%ef%bb%bf%43%6f%6c%6c%65%67%65%20%6f%66%20%45%6e%67%69%6e%65%65%72%69%6e%67
-
KAIST Wins Bid for ‘Physical AI Core Technology Demonstration’ Pilot Project
KAIST (President Kwang Hyung Lee) announced on the 28th of August that, together with Jeonbuk State, Jeonbuk National University, and Sungkyunkwan University, it has jointly won the Ministry of Science and ICT’s pilot project for the “Physical AI Core Technology Proof of Concept (PoC)”, with KAIST serving as the overall research lead. The consortium also plans to participate in a full-scale demonstration project that is expected to reach a total scale of 1 trillion KRW in the future.
In this project, KAIST led the research planning under the theme of “Collaborative Intelligence Physical AI.” Based on this, Jeonbuk National University and Jeonbuk State will carry out joint research and establish a collaborative intelligence physical AI industrial ecosystem within the province. The pilot project will begin on September 1 this year and will run until the end of the year over the next five years. Through this effort, Jeonbuk State aims to be built into a global hub for physical AI.
KAIST will take charge of developing original research technologies, creating a research environment through the establishment of a testbed, and promoting industrial diffusion. Professor Young Jae Jang of the Department of Industrial and Systems Engineering at KAIST, who is the overall project director, has been leading research on collaborative intelligence physical AI since 2016. His “Collaborative Intelligence-Based Smart Manufacturing Innovation Technology” was selected as one of KAIST’s “Top 10 Research Achievements” in 2019.
“Physical AI” refers to cutting-edge artificial intelligence technology that enables physical devices such as robots, autonomous vehicles, and factory automation equipment to perform tasks without human instruction by understanding spatiotemporal concepts.
In particular, collaborative intelligence physical AI is a technology in which numerous robots and automated devices in a factory environment work together to achieve goals. It is attracting attention as a key foundation for realizing “dark factories” in industries such as semiconductors, secondary batteries, and automobile manufacturing.
Unlike existing manufacturing AI, this technology does not necessarily require massive amounts of historical data. Through real-time, simulation-based learning, it can quickly adapt even to manufacturing environments with frequent changes and has been deemed a next-generation technology that overcomes the limitations of data dependency.
Currently, the global AI industry is led by LLMs that simulate linguistic intelligence. However, physical AI must go beyond linguistic intelligence to include spatial intelligence and virtual environment learning, requiring the organic integration of hardware such as robots, sensors, and motors with software. As a manufacturing powerhouse, Korea is well-positioned to build such an ecosystem and seize the opportunity to lead global competition.
In fact, in April 2025, KAIST won first place at INFORMS (Institute for Operations Research and the Management Sciences), the world’s largest industrial engineering society, with its case study on collaborative intelligence physical AI, beating MIT and Amazon. This achievement is recognized as proof of Korea’s global competitiveness in the physical AI technology realm.
Professor Young Jae Jang, KAIST’s overall project director, said, “Winning this large-scale national project is the result of KAIST’s collaborative intelligence physical AI research capabilities accumulated over the past decade being recognized both domestically and internationally. This will be a turning point for establishing Korea’s manufacturing industry as a global leading ‘Physical AI Manufacturing Innovation Model.’”
KAIST President Kwang Hyung Lee emphasized that “KAIST is taking on the role of leading not only academic research but also the practical industrialization of national strategic technologies. Building on this achievement, we will collaborate with Jeonbuk National University and Jeonbuk State to develop Korea into a world-class hub for physical AI innovation.”
Through this project, KAIST, Jeonbuk National University, and Jeonbuk State plan to develop Korea into a global industrial hub for physical AI.
2025.08.29 View 3266 -
KAIST succeeds in controlling complex altered gene networks to restore them to normal
Previously, research on controlling gene networks has been carried out based on a single stimulus-response of cells. More recently, studies have been proposed to precisely analyze complex gene networks to identify control targets. A KAIST research team has succeeded in developing a universal technology that identifies gene control targets in altered cellular gene networks and restores them. This achievement is expected to be widely applied to new anticancer therapies such as cancer reversibility, drug development, precision medicine, and reprogramming for cell therapy.
KAIST (President Kwang Hyung Lee) announced on the 28th of August that Professor Kwang-Hyun Cho’s research team from the Department of Bio and Brain Engineering has developed a technology to systematically identify gene control targets that can restore the altered stimulus-response patterns of cells to normal by using an algebraic approach. The algebraic approach expresses gene networks as mathematical equations and identifies control targets through algebraic computations.
The research team represented the complex interactions among genes within a cell as a "logic circuit diagram" (Boolean network). Based on this, they visualized how a cell responds to external stimuli as a "landscape map" (phenotype landscape).
By applying a mathematical method called the "semi-tensor product,*" they developed a way to quickly and accurately calculate how the overall cellular response would change if a specific gene were controlled.
*Semi-tensor product: a method that calculates all possible gene combinations and control effects in a single algebraic formula
However, because the key genes that determine actual cellular responses number in the thousands, the calculations are extremely complex. To address this, the research team applied a numerical approximation method (Taylor approximation) to simplify the calculations. In simple terms, they transformed a complex problem into a simpler formula while still yielding nearly identical results.
Through this, the team was able to calculate which stable state (attractor) a cell would reach and predict how the cell’s state would change when a particular gene was controlled. As a result, they were able to identify core gene control targets that could restore abnormal cellular responses to states most similar to normal.
Professor Cho’s team applied the developed control technology to various gene networks and verified that it can accurately predict gene control targets that restore altered stimulus-response patterns of cells back to normal.
In particular, by applying it to bladder cancer cell networks, they identified gene control targets capable of restoring altered responses to normal. They also discovered gene control targets in large-scale distorted gene networks during immune cell differentiation that are capable of restoring normal stimulus-response patterns. This enabled them to solve problems that previously required only approximate searches through lengthy computer simulations in a fast and systematic way.
Professor Cho said, “This study is evaluated as a core original technology for the development of the Digital Cell Twin model*, which analyzes and controls the phenotype landscape of gene networks that determine cell fate. In the future, it is expected to be widely applicable across the life sciences and medicine, including new anticancer therapies through cancer reversibility, drug development, precision medicine, and reprogramming for cell therapy.”
*Digital Cell Twin model: a technology that digitally models the complex reactions occurring within cells, enabling virtual simulations of cellular responses instead of actual experiments
KAIST master’s student Insoo Jung, PhD student Corbin Hopper, PhD student Seong-Hoon Jang, and PhD student Hyunsoo Yeo participated in this study. The results were published online on August 22 in Science Advances, an international journal published by the American Association for the Advancement of Science (AAAS).
※ Paper title: “Reverse Control of Biological Networks to Restore Phenotype Landscapes”
※ DOI: https://www.science.org/doi/10.1126/sciadv.adw3995
This research was supported by the Mid-Career Researcher Program and the Basic Research Laboratory Program of the National Research Foundation of Korea, funded by the Ministry of Science and ICT.
2025.08.29 View 3350 -
KAIST–Princeton University Officially Launch “Net-Zero Korea” to Address Climate Crisis
KAIST (President Kwang Hyung Lee) announced on the 27th of August that a research team led by Professor Hae-Won Jeon of the Graduate School of Green Growth and Sustainable Development has signed a memorandum of understanding (MOU) with the Andlinger Center for Energy and the Environment at Princeton University in the United States to promote joint research on carbon neutrality, officially launching the Net-Zero Korea (NZK) project. This project was unveiled at the World Climate Industry EXPO (WCE) held in BEXCO, Busan, and will begin with seed funding from Google.
The NZK project aims, in the short term, to accelerate the transition of Korea’s energy and industrial sectors toward carbon neutrality, and in the mid- to long term, to strengthen Korea’s energy system modeling capabilities for policy formulation and implementation. Energy system modeling plays a critical role in studying the transition to clean energy and carbon neutrality.
In particular, this research plans to apply Princeton’s leading modeling methodologies from the Net-Zero America project—published in 2021 and widely recognized—to the Korean context by integrating them with KAIST’s integrated assessment modeling research.
The Net-Zero Korea project will be supported by funding from Google, KAIST, and Princeton University. This research is characterized by its detailed analysis of a wide range of factors, from regional land-use changes to job creation, and by concretely visualizing the resulting transformations in energy and industrial systems. It will also be conducted through an international collaborative network while reflecting Korea’s specific conditions. In particular, KAIST will develop an optimization-based open-source energy and industrial system model that integrates the effects of international trade, thereby contributing to global academia and policy research.
Therefore, the core of this modeling research is to apply to Korea the precise analysis and realistic approach that drew attention in Net-Zero America. Through this, it will be possible to visualize changes in the energy and industrial systems at high spatial, temporal, sectoral, and technological resolution, and to comprehensively analyze various factors such as regional land-use changes, capital investment requirements, job creation, and health impacts from air pollution. This will provide stakeholders with practical and reliable information.
In addition, the KAIST research team will collaborate with Princeton researchers, who have conducted national-scale decarbonization modeling studies with major research institutions in Australia, Brazil, China, India, Poland, and others, leveraging a global research network for joint studies.
Building on its experience in developing globally recognized integrated assessment models (IAM) tailored to Korea, KAIST will lead a new initiative to integrate international trade impacts into optimization-based open-source energy and industrial system models. This effort seeks to overcome the limitations of existing national energy modeling by reflecting the particularity of Korea, where trade plays a vital role across the economy.
Professor Wei Peng, Princeton’s principal investigator, said: “Through collaboration with KAIST’s world-class experts in integrated assessment modeling, we will be able to build new research that combines the strengths of macro-energy models and integrated assessment models, thereby developing capabilities applicable to many countries where trade plays a crucial role in the economy, such as Korea.”
Antonia Gawel, Director of Partnerships at Google, stated: “We are very pleased to support this meaningful research being conducted by KAIST and Princeton University in Korea. It will greatly help Google achieve our goal of net-zero emissions across our supply chain by 2030.”
Professor Haewon McJeon of KAIST commented: “Through joint research with Princeton University, which has been leading net-zero studies, we expect to provide science-based evidence to support Korea’s achievement of carbon neutrality and sustainable energy.”
President Kwang Hyung Lee of KAIST remarked: “It is deeply meaningful that KAIST, as Korea’s representative research institution, joins hands with Princeton University, a leading institution in the United States, to jointly build a science-based policy support system for responding to the climate crisis. This collaboration will contribute not only to achieving carbon neutrality in Korean society but also to the global response to the climate crisis.”
2025.08.28 View 3693 -
KAIST Develops AI that Automatically Detects Defects in Smart Factory Manufacturing Processes Even When Conditions Change
Recently, defect detection systems using artificial intelligence (AI) sensor data have been installed in smart factory manufacturing sites. However, when the manufacturing process changes due to machine replacement or variations in temperature, pressure, or speed, existing AI models fail to properly understand the new situation and their performance drops sharply. KAIST researchers have developed AI technology that can accurately detect defects even in such situations without retraining, achieving performance improvements up to 9.42%. This achievement is expected to contribute to reducing AI operating costs and expanding applicability in various fields such as smart factories, healthcare devices, and smart cities.
KAIST (President Kwang Hyung Lee) announced on the 26th of August that a research team led by Professor Jae-Gil Lee from the School of Computing has developed a new “time-series domain adaptation” technology that allows existing AI models to be utilized without additional defect labeling, even when manufacturing processes or equipment change.
Time-series domain adaptation technology enables AI models that handle time-varying data (e.g., temperature changes, machine vibrations, power usage, sensor signals) to maintain stable performance without additional training, even when the training environment (domain) and the actual application environment differ.
Professor Lee’s team paid attention to the fact that the core problem of AI models becoming confused by environmental (domain) changes lies not only in differences in data distribution but also in changes in defect occurrence patterns (label distribution) themselves. For example, in semiconductor wafer processes, the ratio of ring-shaped defects and scratch defects may change due to equipment modifications.
The research team developed a method for decomposing new process sensor data into three components—trends, non-trends, and frequencies—to analyze their characteristics individually. Just as humans detect anomalies by combining pitch, vibration patterns, and periodic changes in machine sounds, AI was enabled to analyze data from multiple perspectives.
In other words, the team developed TA4LS (Time-series domain Adaptation for mitigating Label Shifts) technology, which applies a method of automatically correcting predictions by comparing the results predicted by the existing model with the clustering information of the new process data. Through this, predictions biased toward the defect occurrence patterns of the existing process can be precisely adjusted to match the new process.
In particular, this technology is highly practical because it can be easily combined like an additional plug-in module inserted into existing AI systems without requiring separate complex development. That is, regardless of the AI technology currently being used, it can be applied immediately with only simple additional procedures.
In experiments using four benchmark datasets of time-series domain adaptation (i.e., four types of sensor data in which changes had occurred), the research team achieved up to 9.42% improvement in accuracy compared to existing methods.[TT1]
Especially when process changes caused large differences in label distribution (e.g., defect occurrence patterns), the AI demonstrated remarkable performance improvement by autonomously correcting and distinguishing such differences. These results proved that the technology can be used more effectively without defects in environments that produce small batches of various products, one of the main advantages of smart factories.
Professor Jae-Gil Lee, who supervised the research, said, “This technology solves the retraining problem, which has been the biggest obstacle to the introduction of artificial intelligence in manufacturing. Once commercialized, it will greatly contribute to the spread of smart factories by reducing maintenance costs and improving defect detection rates.”
This research was carried out with Jihye Na, a Ph.D. student at KAIST, as the first author, with Youngeun Nam, a Ph.D. student, and Junhyeok Kang, a researcher at LG AI Research, as co-authors. The research results were presented in August 2025 at KDD (the ACM SIGKDD Conference on Knowledge Discovery and Data Mining), the world’s top academic conference in artificial intelligence and data.
※Paper Title: “Mitigating Source Label Dependency in Time-Series Domain Adaptation under Label Shifts”
※DOI: https://doi.org/10.1145/3711896.3737050
This technology was developed as part of the research outcome of the SW Computing Industry Original Technology Development Program’s SW StarLab project (RS-2020-II200862, DB4DL: Development of Highly Available and High-Performance Distributed In-Memory DBMS for Deep Learning), supported by the Ministry of Science and ICT and the Institute for Information & Communications Technology Planning & Evaluation (IITP).
2025.08.26 View 3211 -
KAIST to Host the ‘6th Emerging Materials Symposium’
KAIST (President Kwang Hyung Lee) announced on the 22nd of August that it will host the 6th KAIST Emerging Materials Symposium on the 26th in the Meta Convergence Hall (W13) on its main Daejeon campus, to explore the latest research trends in next-generation promising nanomaterials and discuss future visions.
Launched in 2020, this symposium marks its sixth year and has established itself as KAIST’s flagship academic event by inviting world-renowned scholars on next-generation materials to share groundbreaking achievements.
The event will feature six speakers from four prestigious overseas universities—the Massachusetts Institute of Technology (MIT), Yale University, UCLA, and Drexel University—providing an overview of cutting-edge global research trends in emerging materials, while also showcasing KAIST’s representative achievements.
Notably, Professor Yury Gogotsi of Drexel University, who gained global recognition for the pioneering development of MXene—an emerging material attracting attention for its high electrical conductivity and electromagnetic shielding capability—will deliver a lecture titled “The Future of MXene.”
In the session “Global Frontier in MIT,” three MIT professors will present the institute’s leading research: ▴Professor Ju Li, an authority on AI-robotics-based materials synthesis, ▴Professor Martin Z. Bazant, an expert in the fields of electrochemistry and electronic transport dynamics, and ▴Professor Jeehwan Kim, a leading researcher tackling the limitations of silicon wafer-based semiconductor manufacturing.
In the session “Emerging Materials and New Possibilities,” ▴Professor Yury Gogotsi of Drexel University, ▴Professor Liangbing Hu of Yale University, a pioneer in nanoparticle synthesis through rapid high-temperature thermal processing, and ▴Professor Jun Chen of UCLA, a key researcher in bioelectronic materials using multifunctional flexible materials, will present the development of core emerging materials and future directions.
Additionally, six professors from KAIST’s Department of Materials Science and Engineering will lead the session “KAIST’s MSE Entrepreneurial Spirit” where they will share the process of founding startups based on KAIST’s advanced materials technologies and how nanomaterials have taken root as foundational industries.
The session will include: ▴Professor Il-Doo Kim, founder of the nanofiber and colorimetric gas sensor company IDKLAB; ▴Professor Kibeom Kang, CEO of TDS Innovation, a company specializing in precursors and equipment for 2D material synthesis; ▴Professor Yeonsik Jeong, co-founder of Pico Foundry, a company producing SERS chips; ▴Professor Sang Wook Kim, founder of Materials Creation, which develops products based on high-quality graphene oxide; ▴Professor Jaebeom Jang, founder of Flashomic Inc., a leader in the commercialization of high-speed multiplexed protein imaging technology; and ▴Professor Steve Park, co-CEO of Aldaver, a company developing artificial cadavers (practice organs) that fully replicate the human body. They will each share their entrepreneurial cases, offering vivid lectures on the journey of scientific technologies into the marketplace.
The symposium will also feature a tour of the automated research lab at the Top-Tier KAIST-MIT Future Energy Initiative Research Center, jointly established by KAIST and MIT. The center, designed to build an AI-robotics-based autonomous research laboratory for the rapid development and application of advanced energy materials to help solve the global climate crisis, will operate for ten years. Overseas scholars will also be given an inside look at research and development using automated infrastructure, with discussions to follow on upcoming international collaborations.
Professor Il-Doo Kim of KAIST’s Department of Materials Science and Engineering, who organized the event, emphasized, “This symposium, featuring six global scholars and six KAIST entrepreneurial professors, will be a valuable opportunity to instill an international perspective and entrepreneurial mindset in students. It will also mark a turning point in KAIST’s innovative materials research and international collaborative research network.”
As part of the program, on Wednesday the 27th, KAIST will hold academic exchange sessions with overseas scholars. These will include discussions on international joint research, as well as sessions where KAIST students and early-career researchers can present their work and interact, opening opportunities for future collaborations.
The 6th KAIST Emerging Materials Symposium is open free of charge to all researchers interested in the latest research trends in chemistry, physics, biology, and materials science-related engineering fields.
Participation on the 26th will be available through on-site registration without prior application. Further details are available on the KAIST Department of Materials Science and Engineering EMS website (https://mse.kaist.ac.kr/index.php?mid=MSE_EMS).
2025.08.22 View 2854 -
In KAIST, Robots Now Untie Rubber Bands and Insert Wires Like Humans
The technology that allows robots to handle deformable objects such as wires, clothing, and rubber bands has long been regarded as a key task in the automation of manufacturing and service industries. However, since such deformable objects do not have a fixed shape and their movements are difficult to predict, robots have faced great difficulties in accurately recognizing and manipulating them. KAIST researchers have developed a robot technology that can precisely grasp the state of deformable objects and handle them skillfully, even with incomplete visual information. This achievement is expected to contribute to intelligent automation in various industrial and service fields, including cable and wire assembly, manufacturing that handles soft components, and clothing organization and packaging.
KAIST (President Kwang Hyung Lee) announced on the 21st of August that the research team led by Professor Daehyung Park of the School of Computing developed an artificial intelligence technology called “INR-DOM (Implicit Neural-Representation for Deformable Object Manipulation),” which enables robots to skillfully handle objects whose shape continuously changes like elastic bands and which are visually difficult to distinguish.
Professor Park’s research team developed a technology that allows robots to completely reconstruct the overall shape of a deformable object from partially observed three-dimensional information and to learn manipulation strategies based on it. Additionally, the team introduced a new two-stage learning framework that combines reinforcement learning and contrastive learning so that robots can efficiently learn specific tasks. The trained controller achieved significantly higher task success rates compared to existing technologies in a simulation environment, and in real robot experiments, it demonstrated a high level of manipulation capability, such as untying complicatedly entangled rubber bands, thereby greatly expanding the applicability of robots in handling deformable objects.
Deformable Object Manipulation (DOM) is one of the long-standing challenges in robotics. This is because deformable objects have infinite degrees of freedom, making their movements difficult to predict, and the phenomenon of self-occlusion, in which the object hides parts of itself, makes it difficult for robots to grasp their overall state.
To solve these problems, representation methods of deformable object states and control technologies based on reinforcement learning have been widely studied. However, existing representation methods could not accurately represent continuously deforming surfaces or complex three-dimensional structures of deformable objects, and since state representation and reinforcement learning were separated, there was a limitation in constructing a suitable state representation space needed for object manipulation.
To overcome these limitations, the research team utilized “Implicit Neural Representation.” This technology receives partial three-dimensional information (point cloud*) observed by the robot and reconstructs the overall shape of the object, including unseen parts, as a continuous surface (signed distance function, SDF). This enables robots to imagine and understand the overall shape of the object just like humans.
*Point cloud 3D information: a method of representing the three-dimensional shape of an object as a “set of points” on its surface.
Furthermore, the research team introduced a two-stage learning framework. In the first stage of pre-training, a model is trained to reconstruct the complete shape from incomplete point cloud data, securing a state representation module that is robust to occlusion and capable of well representing the surfaces of stretching objects. In the second stage of fine-tuning, reinforcement learning and contrastive learning are used together to optimize the control policy and state representation module so that the robot can clearly distinguish subtle differences between the current state and the goal state and efficiently find the optimal action required for task execution.
When the INR-DOM technology developed by the research team was mounted on a robot and tested, it showed overwhelmingly higher success rates than the best existing technologies in three complex tasks in a simulation environment: inserting a rubber ring into a groove (sealing), installing an O-ring onto a part (installation), and untying tangled rubber bands (disentanglement). In particular, in the most challenging task, disentanglement, the success rate reached 75%, which was about 49% higher than the best existing technology (ACID, 26%).
The research team also verified that INR-DOM technology is applicable in real environments by combining sample-efficient robotic reinforcement learning with INR-DOM and performing reinforcement learning in a real-world environment.
As a result, in actual environments, the robot performed insertion, installation, and disentanglement tasks with a success rate of over 90%, and in particular, in the visually difficult bidirectional disentanglement task, it achieved a 25% higher success rate compared to existing image-based reinforcement learning methods, proving that robust manipulation is possible despite visual ambiguity.
Minseok Song, a master’s student and first author of this research, stated that “this research has shown the possibility that robots can understand the overall shape of deformable objects even with incomplete information and perform complex manipulation based on that understanding.” He added, “It will greatly contribute to the advancement of robot technology that performs sophisticated tasks in cooperation with humans or in place of humans in various fields such as manufacturing, logistics, and medicine.”
This study, with KAIST School of Computing master’s student Minseok Song as first author, was presented at the top international robotics conference, Robotics: Science and Systems (RSS) 2025, held June 21–25 at USC in Los Angeles.
※ Paper title: “Implicit Neural-Representation Learning for Elastic Deformable-Object Manipulations”
※ DOI: https://www.roboticsproceedings.org/ (to be released), currently https://arxiv.org/abs/2505.00500
This research was supported by the Ministry of Science and ICT through the Institute of Information & Communications Technology Planning & Evaluation (IITP)’s projects “Core Software Technology Development for Complex-Intelligence Autonomous Agents” (RS-2024-00336738; Development of Mission Execution Procedure Generation Technology for Autonomous Agents’ Complex Task Autonomy), “Core Technology Development for Human-Centered Artificial Intelligence” (RS-2022-II220311; Goal-Oriented Reinforcement Learning Technology for Multi-Contact Robot Manipulation of Everyday Objects), “Core Computing Technology” (RS-2024-00509279; Global AI Frontier Lab), as well as support from Samsung Electronics. More details can be found at https://inr-dom.github.io.
2025.08.21 View 3120