연구 > 연구성과 - KAIST 전기 및 전자공학부

Prof. David Hyunchul Shim’s Team Successfully Demonstrates Naval Vessel Steering with physical AI Humanoid Robot ‘PIBOT’

A research team led by Professor David Hyunchul Shim from the School of Electrical Engineering at KAIST has successfully conducted a field demonstration of its humanoid robot, ‘PIBOT,’ operating as a naval helmsman in collaboration with the Republic of Korea (ROK) Navy, proving the applicability of AI and robotics technologies in the defense sector.

This demonstration marks a significant milestone as the military’s first ‘robot crew experiment on a real naval environment,’ showing that a humanoid robot can directly replace the physical tasks of human crew members without any modifications or upgrades to existing vessel systems.

The research team integrated Large Language Model (LLM) technology—capable of understanding context and generating appropriate responses—into PIBOT. This design enables the robot to autonomously recognize and execute complex ship maneuvering procedures and verbal commands. Also, the robot controls the ship to follow the verbally commanded heading as it monitors the ship’s motion in real time.

On July 23, during the Stage 1 land-based maneuvering test at the ROK Navy Education and Training Command’s maneuvering simulator, PIBOT received voice commands from the bridge officer on watch such as “Rudder left 5 degrees, steady on 330 degrees!” PIBOT accurately recognized and repeated the command before manually operating the helm. Once the vessel reached the target heading, PIBOT successfully delivered a completion report, stating, “Steady on 330 degrees completed!”

The one-hour test simulated various complex and harsh navigation conditions, including navigating narrow waterways with heavy ship traffic, harsh offshore weather, and nighttime navigation. The research team plans to thoroughly analyze response times from command input to execution, steering precision, and operator convenience data to further enhance PIBOT’s control algorithms.

Building on the success of this Stage 1 land simulation, Professor Shim’s team plans to gradually expand the scope of verification to tests on ground test vessel (Stage 2), daytime navigation at sea (Stage 3), and long-term day and night navigation at sea (Stage 4).

“This research serves as a crucial stage to demonstrate how physical AI humanoid robots developed in our school can perform high-stake military missions amid the complex global situations environments and also severely shrinking military manpower in our country,” said Professor Shim. “Starting with ship maneuvering, we will expand the capability of the helmsman robot robots to cover broader functionalities.”

The development of PIBOT by Professor Shim’s research team is part of the Future Challenge Defense Technology R&D Project supported by the Agency for Defense Development (ADD), funded with 5.7 billion KRW from the ROK Defense Acquisition Program Administration (DAPA) since 2022. It is also jointly carried by our School’s Prof. Min Jun Kim’s team.

Four Researchers from Prof. Insu Yun’s Lab Named Microsoft MSRC 2026 ‘Most Valuable Researchers’

< (From left) Haein Lee (Ph.D. Student), Prof. Insu Yun, Minwoo Jeong (Master’s Student), Gyeongmin Kim (Master’s Student), (Top right inset) Seung Chan Kim (Undergraduate Student)>

Four students from Professor Insu Yun’s research group in the School of Electrical Engineering have been named to Microsoft’s Security Response Center (MSRC) ‘2026 Most Valuable Researcher (MVR)’ list.

The MSRC MVR program is a prestigious initiative that recognizes the top 100 security researchers worldwide who have made outstanding contributions to reporting and disclosing vulnerabilities in Microsoft products over the past year (July 2025 – June 2026).

The featured members from Professor Yun’s lab are:

Seung Chan Kim (Soc Undergraduate Student) – Ranked 13th
Haein Lee (Ph.D. Student) – Ranked 25th
Gyeongmin Kim (Master’s Student) – Ranked 61st
Minwoo Jeong (Master’s Student) – Ranked 64th

This recognition further solidifies the lab’s world-class standing in software security and vulnerability research. Last year, the team won first place in the DARPA AI Cyber Challenge (AIxCC) in the US and has recently secured a spot in the finals for DEF CON CTF 2026, often dubbed the “Olympics of Cyber Security.”

“I would like to extend my warmest congratulations to the students for their incredible dedication and outstanding achievements,” said Professor Insu Yun. “We will continue to push the boundaries of cybersecurity and lead global technology advancements.”

MSRC 2026 Leaderboard: https://msrc.microsoft.com/leaderboard

Prof. Ian Oakley’s Team Develops ‘ComiXR’, an Extended Reality (XR) Platform for Reading and Creating Comics

< (From left) Ammar Al-Taie, a postdoctoral researcher at the KAIST Information and Electronics Research Institute; Professor Ian Oakley and doctoral student Hyunyoung Han. >

Webtoons are coming to life in the physical world, ushering in a new era in which comics are not merely viewed, but experienced.

A KAIST EE research team has developed the world’s first next-generation extended reality (XR) comics platform that enables a wide range of readers to enjoy immersive, three-dimensional comics in physical space. By expanding webtoons beyond the screen and into the real world, the team has opened up new possibilities for the future of comics.

A research team led by Professor Ian Oakley from the School of Electrical Engineering has proposed core design principles and future directions for next-generation extended reality (XR) comics through a systematic user study involving 15 participants, including human-computer interaction (HCI) experts, professional webtoon creators, and readers.

< Figure 1: A comic summarizing our study: Comics evolve to meet reader needs, and the digital age has further changed their layouts. To address the next stage in their evolution, we investigated how comics may be laid out in eXtended Reality (XR). >

The research team developed ComiXR, a new platform that enables users to both read and create comics in XR environments. Participants used the platform to transform a conventional print comic into an XR comic and explored how different visual, auditory, haptic, and interactive features could be combined.

Comics, which originated in printed books and newspapers, have evolved dramatically with the rise of smartphones. The vertical-scrolling format of webtoons has become particularly successful by adapting comics to the interaction methods of mobile devices.

The research team viewed XR devices as a potential next stage in this evolution. To explore how spatial depth, three-dimensional rendering, spatial audio, eye tracking, and facial expression tracking could be incorporated into comics, the team built ComiXR using a Meta Quest Pro headset.

< Figure 2: Screenshots of participants using ComiXR. Left: ComiXR’s intitial state; passthrough with menus to spawn XR features. Middle: a participant testing eye-tracking to highlight a character. Right: a participant triggering a speech bubble through eye-tracking by looking at the character. >

While wearing the headset, participants freely positioned 3D characters, speech bubbles, sound effects, and other comic elements throughout a physical room. They were able to construct comic environments that they found comfortable, engaging, and immersive.

The results showed that readers strongly preferred designs that actively used the depth of physical space over simply displaying flat comic pages in a virtual environment. Immersion increased significantly when characters were positioned at a different depth from the background and speech bubbles were separated into distinct layers. In particular, an eye-tracking feature that revealed the next line of dialogue only when the reader looked at a specific character proved effective in preventing spoilers.

The platform also demonstrated new sensory experiences that are not possible in conventional comics. Special effects could be triggered in response to readers’ facial expressions, while haptic feedback could convey sensations such as a character’s heartbeat or the impact represented by an onomatopoeic effect.

< Figure 3. The study setup. First, a 15-minute introduction phase with participants using ComiXR to experience different XR comic features. Second, a 45-minute transformation phase, in which participants adapted a print comic (top left) into XR (top right). Finally, participants answered a post-study survey indicating the acceptability of XR comics. >

Based on the study, the research team also proposed four key design concepts for XR comics. The first, “The Panel Gallery,” transforms the walls of a room into a gallery for displaying comic panels. The second, “The Pop-Up,” presents comics like pop-up books on desks or walls. The third, “Around Comic,” places 3D characters and other comic elements in outdoor spaces. The fourth, “Inclusive ComiX,” improves accessibility for a wide range of readers.

The research team expects XR comics to complement, rather than replace, existing smartphone-based webtoons. They could be used for special exhibitions and educational content that allow audiences to experience fictional worlds more vividly, as well as platforms that improve access to cultural content for a wider range of users.

Ammar Al-Taie, a postdoctoral researcher at the KAIST Information and Electronics Research Institute, participated as the first author, while Hyunyoung Han, a doctoral student in the School of Electrical Engineering, participated as a co-author.

The research was presented at the ACM Designing Interactive Systems Conference 2026, or ACM DIS 2026, one of the leading international conferences in human-computer interaction and design. The ComiXR platform has also been released as open-source software for public use.

Paper title : ComiXR: Exploring Comic Layouts in eXtended Reality
DO I: https://doi.org/10.1145/3800645.3812857
Related Video : https://drive.google.com/drive/folders/1D9Efp3T0biDbUm1K5Gu6HLSSy89Uaq8R?usp=sharing
Open-source platform : https://github.com/ammarjamal/ComiXR

The research was supported by the KAIST Jang Young Sil Fel¬lowship Program (Excellence Track). The authors acknowledge support from the IITP (Institute of Information & Communications Technology Planning & Evaluation)-ITRC (Information Technology Research Center) grant funded by the Korean government (Ministry of Science and ICT) (IITP-2026-RS-2024-00436398).

Professor Changick Kim’s Team Develops Key Technology to Make Personalized AI Safer

< (From left) Ph.D. candidate Wonjun Lee, Professor Changick Kim, Ph.D. candidate Seokil Ham, and Ph.D. candidate Jaehyuk Jang. >

“Create an AI assistant trained only on our company’s documents.”

The era of building “personalized AI” by training AI models on individual or corporate documents and data is beginning. However, while such customization can improve task performance, it can also weaken the model’s existing safety safeguards. KAIST researchers have developed a core AI technology that preserves customized performance while further strengthening safety.

A research team led by Professor Changick Kim from its School of Electrical Engineering has developed “Buffer-and-Reinforce,” a training framework for safe fine-tuning that prevents safety degradation when large language models (LLMs), such as ChatGPT, are retrained on data from individuals or companies to better suit their needs.

Until now, one of the biggest challenges in the era of personalized AI has been that fine-tuning improves a model’s ability to perform new tasks, but can also weaken its existing safety rules. The research team focused on prior findings showing that, counterintuitively, fine-tuning an AI model while it is in a temporarily jailbroken state — a state in which it may respond even to dangerous requests it would normally refuse — does not significantly compromise its safety.

The team then devised a new approach in which this jailbroken state is not used in actual services, but is applied only temporarily during the fine-tuning process through a buffering module called “BufferLoRA,” which is removed after training.

The research team was the first to clarify why this phenomenon occurs. They found that, in the temporarily jailbroken state, the AI model becomes less easily influenced by harmful information, while still effectively learning the new task abilities desired by the user. In other words, the model can continue learning useful knowledge without additionally absorbing harmful behaviors.

Based on this insight, the team developed a two-stage learning method consisting of “buffering” and “safety reinforcement.”

First, the temporary buffering module, BufferLoRA, is applied to the AI model during user fine-tuning, where it acts as a protective layer that prevents harmful data from directly affecting the base model. Once fine-tuning is complete, this module is removed.

Next, a safety reinforcement module called “ReinforceLoRA” is applied to restore and strengthen the model’s safety. In this process, the team used QR decomposition, a mathematical technique that separates different types of information and selectively reflects only the necessary components. This allowed the model to retain the new functions learned from user data while selectively reinforcing safety.

Simply put, the researchers first placed a temporary protective layer, BufferLoRA, over the AI model so that harmful data could not directly affect it, while allowing the model to learn the necessary task. They then removed the protective layer and applied ReinforceLoRA to strengthen the model’s safety safeguards. As a result, the model maintained its customized performance while achieving even stronger safety.

< Figure 1. Infographic of the Buffer-and-Reinforce training framework and its applications. >

In experiments, the AI model maintained high safety even in an extreme setting where all user data consisted of harmful questions and answers. After fine-tuning, the rate at which the AI generated harmful responses was about 8%, lower than the roughly 18% observed in the original model that had not been fine-tuned at all. The framework also achieved strong customized performance and state-of-the-art safety without requiring additional safety data during user fine-tuning or significantly increasing computational cost, suggesting its practical applicability to real-world personalized AI services.

Professor Changick Kim stated, “This research provides a key foundational technology that allows anyone to build customized AI with their own data while using it more safely,” adding, “We expect it to contribute significantly to building a trustworthy AI service environment in the era of personalized AI and AI agents.”

This research was led by Seokil Ham, a doctoral student in KAIST’s School of Electrical Engineering, as first author. The paper was selected as a Spotlight presentation at the International Conference on Machine Learning (ICML) 2026, one of the world’s most prestigious conferences in artificial intelligence, an honor given to only about the top 2.2% of all submitted papers, drawing international attention.

Paper title: Jailbreak to Protect: Buffering and Reinforcing via Temporary Jailbreaking for Safe Fine-Tuning in Large Language Models
DOI: 10.48550/arXiv.2605.24550
Author information: Seokil Ham (KAIST, first author), Jaehyuk Jang (KAIST, second author), Wonjun Lee (KAIST, third author), Changick Kim (KAIST, corresponding author)
Related video: https://drive.google.com/file/d/1gfok06dE8699qtiUR7gVsRoVmBGADaWQ/view?usp=sharing

This work was supported by Institute of Information & Communication Technology Planning & Evaluation (IITP) grant funded by Ministry of Science and ICT(MSIT) (No. RS-2025-02215344, Development of AI Technology with Robust and Flexible Resilience Against Risk Factors).

Prof. Jimin Kwon’s Team Develops Optical Image-Based Automation for 2D Semiconductor Screening and Device Fabrication

< (From left) Professor Jimin Kwon, Dr. Haksoon Jung, and Dr. Yongwoo Lee >

The era of researchers manually searching for two-dimensional semiconductors, which are drawing attention as next-generation AI semiconductors, is coming to an end. KAIST researchers have automated semiconductor screening and device fabrication, analyzed thousands of devices, and revealed the relationship between thickness and performance that had long been difficult to identify. This achievement is expected to shift next-generation semiconductor research toward a data-driven approach and accelerate the commercialization of AI semiconductors and ultra-low-power semiconductors.

The research team led by Professor Jimin Kwon of the School of Electrical Engineering and the Department of AI System has developed a technology that automatically identifies two-dimensional semiconductors from optical microscope images alone and connects the process to transistor fabrication, through joint research with UNIST, Hanbat National University, Hanyang University, and Washington University in St. Louis in the United States.

< Figure1. Optical-image-based framework for MoS₂ flake identification and device fabrication >

Two-dimensional semiconductors are ultrathin semiconductors only a few atomic layers thick. They are called “dream semiconductors” because they can enable smaller semiconductors that consume less electricity than conventional silicon semiconductors. Today’s silicon semiconductors are approaching physical limits, as continued miniaturization of circuits leads to greater power loss and heat generation. Two-dimensional semiconductors, which are attracting attention as next-generation materials to overcome these limits, are expected to be used in a wide range of future technologies, including AI semiconductors, smartphones, data centers, wearable devices, foldable or stretchable electronics, and ultra-small medical sensors.

However, in two-dimensional semiconductors made through solution processing, the position, size, and thickness of each small semiconductor flake all differ, requiring researchers to find the desired samples one by one under a microscope. They then had to manually design electrodes according to the identified positions, requiring substantial time and effort, and making it practically difficult to analyze thousands or more devices at once.

The research team used molybdenum disulfide (MoS₂), a representative two-dimensional semiconductor material. By using the fact that the RGB red, green, and blue brightness values seen under a microscope change depending on thickness, the team enabled a computer to automatically identify the desired semiconductor and automatically design the electrodes. Verification using atomic force microscopy (AFM) confirmed that even subtle thickness differences of three to eight layers could be accurately distinguished.

Through this approach, the team successfully selected suitable samples automatically from more than 120,000 semiconductor flakes and fabricated and analyzed 1,615 transistors.

The large-scale analysis also produced meaningful results. The team statistically clarified for the first time that as the semiconductor becomes thicker, current flows more easily, but the ability to switch electricity on and off actually decreases. This characteristic had been difficult to confirm previously because only a small number of samples could be analyzed, but the team revealed it through large-scale data.

< Figure 2. Verification of MoS₂ layer number based on the selection factor (S) and RG-channel brightness clustering >

The greatest significance of this study is that it did not simply automate the fabrication process, but transformed two-dimensional semiconductor research, which had relied on human experience, into data-driven research. Going forward, the technology is expected to enable researchers to fabricate and analyze more semiconductors more quickly, identify high-performance materials, and ultimately expand into research in which AI designs new semiconductors.

This study was conducted with Professor Jimin Kwon, Dr. Haksoon Jung, and Dr. Yongwoo Lee of KAIST as co-corresponding authors, and Sanghyun Lee of UNIST as the first author. The research results were published on April 3 in Advanced Functional Materials, a leading international journal in materials science, and were also selected as an Inside Back Cover article in the field of 2D Materials & Electronics.

Paper title: Statistically Resolving Thickness-Dependent Electrical Characteristics in Multilayer-MoS₂ Transistors
DOI: 10.1002/adfm.202532204
Author information: Professor Jimin Kwon (KAIST, corresponding author), Dr. Haksoon Jung (KAIST, corresponding author), Dr. Yongwoo Lee (KAIST, corresponding author), Sanghyun Lee (UNIST, first author), and participating researchers from partner institutions: Sumin Hong (UNIST), Minho Park (UNIST), Professor Seongju Kim (Hanbat National University), Professor Sang-Hoon Baek (Hanyang University), Professor Joonki Suh (KAIST), Seonguk Yang (KAIST), Professor Sang-Hoon Bae (Washington University in St. Louis), and Dr. Chang-Soo Lee (TDS)

This research was supported by the Individual Basic Research Program of the National Research Foundation of Korea (NRF), funded by the Ministry of Science and ICT (MSIT), and by the Advanced Strategic Industry Super-Gap Technology Development Program of the Korea Planning & Evaluation Institute of Industrial Technology (KEIT), funded by the Ministry of Trade, Industry and Energy (MOTIE).

Prof. Seunghyup Yoo’s Team Implements Distortion-Free Stretchable Display Platforms

< (From left) Professor Hanul Moon of Dong-A University, Dr. Junho Kim, Professor Seunghyup Yoo, and Dr. Su-Bon Kim of KAIST EE >

Beyond bendable and foldable displays, the era of stretchable displays, whose screens can expand freely like rubber, is now emerging. KAIST researchers have developed a core technology that allows text, images, and other on-screen information to retain their original shape even when the screen is stretched by up to 15%. The achievement is expected to help solve the problem of image distortion and accelerate the commercialization of next-generation high-quality stretchable displays.

The research team led by Professor Seunghyup Yoo of the School of Electrical Engineering, in collaboration with Professor Hanul Moon’s team at Dong-A, has successfully implemented an auxetic-based stretchable display platform. Auxetic structures expand in both width and length when pulled, allowing the display to stretch uniformly at the same ratio in all directions without distorting the image on the screen.

Conventional stretchable displays are generally made by forming light-emitting devices on a stretchable substrate, which serves as the base layer of the display. However, when such a substrate is stretched in one direction, it tends to shrink in the opposite direction, causing letters and images on the screen to become flattened or distorted. Auxetic structures have been used to address this problem, but most previous approaches were limited to maintaining the overall horizontal-to-vertical ratio of the screen, while the letters and images within the screen still remained vulnerable to distortion.

Instead of bonding the auxetic structure and the stretchable substrate across the entire surface, as in conventional methods, the research team proposed a new design approach that uses computational analysis to selectively connect only the necessary points that ensure isotropic expansion throughout the substrate.

In the conventional approach, the twisting deformation that occurs as the auxetic structure stretches is directly transferred to the substrate, distorting the image inside the screen. In contrast, the platform developed by the research team was designed so that each region moves evenly outward from its original position. This allows not only the entire screen but also small areas such as letters and images to expand together while maintaining their original shapes.

< Figure1. Image distortion limitations of conventional stretchable displays (Upper row) and the auxetic-based stretchable display design proposed in this study, including its selective bonding strategy. >

The research team verified the platform’s performance by repeatedly stretching a substrate patterned with letters and images in both the horizontal and vertical directions. In the conventional method, the patterns underwent local deformation, whereas in the new platform, the shapes of the letters and images remained intact. This demonstrates that not only the whole screen but also fine images on-screen can expand uniformly without distortion.

The team also integrated an LED array, a structure in which multiple LEDs are arranged at regular intervals, onto the platform to verify its performance as a working stretchable display. Even when stretched by up to 15% in both the horizontal and vertical directions, stable electrical operation and the screen brightness were maintained. After repeated stretching to 15%, the decrease in brightness remained below 2%, confirming the platform’s potential for practical display applications.

< Figure2. Demonstration of distortion-free characteristics of the proposed auxetic-based stretchable display (right) in comparison to those of conventional, fully-bonded auxetic-based stretchable displays (left) >

This technology is expected to serve as a core platform for next-generation electronics with freely changeable shapes, including wearable electronic devices, electronic skin, or e-skin, which refers to electronic devices that stretch like skin while sensing and displaying information, medical biosensors, soft robots, and curved displays for automobiles and aircraft.

Professor Seunghyup Yoo of KAIST said, “For stretchable displays to be used as actual information display devices, they must not only stretch well, but also preserve on-screen information accurately during stretching,” adding, “This platform enables uniform expansion from small areas of the screen to the entire display, and will serve as a key foundational technology for accelerating the commercialization of high-quality stretchable displays.”

This study was led by KAIST Dr. Su-Bon Kim and Dr. Junho Kim as co-first authors, with Professor Hanul Moon of Dong-A University and Professor Seunghyup Yoo of KAIST as co-corresponding authors. The research was published in the international journal Nature Communications on June 10.

Paper title: Hybrid auxetic metamaterial platforms enabling multiscale isotropic expansion for distortion-free stretchable displays
DOI: 10.1038/s41467-026-74141-6
Demonstration Video: https://bit.ly/4gSRf8W

This research was supported by the National Research Foundation of Korea (NRF) Mid-Career Researcher Program, the Future Display Strategic Research Laboratory Program, the Korea Planning & Evaluation Institute of Industrial Technology (KEIT), and theKorea Institute for Advancement of Technology (KIAT) HRD Program.

Professor Minsoo Rhu’s Team First Reveal the “Hidden Power Cost” of AI Agents

<(From left: Byungjun Shin, Master’s student; Jinha Jung, MS-PhD integrated student; Jiin Kim, PhD student; and Professor Minsoo Rhu>

Artificial intelligence (AI) is rapidly evolving beyond simply answering questions into the era of “AI agents,” which can perform complex tasks by autonomously carrying out multi-step reasoning and using external tools. However, the extent to which these advances require additional resources and power has not been properly quantified until now.

A research team led by KAIST School of Electrical Engineering Chair Professor Minsoo Rhu has announced the first quantitative analysis of AI agents’ computational cost, response latency, energy consumption, and their broader impact on data center power demand. The findings were presented at IEEE HPCA 2026, one of the most prestigious conferences in the field of computer architecture, drawing significant attention from the academic community.

The research team focused on the fact that AI agents do not merely increase computational workload, but also impose a fundamentally new burden on data center infrastructure. Unlike conventional “chain-of-thought” reasoning, which proceeds step by step in a manner similar to human reasoning, AI agents operate by repeatedly invoking large language models (LLMs) throughout the task execution process.

The analysis found that AI agents triggered, on average, 9.2 times more LLM invocations than conventional approaches, while processing latency increased by as much as 153.7 times. In addition, during tool-use phases, GPUs inevitably remained idle, with idle time accounting for up to 54.5% of the total execution time. This means that as tasks become more complex, the inefficiency of underutilizing expensive GPUs becomes increasingly severe.

When extrapolated to the scale of data centers, the gap in power consumption becomes even more pronounced. An AI agent using a 70-billion-parameter LLM, comparable to the scale of today’s commercial AI services, consumed an average of 348.41 Wh to process a single query. This is approximately 136.6 times higher than the 2.55 Wh consumed by a conventional one-shot question-answering approach.

The resulting power burden on data centers could grow exponentially. If at least 71.4 million daily active ChatGPT users were to shift from conventional question-answering to AI agents, the required power would jump from 7.6 MW to 1 GW. If approximately 13.7 billion Google searches per day were to be handled by AI agents in the future, the required power would amount to 198.9 GW per day. This is a massive figure equivalent to nearly half of the average total power load of the United States, which stands at 476.9 GW.

“The significance of this study lies in showing that improving AI agent capabilities does not merely require more computation, but also creates a new class of burden across data center infrastructure,” the research team said. “Future research on AI agents must consider not only how to build smarter agents, but also how to operate them efficiently within limited infrastructure and power budgets.”

Kiyoung Choi, former professor at Seoul National University’s Department of Electrical and Computer Engineering and former Minister of Science and ICT, noted that in addition to the challenges surrounding AI data centers, semiconductor shortages, and material constraints, power supply is also a highly difficult issue. He emphasized the significance of this paper in sounding an alarm that endlessly expanding data center scale and cost in response to AI agents’ demands may not be sustainable.

The study was conducted by Master’s student Byungjun Shin, MS-PhD integrated student Jinha Jung, and PhD student Jiin Kim.

※ Paper title: The Cost of Dynamic Reasoning: Demystifying AI Agents and Test-Time Scaling from an AI Infrastructure Perspective

※ GitHub technical open source: https://github.com/VIA-Research/AgentBench

This work was supported by the SW Computing Industry Core Technology Development Program, including the SW Starlab Program, funded by the Ministry of Science and ICT and the Institute of Information & Communications Technology Planning & Evaluation, as well as by the Samsung Science & Technology Foundation.

Major Media Coverage

Sungwon Nah, Ph.D. Candidate in Prof. Hyunchul Shim’s Lab, Claims First Prize at IEEE IV 2026 for Breakthrough EV Aerodynamics

<Ph.D. candidate Sungwon Nah and Professor Hyunchul Shim (top right)>

Future vehicles will control not only their motors but also the flow of air around them. Researchers at KAIST have developed an active aerodynamic technology that continuously adjusts airflow according to driving conditions, improving both the performance and safety of high-performance electric vehicles.

The research team led by Professor Hyunchul Shim in the School of Electrical Engineering, with support from Hyundai Motor Company, has developed a Multi-Surface Active Aerodynamic System capable of integrated control of multiple aerodynamic devices installed at the front and rear of a vehicle. The proposed technology was successfully implemented and validated on a full-scale vehicle under real circuit driving conditions.

The study, led by Sungwon Nah, a Ph.D. candidate and first author, received the First Prize Best Student Paper Award at the 2026 IEEE Intelligent Vehicles Symposium (IV 2026), held in Detroit, Michigan, USA, in June 2026.

This award represents the highest honor presented to the most outstanding student paper at the conference. The research was first selected for an Oral Presentation, a distinction awarded to only approximately 8.5% of accepted papers, and subsequently won the First Prize Best Student Paper Award following the final evaluation of the oral presentations, recognizing both its originality and technical excellence.

The IEEE Intelligent Vehicles Symposium (IV), organized by the IEEE Intelligent Transportation Systems Society (ITSS), is one of the world’s premier international conferences in the field of intelligent vehicles. It brings together leading researchers from universities, research institutes, and the automotive industry to present the latest advances in autonomous driving, vehicle control, artificial intelligence (AI), and future mobility technologies.

To achieve integrated control of four active aerodynamic devices mounted on the front and rear of the vehicle, the research team first established an aerodynamic model based on wind tunnel experiments that accurately captured the aerodynamic characteristics of each device.

Building on this model, the researchers developed a real-time control framework that continuously analyzes driving conditions, including vehicle speed and steering state, to determine the optimal aerodynamic mode. The proposed framework was validated through both high-fidelity vehicle simulations and full-scale vehicle experiments.

Unlike many previous studies that have been limited to simulation-based validation, the team conducted full-scale vehicle testing at the Korea International Circuit (KIC), an FIA Grade 1 circuit certified to host Formula One races, thereby demonstrating the practicality and effectiveness of the proposed technology under real driving conditions.

Experimental results showed that the proposed active aerodynamic system effectively improved lap time, braking performance, cornering performance, and overall vehicle stability in high-performance electric vehicles. Furthermore, the consistent performance improvements observed in both simulation and real-world experiments demonstrate the technology’s strong potential for future applications not only in high-performance electric vehicles but also in autonomous vehicles and Software-Defined Vehicles (SDVs), whose capabilities can be continuously enhanced through software updates.

The first author, Sungwon Nah, also served as the team leader of KAIST EURECAR in the Indy Autonomous Challenge (IAC), an international autonomous racing competition in which Professor Shim’s laboratory has participated since 2021. Under his leadership, the team successfully achieved autonomous driving at speeds of up to 290 km/h, accumulating world-class expertise in high-speed autonomous vehicle control that directly contributed to this research.

Professor Hyunchul Shim said, “This recognition at one of the world’s most prestigious conferences on intelligent vehicles is particularly meaningful because it is built upon the practical experience our laboratory has accumulated through high-speed autonomous racing competitions such as the Indy Autonomous Challenge. We expect this research to contribute not only to improving the performance of high-performance electric vehicles but also to enhancing the safety and driving capabilities of future intelligent vehicles.“

The paper was led by Sungwon Nah (Ph.D. candidate) as first author, with Seungjin Yang, Youngjun Hwang, and Jungha Wang (M.S. students) as co-authors. Researchers Janghan Choi, JungSoo Lee, and JungKi Son from Hyundai Motor Company also participated as collaborators.

Paper Title: Development of an Active Aerodynamic System for Improving Circuit Driving Performance of High-Performance Electric Vehicles

Prof. SangHyeon Kim’s Team Develops Next-Generation Ultra-Compact Optical Modulator to Overcome Power and Speed Limits in AI Data Centers

<(Back row, from left) KAIST Master’s student Yong-Hwan Han, Ph.D. student Dong-Gil Kang
(Front row, from left) KAIST Master’s student Bin Yoon, Master’s student Soo-Hyun Kim
(Top right inset, from left) KAIST Prof. SangHyeon Kim, KAIST Ph.D. student Shin-Hyung Lee, Ph.D. student In-Ki Kim>

As the explosive growth of artificial intelligence (AI) services drives up power consumption and data bottlenecks in data centers, research teams from our department have developed a next-generation optical modulator—a critical device for optical communications that converts electrical signals into light to transmit data. This breakthrough is expected to serve as a foundational technology for future AI data centers by enabling higher data throughput with significantly lower power consumption.

Led by Professor SangHyeon Kim, the research team developed the next-generation optical modulator by integrating the distinct advantages of different semiconductor materials. The study was conducted in collaboration with Dr. Jae-Hoon Han of the Korea Institute of Science and Technology (KIST, President Sang-Rok Oh), Dr. Jong-Min Kim of the Korea Nano Technology Development Center (KANC, President Dae-Suk Byun), and Samsung Electronics’ Advanced Packaging Business unit.

< Conceptual diagram of Co-Packaged Optics (CPO) >

In AI data centers, countless servers and semiconductor chips exchange massive volumes of data in real time. In this architecture, the performance of the optical modulator is key to determining both data transmission speed and energy efficiency. However, conventional silicon-based optical modulators suffer from high heat dissipation and sensitivity to temperature fluctuations. Furthermore, they face a fundamental trade-off: improving efficiency reduces speed, while increasing speed compromises efficiency.

To overcome these limitations, the team utilized a temperature-resilient Mach-Zehnder structure—an optical device configuration that controls signals using phase differences in light paths. On top of the silicon waveguide (the channel through which light travels), they integrated a thin film of indium gallium arsenide phosphide (InGaAsP), a III-V compound semiconductor known for its superior electro-optic responsiveness. By adding this specialized compound semiconductor onto silicon, the team succeeded in reducing device size while achieving vastly superior light signal control.

Previously, academic attempts to use this structure faced a trade-off where improving efficiency caused charge accumulation, which in turn slowed down operating speed. The research team overcame this barrier by combining depletion-mode operation—which boosts speed by reducing internal capacitance and charge buildup—with the Franz-Keldysh effect, a phenomenon where electric fields alter a material’s optical absorption characteristics.

As a result, the team achieved both a world-class figure-of-merit modulation efficiency (0.146 V·cm) and a high-speed operating bandwidth (26.3 GHz) within an ultra-compact device measuring just 500 μm in length—tens of times thinner than a strand of human hair. This milestone proves that high-speed data transmission can be achieved with less power, offering a viable solution to simultaneously boost energy efficiency and processing speeds in AI data centers.

<(Left) Scanning Electron Microscopy (SEM) image of the III-V/Si SISCAP optical modulator
(Right) Cross-sectional Transmission Electron Microscopy (TEM) image>

This study carries significant weight as it provides a practical technology to address the escalating power demand and data transmission bottlenecks in modern AI infrastructure. Given its compact footprint and low-power characteristics, the technology is expected to serve as a core component for next-generation optical communication chips and Co-Packaged Optics (CPO)—an advanced packaging architecture that integrates semiconductor chips and optical components into a single package.

Furthermore, by breaking the long-standing trade-off between efficiency and speed to enhance both metrics simultaneously, the research marks a critical turning point toward the realization of ultra-high-speed optical communication systems and next-generation AI data centers.

Professor SangHyeon Kim stated, “To tackle power consumption and data transmission bottlenecks in AI data centers, our team has been developing various optical device technologies, including Silicon Photonics and Micro-LEDs. This breakthrough simultaneously elevates both the efficiency and speed of optical modulators, and we anticipate its broad application in next-generation AI data centers and ultra-high-speed optical communication systems.”

Ph.D. student Dong-Gil Kang from the KAIST School of Electrical Engineering led the study as the first author. The findings were presented on June 17 at the VLSI Symposium on Technology and Circuits, one of the world’s premier academic conferences in semiconductor devices.

Paper Title: Depletion-Mode III-V/Si SISCAP Mach-Zehnder Modulator Breaking the Efficiency-Bandwidth Trade-Off for Co-Packaged Optics
Abstract Link: https://vlsi26.mapyourshow.com/8_0/sessions/session-details.cfm?scheduleid=202

This research was conducted with support from the National Research Foundation of Korea (NRF) Mid-Career Researcher Program, the Doctoral Student Research Grant Program, and Samsung Electronics.

Prof. Changick Kim’s Team Ranks 1st at CVPR 2026 with Breakthrough AI Vision Tech ‘Upsample Anything’

A research team led by Professor Changick Kim from EE, through joint research with researchers from MIT and Microsoft, has developed ‘Upsample Anything’—a universal technology that enhances AI vision performance even with limited GPU memory.

Submitted to CVPR 2026, the world’s most prestigious computer vision and AI conference, the paper achieved the extraordinary feat of ranking 1st overall among all 4,089 submissions.

Furthermore, in recognition of its highly efficient utilization of computational resources, the technology was awarded the ‘CVPR Compute Gold Star’—an elite distinction presented to only 18 of the submitted papers. The team was also named a ‘Transparency Champion’ for its outstanding contributions to research transparency and reproducibility.

This sweeping success widely recognizes the core elements of responsible AI research, encompassing not only raw performance but also computational efficiency, open-source code disclosure, and experimental reproducibility.

Recently, humanoid robots, autonomous driving systems, and AI based on world models (AI models that learn and predict the physical environment and changes of the real world) have been compressing input images into low-resolution features (core information extracted from images by AI) to increase computational speed and reduce memory usage.

However, during the compression process, a problem occurs where important visual information, such as small objects, thin structures, and minute defects, is lost. Conversely, processing all images at high resolution from the beginning requires massive GPU memory and computational resources, making real-time processing difficult. This has remained an unresolved challenge for a long time in situations where small devices like smartphones or robots, where mobility is crucial, must precisely perceive their surrounding environment.

To overcome these limitations, the research team developed a training-free (requiring no additional data training) upsampling technology that restores low-resolution feature information into high resolution by utilizing the edge and structural information of the input image.

Existing technologies required a separate retraining or complex optimization process to be applied to new environments or data. In contrast, ‘Upsample Anything’ developed by the research team can find the optimal restoration method using just a single input image, allowing it to be immediately applied to various environments.

In addition, by compressing and utilizing only core information instead of storing and processing all visual information at high resolution, GPU memory usage was significantly reduced. Based on a 224×224 size image (approximately 50,000 pixels) widely used in AI research, the research team restored visual information close to the original with a short calculation of about 0.4 seconds, achieving a performance that improves GPU memory efficiency by up to 16 times.

This means that artificial intelligence can perceive its surrounding environment more precisely even with limited computational resources. Therefore, this technology is expected to be widely used in various next-generation artificial intelligence fields, such as small devices like smartphones, as well as humanoid robots that need to accurately identify and manipulate small objects, autonomous driving systems, and on-device AI.

Professor Changick Kim said, “This technology is an algorithm that can significantly increase the visual precision of artificial intelligence with fewer resources, and it is expected to accelerate the commercialization of humanoid robots and on-device AI.” He added, “It is even more meaningful because it was recognized at CVPR not only for its performance but also for its computational efficiency and research transparency.”

This research was participated in by KAIST PhD student Minseok Seo as the first author, and this achievement was presented on June 7 at ‘CVPR 2026,’ the world’s most prestigious conference in the field of artificial intelligence and computer vision.

Paper Title: Upsample Anything: A Simple and Hard to Beat Baseline for Feature Upsampling
DOI:10.48550/arXiv.2511.16301
Author Information: Minseok Seo (First Author), Mark Hamilton (MIT, Microsoft, Second Author), Changick Kim (Corresponding Author)

Archives: 연구 > 연구성과

Prof. David Hyunchul Shim’s Team Successfully Demonstrates Naval Vessel Steering with physical AI Humanoid Robot ‘PIBOT’

Four Researchers from Prof. Insu Yun’s Lab Named Microsoft MSRC 2026 ‘Most Valuable Researchers’

Prof. Ian Oakley’s Team Develops ‘ComiXR’, an Extended Reality (XR) Platform for Reading and Creating Comics

Professor Changick Kim’s Team Develops Key Technology to Make Personalized AI Safer

Prof. Jimin Kwon’s Team Develops Optical Image-Based Automation for 2D Semiconductor Screening and Device Fabrication

Prof. Seunghyup Yoo’s Team Implements Distortion-Free Stretchable Display Platforms

Professor Minsoo Rhu’s Team First Reveal the “Hidden Power Cost” of AI Agents

Sungwon Nah, Ph.D. Candidate in Prof. Hyunchul Shim’s Lab, Claims First Prize at IEEE IV 2026 for Breakthrough EV Aerodynamics

Prof. SangHyeon Kim’s Team Develops Next-Generation Ultra-Compact Optical Modulator to Overcome Power and Speed Limits in AI Data Centers

Prof. Changick Kim’s Team Ranks 1st at CVPR 2026 with Breakthrough AI Vision Tech ‘Upsample Anything’

About Us

Research

EE-X

AI in EE

People & Life

Academics

Admissions

News & Event

Give