%4b%79%75%6e%67%20%4d%69%6e%20%4b%69%6d
-
KAIST Develops Semiconductor Neuron That Tunes Noise to Selectively Process Signals
In electronic devices, irregular fluctuations in signals are generally referred to as “noise.” Because noise interferes with accurate information processing, conventional semiconductor technology has mainly treated it as something to be reduced or eliminated. However, neurons in the human brain do not respond in exactly the same way every time, even to the same stimulus. Tiny internal variations in neurons change when and how often neurons are fired, and this probabilistic operation is one of the brain’s key information-processing features. Inspired by this, KAIST researchers have developed a next-generation semiconductor technology that does not remove current noise generated in memristors, but instead tunes it to a desired level and uses it to process different types of signals.
KAIST (President Choongsik Bae) announced on the August 16 that a research team led by Professor Kyung Min Kim from the Department of Materials Science and Engineering has developed a new neuromorphic neuron technology that uses noise generated in semiconductor devices for information processing, enabling selective encoding of time-series signals across different frequency bands.
※ Neuromorphic technology: A technology that processes information by mimicking the way the human brain and neurons operate.
In general, noise generated in semiconductors is regarded as an obstacle to accurate signal processing. For this reason, most electronic devices are designed to reduce or eliminate noise as much as possible. The human brain, however, works differently. Neurons, the nerve cells of the brain, do not always respond in the same way to the same stimulus because of internal probabilistic fluctuations. This irregularity actually helps the brain flexibly respond to a wide range of situations and sensory signals.
The research team used a memristor in this study. A memristor is a semiconductor device that changes its resistance state in response to electrical stimulation and remembers that state. Until now, current noise generated in memristors has mainly been used for random number generation, which creates unpredictable numbers, or for probabilistic computing.
However, previous studies have largely focused on using the inherent randomness of memristors as it is. Technologies that can tune probabilistic response characteristics according to need had not been sufficiently realized.
The key insight of this study is that when the resistance state of a memristor is changed, the magnitude and behavior of its current noise also change. By presetting the resistance state of the memristor, the probability of spike generation and the response range can vary even under the same input. Using this principle, the research team implemented a “programmable probabilistic neuron (PPN)” that treats noise not simply as instability, but as an information-processing resource that can be tuned in a desired way.
This neuron can be configured to respond differently depending on how rapidly an input signal changes, in other words, its frequency. By changing only the resistance state of the memristor, the same circuit can be switched to respond sensitively to slow human activity signals in the hertz (Hz) range or fast speech signals in the kilohertz (kHz) range. Hz and kHz are units that indicate how many times a signal repeats per second, with 1 kHz equal to 1,000 Hz.
In simple terms, a single artificial neuron can be reconfigured according to the speed of the signal it needs to process. When processing slowly changing signals such as human movement, it can operate in a way suited to slow variations; when processing rapidly changing signals such as speech, it can be adjusted to capture short and fast changes effectively.
The research team verified the technology using signals with different frequency ranges. The system encoded and classified human activity signals in the Hz range and speech signals in the kHz range, achieving accuracies of 94.8% in human activity recognition and 95.0% in speech recognition.
Professor Kyung Min Kim said, “The significance of this study lies in demonstrating that memristor noise can be harnessed as a tunable information-processing resource, rather than simply treated as an error or instability,” adding, “Because the same hardware can be reconfigured for signals of different speeds and frequencies, it could be used as a signal-processing technology for future low-power edge neuromorphic systems.”
This study was led by Dr. Do Hoon Kim from the Department of Materials Science and Engineering as first author, and was published in the internationally renowned materials science journal Advanced Materials on August 05.
Paper title: Noise-Tunable Memristor Enabling Programmable Probabilistic Neurons for Frequency-Selective Time-Series Signal Encoding,
DOI: https://doi.org/10.1002/adma.74529
This research was supported by the Basic Research Program in Science and Engineering and the PIM Artificial Intelligence Semiconductor Core Technology Development Program of the Ministry of Science and ICT and the National Research Foundation of Korea.
2026.08.18 View 1029 -
KAIST Develops Semiconductor Neuron that Remembers and Responds Like the Brain
<(From left, clockwise) Professor Kyung Min Kim, Min-Gu Lee, Dae-Hee Kim, Dr. Han-Chan Song, Tae-Uk Ko, Moon-Gu Choi, and Eun-Young Kim>
The human brain does more than simply regulate synapses that exchange signals; individual neurons also process information through “intrinsic plasticity,” the adaptive ability to become more sensitive or less sensitive depending on context. Existing artificial intelligence semiconductors, however, have struggled to mimic this flexibility of the brain. A KAIST research team has now developed next-generation, ultra-low-power semiconductor technology that implements this ability as well, drawing significant attention.
KAIST (President Kwang Hyung Lee) announced on September 28 that a research team led by Professor Kyung Min Kim of the Department of Materials Science and Engineering developed a “Frequency Switching Neuristor” that mimics “intrinsic plasticity,” a property that allows neurons to remember past activity and autonomously adjust their response characteristics.
“Intrinsic plasticity” refers to the brain’s adaptive ability- for example, becoming less startled when hearing the same sound repeatedly, or responding more quickly to a specific stimulus after repeated training. The “Frequency Switching Neuristor” is an artificial neuron device that autonomously adjusts the frequency of its signals, much like how the brain becomes less startled by repeated stimuli or, conversely, increasingly sensitive through training.
The research team combined a “volatile Mott memristor,” which reacts momentarily before returning to its original state, with a “non-volatile memristor,” which remembers input signals for long periods of time. This enabled the implementation of a device that can freely control how often a neuron fires (its spiking frequency).
<Figure 1. Conceptual comparison between a neuron and a frequency-tunable neuristor. The intrinsic plasticity of brain neurons regulates excitability through ion channels. Similarly, the frequency-tunable neuristor uses a volatile Mott device to generate current spikes, while a non-volatile VCM device adjusts resistance states to realize comparable frequency modulation characteristics>
In this device, neuronal spike signals and memristor resistance changes influence each other, automatically adjusting responses. Put simply, it reproduces within a single semiconductor device how the brain becomes less startled by repeated sounds or more sensitive to repeated stimuli.
To verify the effectiveness of this technology, the researchers conducted simulations with a “sparse neural network.” They found that, through the neuron’s built-in memory function, the system achieved the same performance with 27.7% less energy consumption compared to conventional neural networks.
They also demonstrated excellent resilience: even if some neurons were damaged, intrinsic plasticity allowed the network to reorganize itself and restore performance. In other words, artificial intelligence using this technology consumes less electricity while maintaining performance, and it can compensate for partial circuit failures to resume normal operation.
Professor Kyung Min Kim, who led the research, stated, “This study implemented intrinsic plasticity, a core function of the brain, in a single semiconductor device, thereby advancing the energy efficiency and stability of AI hardware to a new level. This technology, which enables devices to remember their own state and adapt or recover even from damage, can serve as a key component in systems requiring long-term stability, such as edge computing and autonomous driving.”
This research was carried out with Dr. Woojoon Park (now at Forschungszentrum Jülich, Germany) and Dr. Hanchan Song (now at ETRI) as co-first authors, and the results were published online on August 18 in Advanced Materials (IF 26.8), a leading international journal in materials science.
※ Paper title: “Frequency Switching Neuristor for Realizing Intrinsic Plasticity and Enabling Robust Neuromorphic Computing,” DOI: 10.1002/adma.202502255
This research was supported by the National Research Foundation of Korea and Samsung Electronics.
2025.09.30 View 4840 -
KAIST Team Develops an Insect-Mimicking Semiconductor to Detect Motion
The recent development of an “intelligent sensor” semiconductor that mimics the optic nerve of insects while operating at ultra-high speeds and low power offers extensive expandability into various innovative technologies. This technology is expected to be applied to various fields including transportation, safety, and security systems, contributing to both industry and society.
On February 19, a KAIST research team led by Professor Kyung Min Kim from the Department of Materials Science and Engineering (DMSE) announced the successful developed an intelligent motion detector by merging various memristor* devices to mimic the visual intelligence** of the optic nerve of insects.
*Memristor: a “memory resistor” whose state of resistance changes depending on the input signal
**Visual intelligence: the ability to interpret visual information and perform calculations within the optic nerve
With the recent advances in AI technology, vision systems are being improved by utilizing AI in various tasks such as image recognition, object detection, and motion analysis. However, existing vision systems typically recognize objects and their behaviour from the received image signals using complex algorithms. This method requires a significant amount of data traffic and higher power consumption, making it difficult to apply in mobile or IoT devices.
Meanwhile, insects are known to be able to effectively process visual information through an optic nerve circuit called the elementary motion detector, allowing them to detect objects and recognize their motion at an advanced level. However, mimicking this pathway using conventional silicon integrated circuit (CMOS) technology requires complex circuits, and its implementation into actual devices has thus been limited.
< Figure 1. Working principle of a biological elementary motion detection system. >
Professor Kyung Min Kim’s research team developed an intelligent motion detecting sensor that operates at a high level of efficiency and ultra-high speeds. The device has a simple structure consisting of only two types of memristors and a resistor developed by the team. The two different memristors each carry out a signal delay function and a signal integration and ignition function, respectively. Through them, the team could directly mimic the optic nerve of insects to analyze object movement.
< Figure 2. (Left) Optical image of the M-EMD device in the left panel (scale bar 200 μm) and SEM image of the device in the right panel (scale bar: 20 μm). (Middle) Responses of the M-EMD in positive direction. (Right) Responses of the M-EMD in negative direction. >
To demonstrate its potential for practical applications, the research team used the newly developed motion detector to design a neuromorphic computing system that can predict the path of a vehicle. The results showed that the device used 92.9% less energy compared to existing technology and predicted motion with more accuracy.
< Figure 3. Neuromorphic computing system configuration based on motion recognition devices >
Professor Kim said, “Insects make use of their very simple visual intelligence systems to detect the motion of objects at a surprising high speed. This research is significant in that we could mimic the functions of a nerve using a memristor device.” He added, “Edge AI devices, such as AI-topped mobile phones, are becoming increasingly important. This research can contribute to the integration of efficient vision systems for motion recognition, so we expect it to be applied to various fields such as autonomous vehicles, vehicle transportation systems, robotics, and machine vision.”
This research, conducted by co-first authors Hanchan Song and Min Gu Lee, both Ph.D. candidates at KAIST DMSE, was published in the online issue of Advanced Materials on January 29.
This research was supported by the Mid-Sized Research Project by the National Research Foundation of Korea, the Next-Generation Intelligent Semiconductor Technology Development Project, the PIM Artificial Intelligence Semiconductor Core Technology Development Project, the National Nano Fab Center, and the Leap Research Project by KAIST.
2024.02.29 View 15067 -
Energy-Efficient AI Hardware Technology Via a Brain-Inspired Stashing System
Researchers demonstrate neuromodulation-inspired stashing system for the energy-efficient learning of a spiking neural network using a self-rectifying memristor array
Researchers have proposed a novel system inspired by the neuromodulation of the brain, referred to as a ‘stashing system,’ that requires less energy consumption. The research group led by Professor Kyung Min Kim from the Department of Materials Science and Engineering has developed a technology that can efficiently handle mathematical operations for artificial intelligence by imitating the continuous changes in the topology of the neural network according to the situation. The human brain changes its neural topology in real time, learning to store or recall memories as needed. The research group presented a new artificial intelligence learning method that directly implements these neural coordination circuit configurations.
Research on artificial intelligence is becoming very active, and the development of artificial intelligence-based electronic devices and product releases are accelerating, especially in the Fourth Industrial Revolution age. To implement artificial intelligence in electronic devices, customized hardware development should also be supported. However most electronic devices for artificial intelligence require high power consumption and highly integrated memory arrays for large-scale tasks. It has been challenging to solve these power consumption and integration limitations, and efforts have been made to find out how the human brain solves problems.
To prove the efficiency of the developed technology, the research group created artificial neural network hardware equipped with a self-rectifying synaptic array and algorithm called a ‘stashing system’ that was developed to conduct artificial intelligence learning. As a result, it was able to reduce energy by 37% within the stashing system without any accuracy degradation. This result proves that emulating the neuromodulation in humans is possible.
Professor Kim said, "In this study, we implemented the learning method of the human brain with only a simple circuit composition and through this we were able to reduce the energy needed by nearly 40 percent.”
This neuromodulation-inspired stashing system that mimics the brain’s neural activity is compatible with existing electronic devices and commercialized semiconductor hardware. It is expected to be used in the design of next-generation semiconductor chips for artificial intelligence.
This study was published in Advanced Functional Materials in March 2022 and supported by KAIST, the National Research Foundation of Korea, the National NanoFab Center, and SK Hynix.
-Publication:
Woon Hyung Cheong, Jae Bum Jeon†, Jae Hyun In, Geunyoung Kim, Hanchan Song, Janho An, Juseong Park, Young Seok Kim, Cheol Seong Hwang, and Kyung Min Kim (2022)
“Demonstration of Neuromodulation-inspired Stashing System for Energy-efficient Learning of Spiking Neural Network using a Self-Rectifying Memristor Array,” Advanced FunctionalMaterials March 31, 2022 (DOI: 10.1002/adfm.202200337)
-Profile:
Professor Kyung Min Kimhttp://semi.kaist.ac.kr https://scholar.google.com/citations?user=BGw8yDYAAAAJ&hl=ko
Department of Materials Science and EngineeringKAIST
2022.05.18 View 18828