A5088806568- Advanced Memory and Neural Computing
- Phase-change materials and chalcogenides
- Ferroelectric and Negative Capacitance Devices
- Korean Urban and Social Studies
- Neural Networks and Reservoir Computing
- Polymer composites and self-healing
- Natural Language Processing Techniques
- Neural dynamics and brain function
- Semiconductor materials and devices
- Transition Metal Oxide Nanomaterials
- Topic Modeling
- Liquid Crystal Research Advancements
- CCD and CMOS Imaging Sensors
- Spine and Intervertebral Disc Pathology
- Advanced Text Analysis Techniques
- Engineering Applied Research
- Orthopedic Surgery and Rehabilitation
- Flame retardant materials and properties
- Advanced Data Storage Technologies
- Stroke Rehabilitation and Recovery
- Chalcogenide Semiconductor Thin Films
- Photopolymerization techniques and applications
- Urban and spatial planning
- Synthesis and properties of polymers
- Web Data Mining and Analysis
Seoul National University
2014-2024
Gwangju Institute of Science and Technology
2011-2024
Korea University
2002-2024
IC Design Education Center
2024
Honam University
2024
Chungnam National University
2019-2023
Kyungil University
2023
Ajou University
2009-2022
Institute for Basic Science
2020-2022
IBM (United States)
2011-2021
In this paper, recent progress of phase change memory (PCM) is reviewed. The electrical and thermal properties materials are surveyed with a focus on the scalability their impact device design. Innovations in structure, cell selector, strategies for achieving multibit operation 3-D, multilayer high-density arrays described. scaling PCM illustrated experimental results using special test structures novel material synthesis. Factors affecting reliability discussed.
Dense crossbar arrays of non-volatile memory (NVM) devices represent one possible path for implementing massively-parallel and highly energy-efficient neuromorphic computing systems. We first review recent advances in the application NVM to three paradigms: spiking neural networks (SNNs), deep (DNNs), 'Memcomputing'. In SNNs, synaptic connections are updated by a local learning rule such as spike-timing-dependent-plasticity, computational approach directly inspired biology. For DNNs, can...
While naive Bayes is quite effective in various data mining tasks, it shows a disappointing result the automatic text classification problem. Based on observation of for natural language text, we found serious problem parameter estimation process, which causes poor results domain. In this paper, propose two empirical heuristics: per-document normalization and feature weighting method. these are somewhat ad hoc methods, our proposed classifier performs very well standard benchmark...
We survey progress in the PCM field over past five years, ranging from large-scale demonstrations to materials improvements for high–temperature retention and faster switching. Both new cell designs that support lower-power switching are discussed, as well higher reliability long cycling endurance. Two paths towards density discussed: through 3D integration by combination of 3D-capable access devices, multiple bits per cell, understanding managing resistance drift caused structural...
There is a significant need to build efficient non-von Neumann computing systems for highly data-centric artificial intelligence related applications. Brain-inspired one such approach that shows promise. Memory expected play key role in this form of and, particular, phase-change memory (PCM), arguably the most advanced emerging non-volatile technology. Given lack comprehensive understanding working principles brain, brain-inspired likely be realized multiple levels inspiration. In first...
Recent advances in neuroscience together with nanoscale electronic device technology have resulted huge interests realizing brain-like computing hardwares using emerging memory devices as synaptic elements. Although there has been experimental work that demonstrated the operation of element at single level, network level studies limited to simulations. In this work, we demonstrate, experiments, array associative learning phase change connected a grid like configuration similar organization...
We demonstrate a neuromorphic core with 64k-cell phase change memory (PCM) synaptic array (256 axons by 256 dendrites) in-situ learning capability. configurable on-chip neuron circuits perform leaky integrate and fire (LIF) weight update based on spike-timing dependent plasticity (STDP). 2T-1R PCM unit cell design separates LIF STDP paths, minimizing circuit size. The implementation of algorithm along structure enables both to operate asynchronously simultaneously within the array, avoiding...
Background and Purpose —After stroke, many individuals have chronic unilateral motor dysfunction in the upper extremity that severely limits their functional movement control. The purpose of this study was to determine effect electromyography-triggered neuromuscular electrical stimulation on wrist finger extension muscles who had a stroke ≥1 year earlier. Methods —Eleven volunteered participate were randomly assigned either experimental group (7 subjects) or control (4 subjects). After...
Background and Purpose — Overcoming chronic hemiparesis from a cerebrovascular accident (CVA) can be challenging for many patients, especially after the first 12 months CVA. With use of established motor control theories, present study investigated electromyogram (EMG)-triggered neuromuscular stimulation bilateral coordination training. Methods Twenty-five CVA subjects volunteered to participate in this recovery protocol study. Subjects were randomly assigned 1 3 groups: (1) coupled...
Feasibility of a high speed pattern recognition system using 1k-bit cross-point synaptic RRAM array and CMOS-based neuron chip has been experimentally demonstrated. Learning capability neuromorphic comprising synapses CMOS neurons confirmed experimentally, for the first time.
Organic neuromorphic computing/sensing platforms are a promising concept for local monitoring and processing of biological signals in real time. Neuromorphic devices sensors with low conductance power consumption high low-impedance sensing desired. However, it has been struggle to find materials fabrication methods that satisfy both these properties simultaneously single substrate. Here, nanofiber channels self-formed ion-blocking layer fabricated create organic electrochemical transistors...
Abstract Memristors, or memristive devices, have attracted tremendous interest in neuromorphic hardware implementation. However, the high electric-field dependence conventional filamentary memristors results either digital-like conductance updates gradual switching only a limited dynamic range. Here, we address parameter, reduction probability of Ag cations medium, and ultimately demonstrate cluster-type analogue memristor. Ti nanoclusters are embedded into densified amorphous Si for...
Thermal conduction in GeSbTe films strongly influences the writing energy and time for phase change memory (PCM) technology. This study measures thermal conductivity of Ge2Sb2Te5 between 25 340°C layers with thicknesses near 60, 120, 350nm. A strong thickness dependence is attributed to a combination boundary resistance (TBR) microstructural imperfections. Stoichiometric variations significantly alter transition temperatures but do not impact at given temperature. work makes progress on...
<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> Thermal interfaces play a key role in determining the programming energy of phase-change memory (PCM) devices. This letter reports picosecond thermoreflectance measurements thermal boundary resistance (TBR) at TiN/GST and Al/TiN interfaces, as well intrinsic conductivity fcc GST between 30 <formula formulatype="inline"><tex Notation="TeX">$^{\circ}\hbox{C}$</tex></formula> 325 Notation="TeX">$^{...
We performed a multicenter, open, randomized, clinical study of autologous cultured osteoblast injection for long-bone fracture, to evaluate the fracture healing acceleration effect and safety osteoblasts. Sixty-four patients with fractures were randomly divided into two groups, i.e. those who received no treatment. The sum difference in callus formation scores after four eight weeks, was used as first efficacy variable. group showed statistical significance, there specific patient...
While atomic vibrations dominate thermal conduction in the amorphous and face-centered cubic phases of Ge2Sb2Te5, electrons hexagonal closed-packed (hcp) phase. Here we separate electron phonon contributions to interface volume resistances for three using time-domain thermoreflectance electrical contact resistance measurements. Even when film-normal (i.e., 70% hcp phase), their contribution heat is overwhelmed by phonons high-quality interfaces with metallic TiN.
Optical diffraction tomography (ODT) provides label-free three-dimensional (3D) refractive index (RI) measurement of biological samples.However, due to the nature RI values specimens, ODT has limited access molecular specific information.Here, we present an optical setup combining with three-channel 3D fluorescence microscopy, enhance specificity measurement.The distribution and deconvoluted images HeLa cells NIH-3T3 are measured, cross-correlative analysis between live presented.
We present for the first time in-depth analysis of outstanding endurance characteristics an ALD-based confined phase change memory (PCM) [1] with a thin metallic liner. Experimental results confirm that both proper liner and pore cell structure are required reliability advantage. This PCM is found to be immune classic failure mechanisms. The void-free yields new record (2×10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">12</sup> cycles)...
Discovering and understanding new materials with desired properties are at the heart of science research, machine learning (ML) has recently offered special shortcuts to ultimate goal. Thanks nourishment computer hardware computational chemistry, development calculated scientific data repositories could fuel ML models investigate vast space. At this moment, revolutionary paradigm is urgent, Review aims deliver comprehensive information about collaboration science. This summarizes recent...
Abstract Nonvolatile memory (NVM)‐based neuromorphic computing has been attracting considerable attention from academia and the industry. Although it is not completely successful yet, remarkable achievements have reported pertaining to synaptic devices that can leverage NVM capable of storing multiple states. The analog performing computation similar biological nerve systems are crucial in energy‐efficient systems. To use as an device, researchers focus on improving device characteristics....