California, San Diego (UCSD) has identified the core principle by which specific inhibitory neurons in the prefrontal cortex regulate cocaine-seeking behavior.
In particular, the research team focused on parvalbumin-positive (PV) inhibitory neurons, which regulate the balance of neural signals by suppressing the activity of other neurons in the brain. They confirmed that these cells act as a "brake gate" that controls excitatory signals in the brain and serve as a crucial factor in determining drug-seeking behavior that emerges after withdrawal.
The prefrontal cortex (PFC) of our brain can properly perform its "braking" function to suppress impulses when excitatory and inhibitory signals are in balance. To investigate how chronic drug exposure disrupts this balance, the research team conducted cocaine administration experiments on mice. During this process, they tracked when inhibitory neurons in the PFC were activated and how they sent signals to downstream brain regions.
The experimental results showed that parvalbumin (PV) cells, which account for about 60-70% of the inhibitory neurons in the PFC, were highly active when the mice attempted to seek cocaine. However, when "extinction training"—training to stop seeking the drug—was conducted, the activity of these cells significantly decreased. This demonstrates that the activity patterns of PV cells are not permanently fixed by addiction but can be readjusted through the extinction process.
<Figure 1. Experimental design illustrating cocaine self-administration and longitudinal tracking of prefrontal cortical neural activity during cocaine-seeking behavior>
The research team confirmed that artificially suppressing PV cell activity significantly reduced cocaine-seeking behavior in mice. Conversely, activating these cells caused the drug-seeking behavior to persist even after the extinction process. This effect was specifically observed in drug-addiction behavior and did not appear with general rewards like sugar water. Furthermore, this phenomenon was not observed in somatostatin (SOM) cells—another type of inhibitory neuron—indicating that PV cells selectively regulate drug addiction behavior.
<Figure 2. Comparison of single-neuron activity, population activity patterns, and behavioral modulation of prefrontal inhibitory neurons across different stages of cocaine-seeking behavior>
The team also identified the specific brain circuit through which these PV cells operate. Signals originating from the prefrontal cortex are transmitted to the reward circuit of the Ventral Tegmental Area (VTA), a key brain region related to reward. This pathway emerged as the central channel for regulating addiction behavior, determining whether or not to seek the drug again. In this process, PV neurons act as a "regulatory switch," controlling the flow of signals to influence dopamine signaling and deciding whether to maintain or suppress addictive behavior.
In short, the study revealed that addiction relapse is not due to an overall functional decline of the prefrontal cortex, but is determined by whether PV neurons regulate the neural pathway connecting the PFC to the reward circuit.
<Figure 3. Schematic illustrating the prefrontal–reward circuit mechanism that determines drug-seeking behavior>
Prof. Se-Bum Paik stated, "This research shows that drug addiction is a circuit-level problem arising from a collapse in the regulatory balance of specific neurons and downstream neural circuits. The discovery that parvalbumin (PV) cells act as a 'gate' for addictive behavior will provide a crucial lead for developing precision-targeted treatment strategies in the future."
This study was led by Dr. Minju Jeong (UCSD) as the first author, with Prof. Byung Kook Lim (UCSD) and Prof. Se-Bum Paik (KAIST) serving as co-corresponding authors. The findings were published online on February 26 in Neuron, a premier journal in the field of neuroscience.
Paper Title: Distinct Interneuronal Dynamics Selectively Gate Target-Specific Cortical Projections in Drug Seeking
DOI: 10.1016/j.neuron.2026.01.002
Full Author List: Minju Jeong, Seungdae Baek, Qingdi Wang, Li Yao, Eun Ji Lee, Arturo Marroquin Rivera, Joann Jocelynn Lee, Hyeonseok Jang, Dhananjay Bambah-Mukku, Christine Hyun-Seung Mun, Tyler Boesen, Sumit Nanda, Cheol Ryong Ku, Hong-wei Dong, Benoit Labonté, Se-Bum Paik, and Byung Kook Lim.
This research was conducted with the support of the Basic Research Program in Science and Engineering of the National Research Foundation of Korea.
2026.03.10 View 57 -
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 40 -
KAIST Explores Solutions for African Youth Employment with World Bank and African Union
< Group photo of meeting participants >
KAIST announced on the 6th that the 'Jobs for Youth in Africa Knowledge Exchange' platform was held in Nairobi, Kenya, from March 3 to 5 (local time). The event was hosted by the Kenyan government and co-organized by the World Bank Group, the African Union, and the KAIST Global Center for Development and Strategy (G-CODEs).
As a high-level policy implementation platform dedicated to addressing youth employment challenges in Africa, the event drew approximately 200 participants, including government officials from over 20 African nations, international organizations, the private sector, academia, and development cooperation partners. KAIST participated as a key global partner linking technology and policy, presenting innovation models for employment systems based on digital and Artificial Intelligence (AI) technologies.
< Scene from the meeting hosted by the Kenyan government >
With Africa’s youth population projected to double by 2050, the continent faces significant hurdles such as high unemployment rates and informal employment. This event marked the second face-to-face meeting of the 'Jobs for Youth in Africa Community of Practice (CoP),' which was launched in Kigali, Rwanda, in 2025. The meeting aimed to share policy experiences among member states and materialize scalable implementation models. Salim Mvurya, Kenya's Cabinet Secretary for Youth Affairs, Creative Economy, and Sports, attended the opening ceremony and emphasized that youth job creation is a critical priority at both national and continental levels.
The program focused on several key themes:
Evidence-based youth employment strategies
Innovation in employment systems through digital and AI technologies
Improving labor market outcomes through Recognition of Prior Learning (RPL)
Business environment reforms and strengthening value chain linkages
Notably, in the session titled "Digital and AI-based Employment System Innovation," Professor Kyung Ryul Park of KAIST shared Korea’s digital transformation experiences and AI application cases, proposing directions for data-driven policy design and the development of technology-based employment platforms. Additionally, KAIST Professor Ga-young Park facilitated mutual learning and connected cases of scalable youth employment projects across countries during the "Global Cafe Session."
< Professor Kyung Ryul Park of KAIST delivering a presentation >
Participants visited the project site of "National Youth Opportunities Towards Advancement (NYOTA)," an initiative pursued by the Kenyan government and the World Bank. There, they observed a comprehensive youth employment model that integrates vocational training, job matching, and entrepreneurship support. The site visit served as a practical learning opportunity to share the processes of policy design and execution.
Since last year, KAIST has been involved in digital innovation projects for youth employment in East Africa through the Korea-World Bank Partnership Facility (KWPF). Through this event, the university reaffirmed its status as a global cooperation hub leading technology-based policy innovation.
"The issue of youth employment is a structural challenge that combines digital transformation, industrial strategy, and educational reform," stated Professor Kyung Ryul Park. "KAIST will continue to present actionable policy models based on data and technology while strengthening international cooperation."
This Knowledge Exchange platform is evaluated as a significant milestone that reaffirmed the African youth employment agenda as a core priority of international cooperation and solidified the foundation for enhancing policy implementation capabilities. A follow-up workshop is scheduled to be held early next year at the Kenya Advanced Institute of Science and Technology (Kenya-AIST) campus in Konza, Nairobi, which is modeled after KAIST.
2026.03.10 View 45 -
KAIST Develops mRNA Platform That Remains Effective Even in Aging and Obesity
<(From Left) Dr. Subin Yoon, Ph.D candidate Hyeonggon Cho, Prof. Jae-Hwan Nam, Prof. Young-suk Lee>
Since the COVID-19 pandemic, mRNA vaccines have gained attention as a next-generation pharmaceutical technology. mRNA therapeutics work by delivering genetic instructions that enable cells to produce specific proteins for therapeutic effects. However, their efficacy has been reported to decline in elderly individuals or patients with obesity. To address this limitation, Korean researchers have newly designed a key regulatory region of mRNA that improves therapeutic protein production efficiency, developing a next-generation mRNA platform that maintains effectiveness even in aging and obesity conditions.
KAIST (President Kwang Hyung Lee) announced on the 10th of March that a joint research team led by Professor Young-suk Lee of the Department of Bio and Brain Engineering and Professor Jae-Hwan Nam of The Catholic University of Korea (President Jun-Gyu Choi) has developed a new mRNA platform by precisely designing the sequence of the 5′ untranslated region (5′UTR)*, a key regulatory region of mRNA.*5′ untranslated region (5′UTR): A region of mRNA that initiates and regulates protein production. The design of this region influences both the amount and speed of protein synthesis.
The research team analyzed large-scale bioinformatics datasets to identify 5′UTR sequences that enable proteins to be produced more efficiently across diverse cellular environments. When applied, the designed sequences significantly enhanced protein production and immune responses even in preclinical models of aging and obesity.
mRNA is a long single-stranded RNA molecule that serves as the blueprint for producing proteins required by the body. It consists of several components: the 5′UTR, which initiates and regulates the rate of protein production; the coding sequence (CDS), which contains the genetic information for a specific protein; the 3′ untranslated region (3′UTR), which helps maintain mRNA stability within cells; and the poly(A) tail, which further enhances stability and supports protein synthesis.
Among these components, the 5′UTR and 3′UTR do not determine the type of protein produced, but they play a critical role in regulating how efficiently the protein is synthesized. For this reason, these regions are receiving increasing attention as key bioengineering platforms for improving the performance of various mRNA therapeutics, including vaccines and treatments.
<Schematic Diagram of mRNA Therapeutic Design and Validation Using Bioinformatics>
To identify highly efficient 5′UTR sequences capable of promoting protein production across multiple tissues and cellular environments, the team conducted an integrated analysis of large-scale biological datasets. This included multiple analytical approaches such as RNA sequencing (RNA-seq) for analyzing gene activity across tissues, single-cell RNA sequencing (scRNA-seq) for examining gene expression at the individual cell level, and ribosome profiling (Ribo-seq) for measuring actual protein translation efficiency.
The researchers also focused on the fact that in aging or obesity conditions, cells often experience high levels of stress—particularly oxidative stress—which can reduce their ability to synthesize proteins. When the newly designed mRNA therapeutics were applied to preclinical models of aging and obesity, the results showed significantly improved protein production and immune responses compared with existing approaches. This research is expected to be applicable not only to mRNA vaccines but also to a wide range of biopharmaceutical technologies, including gene therapies and immunotherapies.
<Multimodal Bio–Big Data Analysis–Based mRNA Therapeutic Design (AI-Generated Image)>
Professor Young-suk Lee of KAIST Department of Bio and Brain Engineering stated, “This study identified a design strategy that enables mRNA to produce proteins more efficiently by analyzing large-scale biological data,” adding, “This technology will provide an important foundation for ensuring that mRNA vaccines and therapeutics remain effective even in environments where drug efficacy may decline, such as in elderly or obese patients.”
In this study, Dr. Subin Yoon from The Catholic University of Korea and doctoral candidate Hyeonggon Cho from KAIST participated as co-first authors. The research findings were published online on January 2 in the internationally renowned journal Molecular Therapy (IF = 12.0), a leading journal in gene and cell therapy.
(Paper title: ”Designing 5′UTR sequences improves the capacity of mRNA therapeutics in preclinical models of aging and obesity” DOI: https://doi.org/10.1016/j.ymthe.2025.12.060)
This research was supported by the Excellent Young Researcher Program and the Bio-Medical Technology Development Program of the National Research Foundation of Korea funded by the Ministry of Science and ICT, the Infectious Disease Response Innovative Technology Support Program of the Ministry of Food and Drug Safety, and the Infectious Disease Prevention and Therapeutics Technology Development Program of the Korea Health Industry Development Institute.
2026.03.10 View 68 -
KAIST Surpasses the Limits of AlphaFold… AI Now Predicts Whether Drugs Actually Work
<(From Left) Ph.D candidate Hyojin Son, Professor Gwan-su Yi>
Proteins in our body function like switches. When a drug binds to a protein, the structure at the binding site changes, and this structural change propagates throughout the protein, turning its function on or off. Google DeepMind’s AlphaFold3 successfully predicted whether drugs bind to proteins and the three-dimensional structure of binding sites. However, it could not predict how signals propagate inside the protein after drug binding, how the entire structure changes, or whether the protein’s function is ultimately activated or inhibited. KAIST researchers have developed an AI that predicts not only whether a drug binds but whether it actually works.
KAIST (President Kwang Hyung Lee) announced on the 4th of March that a research team led by Professor Gwan-Su Yi of the Department of Bio and Brain Engineering has developed an artificial intelligence model called “GPCRact” that predicts whether candidate molecules not only bind to G-protein-coupled receptors (GPCRs)—a major drug target—but also actually activate the protein.
GPCRs act as “signal receivers” on the surface of cells. When hormones, neurotransmitters, or drugs send signals from outside the cell, GPCRs function as gates that receive these signals and transmit them into the cell. There are about 800 types of GPCRs in the human body, and roughly 30–40% of currently marketed drugs target them. They are key proteins involved in numerous physiological functions, including heart rate regulation, blood pressure control, pain sensing, immune responses, and emotional regulation.
However, a drug binding to a GPCR does not always trigger the desired biological function. Structural changes inside the protein and subsequent signal transmission determine whether the drug actually produces an effect. This process is known as allosteric signal propagation.
The research team designed the AI to learn the drug action process in two stages: ① the drug–target binding stage, and ② the intracellular signal propagation stage within the protein. The three-dimensional protein structure was represented as an atom-level graph, and an attention mechanism was applied to enable the model to learn important signaling pathways. Through this approach, the AI analyzes not only the drug binding signal but also the internal signaling pathways of the protein to predict whether the protein becomes activated.
As a result, the model significantly improved the prediction performance of drug activity even in proteins with complex structures that existing models struggled to analyze. Importantly, the model does not simply output “active” or “inactive.” It also presents the key internal signaling pathways that form the basis of its predictions, overcoming the limitations of so-called “black-box AI.”
<Schematic diagram of drug activity prediction and mechanism interpretation using the GPCRact artificial intelligence model>
This represents an important advance, as it allows researchers to interpret and verify predictions while simultaneously improving the reliability and efficiency of drug discovery. In the future, the model is expected to serve as a precision drug discovery AI platform capable of predicting not only whether drugs bind to GPCRs but also whether they truly activate them in various diseases targeting GPCRs.
<AI-generated image to help illustrate the research>
Professor Gwan-Su Yi explained, “Allosteric structural change refers to a phenomenon in which a drug binds to one part of a protein and its influence propagates internally, altering the function of other regions,” adding, “The key contribution of this research is incorporating this operational principle into deep learning.” He further noted, “We plan to expand the model to various proteins and ultimately develop technologies capable of predicting cellular and whole-body responses.”
Ph.D candidate Hyojin Son participated as the first author in this study. The paper was published on January 15 in the international journal Briefings in Bioinformatics, one of the leading journals in the field of bioinformatics.
※ Paper title: “GPCRact: a hierarchical framework for predicting ligand-induced GPCR activity via allosteric communication modeling”
DOI: https://doi.org/10.1093/bib/bbaf719※ Author information: Hyojin Son (KAIST, first author), Gwan-Su Yi (KAIST, corresponding author)
This research was supported by the Basic Research Program for Individual Research funded by the Ministry of Science and ICT and the National Research Foundation of Korea (RS-2025-24533057).
2026.03.09 View 167 -
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 65 -
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 335 -
Earth’s Safety Limit Already Exceeded… Carbon Emissions More Than Double the Planetary Boundary
<(From Left) Professor Haewon McJeon, Dr. Paul Wolfram>
Earth is not infinite. Pollution beyond certain levels threatens the climate and ecosystems. To prevent this, scientists have proposed “Planetary Boundaries,” defining the safe operating limits of the Earth system. A KAIST research team recalculated climate change and nitrogen pollution using the same standard and found that current carbon emissions already exceed the planet’s sustainable limit by more than double.
KAIST (President Kwang Hyung Lee) announced on the 6th of March that Professor Haewon McJeon of the Graduate School of Green Growth and Sustainability, in collaboration with Dr. Paul Wolfram’s team at the Pacific Northwest National Laboratory (PNNL) of the U.S. Department of Energy, recalculated the carbon dioxide emission boundary using an annual emissions (flow) framework rather than the traditional cumulative carbon stock framework.
Until now, climate change has been evaluated based on how much CO₂ accumulates in the atmosphere (stock). In contrast, nitrogen and phosphorus pollution have been assessed based on how much is emitted each year (flow). Because these problems were measured using different metrics, it was difficult to fairly compare their relative severity. The research team therefore recalculated carbon emissions using the same annual emissions framework used for nitrogen pollution.
Based on the condition of limiting the rise in global average temperature to within 1.5°C, the analysis showed that the Earth’s safe limit for annual CO₂ emissions is approximately 4–17 gigatons (Gt CO₂ per year). However, humanity’s current annual emissions amount to about 37 gigatons (Gt CO₂ per year). This level exceeds the Earth’s safe operating space by more than twofold.
Professor Haewon McJeon stated, “When carbon emissions are compared using the same framework as nitrogen pollution, the severity of climate change becomes much clearer,” adding, “This study helps place different environmental problems on the same analytical basis, which can contribute to setting clearer policy priorities.”
<Comparative Measurement of Planetary Boundaries and Proposal for Flow-Based Carbon Emission Limits>
<Scope and Sensitivity of Flow-Based Carbon Emission Limits>
He further emphasized, “The need for integrated strategies that simultaneously consider carbon, nitrogen, and phosphorus pollution is growing,” adding that global efforts toward decarbonization must accelerate further.
The study was jointly led by Professor Haewon McJeon and Dr. Paul Wolfram as co-corresponding authors, with Hassan Niazi, Page Kyle, and other researchers from PNNL participating as collaborators. The research results were published on February 16 in the international journal Nature Sustainability.
※ Paper title: “Ensuring consistency between biogeochemical planetary boundaries”
DOI: https://doi.org/10.1038/s41893-026-01770-6
This research was supported by the project “Development of an AI-Based Next-Generation Integrated Assessment Model for Climate–Human Interactions” funded by the Ministry of Science and ICT and the National Research Foundation of Korea.
In a Science commentary published on March 5 titled “Thirty-six solutions to stabilize Earth’s climate,” Professor McJeon revisited the progress of climate technologies over the past 20 years. He pointed out that although humanity has possessed many of the necessary technologies, they have not been implemented quickly enough, allowing the climate crisis to intensify. He also emphasized that the pace of decarbonization must accelerate to achieve carbon neutrality.
※ Commentary: “Thirty-six solutions to stabilize Earth’s climate”
Link: https://doi.org/10.1126/science.aed5212
2026.03.06 View 269 -
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 218 -
Simultaneous Decoding of Genetic Maps Inside Cells... A Game Changer for Understanding Complex Human Diseases
< (Clockwise from top left) Professor Inkyung Jung (KAIST), Dr. Dongchan Yang (KAIST), Dr. Kyukwang Kim (KAIST), Dr. Yueyuan Xu (Duke University), Dr. Xiaolin Wei (Duke University), Professor Yarui Diao (Duke University) >
The origin of many diseases begins at the cellular level and involves multiple molecular interactions. However, previous methods have struggled to accurately observe changes in individual cells. Analyzing average values across thousands of cells made it challenging to detect the early signals of disease.
Our university's research team has pioneered groundbreaking technology that decodes the genetic blueprint within a cell in 3D, akin to zooming in on Earth using Google Earth. This innovation is poised to transform research into complex diseases such as cancer, dementia, and Parkinson's disease.
KAIST announced on March 4th that Professor Inkyung Jung's research team from the Department of Biological Sciences, in collaboration with Professor Yarui Diao's team at Duke University, has developed scHiCAR (single-cell Hi-C with assay for transposase-accessible chromatin and RNA sequencing). This is the world’s first ultra-high throughput & precise molecular map decoding technology that simultaneously analyzes gene expression (transcriptome), the epigenome, and the 3D genome structure within a single cell.
The key to determining a cell's state lies in how its genes operate. Genes are not simply switches that turn on and off. The destiny of a cell is determined by which genes are actually active (transcriptome), why they are active (epigenome), and within what spatial structure they operate (3D genome structure). Existing technologies required obtaining this information from different cells separately and then matching them afterward, which could lead to the distortion or omission of subtle changes.
The research team introduced ‘Trimodal Multi-omics’ technology, an integrated precision analysis method that concurrently examines these three types of genetic information within a single cell. By incorporating Artificial Intelligence (AI) analysis, they significantly enhanced accuracy and reproducibility, culminating in a unified platform that reads internal cellular genetic information akin to a ‘single 3D map.’
<Ultra-precision Single-cell Molecular Map>
Notably, the team succeeded in lowering the analysis cost to approximately $0.04 (approx. 50 KRW) per cell. Using this, they constructed a high-resolution molecular map of 1.6 million cells in mouse brain tissue. This means it is now possible to precisely identify when, where, and within what structure disease genes are turned on or off at the cellular level.
The research team applied this technology to brain tissue and the muscle regeneration process, revealing distinct gene operation principles across 22 major cell types. Notably, they successfully tracked in real-time how the 3D structure of genes dynamically changes to influence cell fate during muscle stem cell regeneration. This advancement is expected to lay a crucial foundation for developing treatment strategies for aging and incurable diseases.
<Research Result Image (AI-generated)>
Professor Inkyung Jung remarked, ‘This research transcends mere observation of cells; it opens the door to precisely reading and controlling the genomic blueprints within them. It represents a significant turning point in elucidating the developmental mechanisms of complex diseases like Parkinson's and cancer, as well as identifying target points for patient-specific new drugs.’
The study was published on February 19th in the international academic journal Nature Biotechnology (IF=46.9).
Paper Title: Trimodal single-cell profiling of transcriptome, epigenome and 3D genome in complex tissues with scHiCAR
DOI: 10.1038/s41587-026-03013-7
Meanwhile, this research was conducted with support from the Suh Kyungbae Foundation, the Samsung Science and Technology Foundation, and the Basic Research Program and Bio-Medical Technology Development Program of the National Research Foundation of Korea (Ministry of Science and ICT).
2026.03.06 View 174 -
Campus in Spring: KAIST Art Museum Hosts ‘Oblique Time’ Exhibition
< (Left: Stair-shaped work) Divided Horizon, (Right: Circular mirror work) In-between >
KAIST invites visitors to go beyond simply 'looking' at art, offering a space to experience ‘Oblique Time’ while strolling through the venue. ‘Oblique Time’ refers to a different texture of time—stepping away from the linear flow of daily life to a place where senses and contemplation slowly intersect.
KAIST announced that it will host ‘Oblique Time,’ a special installation art exhibition by graphic designer Na Kim, at the KAIST Art Museum on the main Daejeon campus starting on the 3rd.
The exhibition takes place on the newly opened rooftop of the museum. The three installed works awaken a new sense of space through the media of wind, light, gaze, and movement. Upon entering the rooftop, soaring columns greet visitors; as one walks slowly between them, the traces of the wind brushing past can be felt by the body.
Ascending and descending the stairs reveals entirely different scenes within the same space, while circular mirrors on the floor simultaneously reflect the sky, the clouds, and the viewer standing before them. Visitors shift from being mere observers to becoming part of the landscape within the artwork. The space is no longer a fixed structure but transforms into a fluid field of time.
All three works are experiential installations that rely on the participation and movement of the audience. Rather than a fixed viewpoint, the artist uses a "wavering gaze" to twist our conventional senses of space and time. It is an exhibition that leaves behind experience over explanation, and contemplation over definitive answers.
Na Kim is a graphic designer and visual artist who graduated from the Department of Industrial Design at KAIST and studied at Hongik University and ArtEZ University of the Arts in the Netherlands. Based in design, she has built a unique artistic world that crosses into the realm of fine arts. By extracting and reconstructing memories and contexts tied to objects, images, and text, she encourages viewers to reflect on themselves and conjure their own self-portraits. She has received the Korea Institute of Design Promotion’s ‘Next Generation Design Leader’ award, the Doosan Yonkang Art Award, and the Ministry of Culture, Sports and Tourism’s ‘Today’s Young Artist Award.’ Currently represented by Kukje Gallery, she operates the project space ‘LOOM’ in Berlin, Germany.
"The creation of an art museum within the KAIST campus felt very meaningful to me," said Kim. "I am honored to participate in the first exhibition for this newly opened rooftop space. I hope this exhibition provides students with an opportunity to experience art and reflect on themselves."
Since its opening in December 2024, the KAIST Art Museum has operated three exhibition halls on the first and second floors. With the recent completion of interior work on the third floor, it has evolved into a full-scale museum with a total of seven galleries (five indoor, two outdoor). This exhibition, held in the rooftop area (Galleries 6 & 7) being revealed for the first time since the museum's opening, symbolizes the spatial expansion of the institution. This stage—extending from indoors to outdoors and from fixed galleries to the open sky—marks the starting point of a new experiment envisioned by the museum.
Hyeon-Jeong Suk, Director of the KAIST Art Museum (and Head of the Department of Industrial Design), stated, "I am delighted to introduce Na Kim, who is also a junior from our department, to KAIST. Her original artistic world, rooted in the language of design, will diversely expand the museum's exhibitions."
President Kwang Hyung Lee remarked, "I am proud that a KAIST graduate has grown into a world-class artist and returned to her alma mater. I look forward to this exhibition serving as a leap forward for both the KAIST Art Museum and the artist."
Starting with the museum construction fund and art donations from the late Chairman Moon-Soul Chung, KAIST has consistently received artworks from various social figures, artists, and bereaved families. Na Kim’s installation works will also belong to the KAIST Art Museum after the exhibition concludes, remaining as artistic assets for the campus.
The exhibition ‘Oblique Time’ opens at 3:00 PM on the 3rd and will be open to the public free of charge until August 28th. Viewing hours are weekdays from 10:00 AM to 5:00 PM.
< Vertical Texture (5m size)>
< Divided Horizon (8m size) >
< In-between (Circular diameters of 1.2m and 1.5m) >
2026.03.04 View 172 -
Designing the Heart of Hydrogen Cars with AI... Development of Next-Generation Super Catalyst
<(From left) KAIST Ph.D. Candidate HyunWoo Chang, Professor EunAe Cho. (Top, from left) Seoul National University Professor Won Bo Lee, Dr. Jae Hyun Ryu.>
In the era of climate crisis, hydrogen vehicles are emerging as an alternative for eco-friendly mobility. However, the fuel cell, known as the ‘heart of the hydrogen car,’ still faces limitations of high cost and short lifespan. The core cause is the platinum catalyst. While it is a decisive material for generating electricity, the reaction is slow, performance degrades over time, and manufacturing costs are high. Korean researchers have presented a clue to solving this difficult problem.
KAIST announced on February 26th that the research team led by Professor EunAe Cho of the Department of Materials Science and Engineering, together with the team of Professor Won Bo Lee of the School of Chemical and Biological Engineering at Seoul National University, has developed a technology that predicts the ‘atomic arrangement’ tendency of catalysts using artificial intelligence (AI).
This technology is akin to calculating beforehand which combination is advantageous for completing a puzzle before putting it together. By having AI calculate the arrangement speed of metal atoms first, it has become possible to efficiently design catalysts with better performance. The core of this research is that ‘AI revealed the fact that zinc plays a decisive role in the platinum-cobalt atomic arrangement.’
<Schematic diagram of AI-based atomic alignment prediction>
Despite the high performance of existing platinum-cobalt (Pt-Co) alloy catalysts, very high-temperature heat treatment was required to create the ‘intermetallic (L1₀)’ structure, where atoms are regularly arranged. In this process, particles would clump together, or the structure would become unstable, posing limitations for actual fuel cell application.
To solve this problem, the research team introduced machine learning-based quantum chemistry simulations. Through AI, they precisely predicted how atoms move and arrange themselves inside the catalyst.
As a result, they discovered that zinc (Zn) acts as a mediating element that promotes atomic arrangement. The principle is that when zinc is introduced, atoms find their places more easily, forming a more sophisticated and stable structure. In other words, AI has found the ‘optimal path for atomic arrangement creation’ in advance.
< Synthesis process of Zinc-introduced Platinum-Cobalt catalyst>
The zinc-platinum-cobalt catalyst, synthesized based on AI predictions, secured both higher activity and superior long-term durability compared to commercial platinum catalysts. This is a case proving that the ‘virtual blueprint’ calculated by artificial intelligence can be implemented as a high-performance catalyst in an actual laboratory.
In particular, this technology is expected to contribute to extending catalyst lifespan and reducing manufacturing costs across core carbon-neutral industries, such as hydrogen passenger cars, hydrogen trucks requiring long-distance operation, hydrogen ships, and energy storage systems (ESS).
< Conceptual diagram of AI-based catalyst development (AI-generated image) >
Professor EunAe Cho stated, “This research is a case of utilizing machine learning to predict the atomic arrangement tendency of catalysts in advance and implementing this through actual synthesis,” and added, “AI-based material design will become a new paradigm for the development of next-generation fuel cell catalysts.”
Ph.D. Candidate HyunWoo Chang from KAIST’s Department of Materials Science and Engineering and Dr. Jae Hyun Ryu from Seoul National University’s School of Chemical and Biological Engineering participated as co-first authors in this research. The research results were published on January 15, 2026, in ‘Advanced Energy Materials,’ a world-renowned academic journal in the energy materials field. ※ Paper Title: Machine Learning-Guided Design of L1₀-PtCo Intermetallic Catalysts: Zn-Mediated Atomic Ordering, DOI: https://doi.org/10.1002/aenm.202505211
This research was conducted with the support of the National Research Foundation of Korea’s Nano & Material Technology Development Program and the Korea Institute of Energy Technology Evaluation and Planning’s Energy Innovation Research Center for Fuel Cell Technology.
2026.02.27 View 426