- KAIST MatriX
- The official newsletter of KAIST Institutes
- kmatrix.kaist.ac.kr
-
share newsletter
-
- Editorial Committee:
- Wanyeong Jung
- Namil Lee
- Minkyun Noh
- Haewon C. McJeon
- Kyoung-Ho Ha
A Neuromodulation System with Real-Time Neural Signals Recovery Overlapped Temporally and Spectrally with Stimulation Artifacts
Professor Chul Kim’s group at KAIST has developed a neuromodulation system that recovers neural signals in real time even when electrical stimulation artifacts overlap with them in both time and frequency. The system removes artifacts directly on a compact chip without prior calibration, supporting continuous monitoring during stimulation. The study was published online on January 20, 2026, in IEEE Transactions on Biomedical Circuits and Systems. Addressing signal loss during stimulation Electrical stimulation produces artifacts that are much larger than the neural signals researchers wish to observe. Discarding contaminated recording periods also removes neural information, while conventional frequency filtering cannot readily separate signals and artifacts that share the same frequencies. These limitations make it difficult to monitor neural activity continuously and use it to guide stimulation. Adaptive recovery directly on the chip The team developed an adaptive algorithm that estimates the artifact waveform and subtracts it from incoming recordings. ...read more
Featured Activities
-
Introduction to KI
PIM Semiconductor Design Research Center
The PIM Semiconductor Design Research Center at KAIST develops ultra-efficient AI hardware for LLMs, on-device AI, and embodied agents to overcome critical computing and energy bottlenecks.
-
Introduction to KI
Augmented Reality Research Center
ARRC at KAIST KI-ITAIC is expanding AR into augmented human research and pursuing symbiotic AIR4BTS through an open AI+XR platform for sharing experiences.
-
Research Highlight
Agricultural Workforce as a Potential Bottleneck of Future Cropland Availability
Professor Hyungjun Kim’s group developed a cropland projection model that simulates future agricultural workforce shortages and future cropland availability.
-
Research Highlight
Artificial intelligence to support cross-disciplinary climate change research
Professor McJeon’s team published a commentary for an AI foundation-model architecture that integrates climate, impact, and socioeconomic analysis for decision support.
-
Research Highlight
Stretchable High-Density Silicon–Liquid Metal Electronics for Robotic Visions
Professor Song’s group has developed a robotic vision system combining varifocal and lens-free imaging, enabling multiscale visual acquisition and depth sensing.
-
Research Highlight
High-titer, antibiotic-free, pilot-scale production of 1,3-propanediol by engineered Corynebacterium
Professor Lee’s group developed microbial cell factories for high-level 1,3-PDO production, with scalability confirmed at 300-L pilot scale.
-
Research Highlight
TEMPER Enables Programmable One-Pot DNA Synthesis through Temperature Control
Professor Yeongjae Choi’s group developed TEMPER, a one-pot DNA synthesis platform programmed solely by temperature without repeated solution exchange
-
Research Highlight
UNCHARTED: A Universal ML-based Framework for Circuit Simulation and Inverse Design
Professor Jung’s group developed UNCHARTED, an AI framework that learns component-level circuit behavior for fast transient simulation and inverse design.
-
Research Highlight
A 14.4-Gb/s 7.3-pJ/b Universal Vision Transformer OFDM Channel Estimation Processor for B5G/6G Communication Systems
Professor Lee’s team developed UniC-Vision, the first silicon-proven Vision Transformer channel estimation processor for B5G/6G AI-RAN systems.
-
Research Highlight
Beyond the Bell Curve Towards Safer, Faster Control for Autonomous Systems Under Real-World Uncertainty
Real-world disturbances do not always follow Gaussian distributions. The creates a challenging trade-off for safety-critical autonomous control: simple risk models can be misleading, while more flexible alternatives can be too computationally demanding for real-time use. Professor Hyo-Sang Shin's team at KAIST developed Spectral-IRA, combining non-Gaussian uncertainty propagation, data-driven calibration, and adaptive risk allocation to turn nonlinear probabilistic safety constraints into a tractable convex control problem. Theoretical analysis and simulations show that Spectral-IRA can keep risk within prescribed limits while remaining fast enough for real-time control.
-
Research Highlight
Human-Like Visual Attention Improves Gaze-Augmented Reward Modeling for Large Language Models
Professor Sang Ho Lee’s group developed a gaze-augmented reward modeling approach for LLMs that incorporates a psychologically grounded account of visual attention. The new approach improved human preference prediction across tested LLM models.
-
Activity
Featured Activities in Fall 2026
Visitor KIHAC 2026.03.04 Joint Research Meeting with Imperial College London KIHAC, Tomocube, Imperial College London, and Nexpring discussed collaboration on IVF and embryo analysis technologies. KIDR 2026.03.06 Prof. Junmin Wang, The University of Texas at Austin Prof. Junmin Wang at UT Austin visited KI-DR and gave as seminar on human-centered vehicle motion control and automation.
291 Daehak-ro Yuseong-gu Daejeon, 34141, Republic of Korea
Partnered with KAIST Breakthroughs and KAIST Compass