Search

Studies on neuromorphic computing systems are becoming increasingly important in the big-data-processing era as these systems are capable of energy-efficient parallel data processing and can overcome the present limitations owing to the von Neumann bottleneck. The Pt/WOx/ITO resistive random-access memory device can be used to implement versatile synapse functions because it possesses both volatile and nonvolatile characteristics. The gradual increase and decrease in the current of the Pt/WOx/ITO device with its uniform resistance state for endurance and retention enables additional synaptic applications that can be controlled using electric pulses. If the volatile and nonvolatile device properties are set through rehearsal and forgetting processes, the device can emulate various synaptic behaviors, such as potentiation and depression, paired-pulse facilitation, post-tetanic potentiation, image training, Hebbian learning rules, excitatory postsynaptic current, and Pavlov’s test. Furthermore, reservoir computing can be implemented for applications such as pattern generation and recognition. This emphasizes the various applications of future neuromorphic devices, demonstrating the various favorable characteristics of pulse-enhanced Pt/WOx/ITO devices.

We propose a hardware-friendly architecture of a convolutional neural network using a 32 × 32 memristor crossbar array having an overshoot suppression layer. The gradual switching characteristics in both set and reset operations enable the implementation of a 3-bit multilevel operation in a whole array that can be utilized as 16 kernels. Moreover, a binary activation function mapped to the read voltage and ground is introduced to evaluate the result of training with a boundary of 0.5 and its estimated gradient. Additionally, we adopt a fixed kernel method, where inputs are sequentially applied to a crossbar array with a differential memristor pair scheme, reducing unused cell waste. The binary activation has robust characteristics against device state variations, and a neuron circuit is experimentally demonstrated on a customized breadboard. Thanks to the analogue switching characteristics of the memristor device, the accurate vector–matrix multiplication (VMM) operations can be experimentally demonstrated by combining sequential inputs and the weights obtained through tuning operations in the crossbar array. In addition, the feature images extracted by VMM during the hardware inference operations on 100 test samples are classified, and the classification performance by off-chip training is compared with the software results. Finally, inference results depending on the tolerance are statistically verified through several tuning cycles.

In the year of 1991, I was a senior student at Seoul National University of Korea. I saw a young new professor working at a computer room of the Inter-university Semiconductor Research Center (ISRC). He brought his own Macintosh with a tiny 9-in monochrome monitor into the computer room, and he was making documents. He did not have any lab, and he didn’t have any students at that time. After I talked to him, I really wanted to become his student. He accepted me as one of the first lab members.

Synaptic devices store the synaptic weight in spiking neural networks (SNNs). However, because synaptic devices are based on memory cells, their synaptic weights are vulnerable to temperature variations, which significantly degrade network accuracy. To implement temperature-robust asynchronous SNNs, neurons with captive synaptic devices (CSD neurons) are proposed in this study. Captive synaptic devices that mimic the temperature characteristics of synaptic devices are located parallel to the integration capacitors. By using the captive synaptic devices for the membrane potential reset, the CSD neurons are inherently equipped to counteract temperature-induced variations. The validity of the CSD neurons was experimentally confirmed using two types of synaptic devices: poly Si-channel charge-trap flash synaptic devices and resistive random access memory devices. When applied to the street view house numbers (SVHN) dataset, the SNNs with CSD neurons successfully maintained a constant inference accuracy value (95.00 %) from 300 to 360 K, even if the output current of the synaptic devices doubled.

A lack of the T cell-inflamed tumor microenvironment limits the efficacy of immune checkpoint inhibitors (ICIs). Activation of stimulator of interferon genes (STING)-mediated innate immunity has emerged as a novel therapeutic approach in cancer therapy. 2′,3′-Cyclic GMP–AMP (cGAMP) is a natural STING agonist; however, cGAMP is subjected to endogenous degradation by ecto-nucleotide pyrophosphatase phosphodiesterase 1 (ENPP1). To improve the ICI response rate, we developed 29f, a novel ENPP1 inhibitor with phthalazin-1(2H)-one as the core scaffold. 29finhibited the cGAMP hydrolysis by ENPP1 in vitro (IC50= 68 nM) and enhanced the STING-mediated type I interferon response in both immune and tumor cells. 29fdemonstrated excellent metabolic stability and bioavailability (F= 65%). Orally administered 29fpromoted tumor growth inhibition in a CT26 syngeneic model and increased the anti-PD-L1 response. Furthermore, 29f-induced immunological memory prevented the tumor relapse against tumor rechallenge, suggesting the promising therapeutic potential of 29f.

In this study, we developed a fuse device for pruning implementation in a hardware neural network. A line-shaped fuse device was fabricated with aluminum metal and characterized in a one fuse and one resistive random-access memory (1F1R) structure. A blow time of 0.4 $\mu \text{s}$ and read endurance of $> 10^{{7}}$ were achieved, and the cut-off operation of the fuse was successfully verified in 1F1R. In addition, we developed a fuse design method for blow voltage and current via adjustment of the length, width, and thickness of the fuse. The adjustments were adopted to utilize the fuse in various synaptic devices. Finally, using simulations, we demonstrated a performance improvement due to the network pruning wherein the defective devices are disconnected by the fuse operations.

This paper suggests the practical implications of utilizing a high-density crossbar array with self-compliance (SC) at the conductive filament (CF) formation stage. By limiting the excessive growth of CF, SC functions enable the operation of a crossbar array without access transistors. An AlOx/TiOy, internal overshoot limitation structure, allows the SC to have resistive random-access memory. In addition, an overshoot-limited memristor crossbar array makes it possible to implement vector-matrix multiplication (VMM) capability in neuromorphic systems. Furthermore, AlOx/TiOystructure optimization was conducted to reduce overshoot and operation current, verifying uniform bipolar resistive switching behavior and analog switching properties. Additionally, extensive electric pulse stimuli are confirmed, evaluating long-term potentiation (LTP), long-term depression (LTD), and other forms of synaptic plasticity. We found that LTP and LTD characteristics for training an online learning neural network enable MNIST classification accuracies of 92.36%. The SC mode quantized multilevel in offline learning neural networks achieved 95.87%. Finally, the 32 × 32 crossbar array demonstrated spiking neural network-based VMM operations to classify the MNIST image. Consequently, weight programming errors make only a 1.2% point of accuracy drop to software-based neural networks.

This work focused on plasma-induced nitrogen-doped indium gallium zinc oxide (InGaZnO:N) based resistive switching devices. Nitrogen atoms in the InGaZnO:N can reduce the randomness of oxygen vacancy formation and conducting filaments, resulting in stable and uniform resistive switching (RS) and artificial synaptic performance. Optimized nitrogen and oxygen ion concentration improves the redox reaction depending on their reactivity with InGaZnO due to optimal VOconcentration in the switching layer confirmed by X-ray photoelectron spectroscopic (XPS) analysis. Device-to-device promising RS variability with ION/IOFFratio >10 confirmed with multiple devices with low average SET/RESET voltages of 1.35 /1.5V. Pulse-induced short-term memory learning characteristics with linear weight-update, including synaptic function of short-term potentiation (STP), paired-pulse facilitation (PPF), spike rate-dependent plasticity (SRDP), and experience dependent synaptic weight modification are successfully emulated and demonstrated. The optimization of InGaZnO:N-based memristors provides efficient temporal information processing capability, serving as the physical reservoir computing system, demonstrated through experimental pattern recognition. The potentiaiton/depression of conductance was used for for on-chip learning with the street view house numbers (SVHN) dataset. This study suggests that moderate nitrogen-doped ITO/InGaZnO:N/ITO memristors possess potential as synaptic devices for neuromorphic systems.

In this study, we demonstrate the implementation of programmable threshold logics using a 32 × 32 memristor crossbar array. Thanks to forming-free characteristics obtained by the annealing process, its accurate programming characteristics are presented by a 256-level grayscale image. By simultaneous subtraction between weighted sum and threshold values with a differential pair in an opposite way, 3-input and 4-input Boolean logics are implemented in the crossbar without additional reference bias. Also, we verify a full-adder circuit and analyze its fidelity, depending on the device programming accuracy. Lastly, we successfully implement a 4-bit ripple carry adder in the crossbar and achieve reliable operations by read-based logic operations. Compared to stateful logic driven by device switching, a 4-bit ripple carry adder on a memristor crossbar array can perform more reliably in fewer steps thanks to its read-based parallel logic operation.

This study describes a novel chemical route to grow ZnSthin films via atomic layer deposition (ALD). By using diethylzinc and 1,5‐pentanedithiol as precursors of Zn and S, respectively, ZnSfilms are grown on substrates with an atomic precision by repeating self‐limiting chemisorption of each precursor. The growth‐per‐cycle of the ALDprocess is around 0.1 Å per cycle at 150°C, and the as‐grown films are characterized to be amorphous ZnSby X‐ray diffraction and X‐ray photoelectron spectroscopy.

One-dimensional (1D) materials demonstrate anisotropic in-plane physical properties that enable a wide range of functionalities in electronics, photonics, valleytronics, optoelectronics, and catalysis. Here, we undertake an in-depth study of the growth mechanism for equimolar midentropy alloy of (NbTaTi)0.33S3nanoribbons as a model system for 1D transition metal trichalcogenide structures. To understand the thermodynamic and kinetic effects in the growth process, the energetically preferred phases at different synthesis temperatures and times are investigated, and the phase evolution is inspected at a sequence of growth steps. It is uncovered that the dynamics of the growth process occurs at four different stages via preferential incorporation of chemical species at high-surface-energy facets. Also, a sequence of temperature and time dependent nonuniform to uniform phase evolutions has emerged in the composition and structure of (NbTaTi)0.33S3which is described based on an anisotropic vapor–solid (V–S) mechanism. Furthermore, direct evidence for the 3D structure of the charge density wave (CDW) phase (width less than 100 nm) is provided by three-dimensional electron diffraction (3DED) in individual nanoribbons at cryogenic temperature, and detailed comparisons are made between the phases obtained before and after CDW transformation. This study provides important fundamental information for the design and synthesis of future 1D alloy structures.

• Multi-level operation of Al 2 O 3 /TiO x memristor devices for neuromorphic system • Analyzing analogue-grade weight modulation by adjusting compliance current up to 64 levels • Precisely controlled device state with compliance current and off-chip learning verification • Evaluating the recognition performance of convolutional neural network considering the number of quantization level and accuracy Multilevel operation is one of the most essential properties for synaptic devices to realize hardware artificial neural networks. Compliance current (I cc) adjustment is a multilevel programming method that can be utilized for a large-scale one-transistor and one-resistor (1T1R) array. It protects the devices from permanent breakdown by regulating abrupt switching. However, according to the reported literature so far, the number of conductance states in the I cc control method is insufficient to implement off-chip-trained neuromorphic systems. Therefore, we experimentally explore the feasibility of a larger number of conductance states using the I cc control method. We fabricated an Al 2 O 3 /TiO x -based resistive switching memory array, observed the conductance change while increasing I cc during set operations, and 64-level conductance states were statistically demonstrated. Furthermore, we verified that the 64-level states showed recognition performance close to that of a software-based neural network through off-chip learning of the convolutional neural network (CNN) structure. The fabricated synaptic device array with the I cc -control programming method is expected to contribute to the development of hardware neural network by reducing the information loss in the transfer process. [ABSTRACT FROM AUTHOR]

Due to their ability to function as dominant oncogenes, protein kinases have become favored targets in the quest for ‘molecularly-targeted’ cancer chemotherapeutics. The discovery of a large number of cancer-associated mutations in the kinome, and the progress in developing specific small-molecule kinase inhibitors has increased the need for accurate, reproducible, and efficient kinase activity-dependent cellular assay systems.

FCP1, a phosphatase specific for the carboxy-terminal domain of RNA polymerase II (RNAP II), was found to stimulate transcript elongation by RNAP II in vitro and in vivo. This activity is independent of and distinct from the elongation-stimulatory activity associated with transcription factor IIF (TFIIF), and the elongation effects of TFIIF and FCP1 were found to be additive. Genetic experiments resulted in the isolation of several distinct fcp1alleles. One of these alleles was found to suppress the slow-growth phenotype associated with either the reduction of intracellular nucleotide concentrations or the inhibition of other transcription elongation factors. Importantly, this allele of fcp1was found to be lethal when combined individually with two mutations in the second-largest subunit of RNAP II, which had been shown previously to affect transcription elongation.

A transcriptional repressor complex encoded by two essential genes,YDR1and BUR6, was isolated from Saccharomyces cerevisiaeand shown to be the functional counterpart of the human repressor complex Dr1-DRAP1. To elucidate the mechanism of repression by this complex, altered forms of Ydr1 and Bur6 were studied in vitro and in vivo. Deletion of the C-terminal 41 amino acids of Ydr1 resulted in loss of repressor activity and a growth defect, suggesting that the C-terminal domain of Ydr1 functions as a potent transcriptional repressor. A screen for extragenic suppressors of a cold-sensitive ydr1(ydr1cs) mutant led to the identification of recessive mutations in the SIN4gene, which encodes a component of the SRB-MED complex. The sin4alleles suppressed not onlyydr1csmutations but alsobur6csmutations. In contrast, deletion of the gal11gene, whose product is also a member of the SRB-MED complex, failed to suppress ydr1csandbur6csmutations, indicating that suppression is not due to general defects in the SRB-MED complex. Moreover, one of the sin4alleles, but not the sin4deletion, was found to specifically suppress the inviability of aydr1deletion, demonstrating that the essential function of Ydr1 becomes dispensable in a sin4mutant background. Biochemical analysis of the SRB-MED complex from the sin4suppressor strain revealed a structurally distinct form of the SRB-MED complex that lacks a subset of mediator subunits. These results define a delicate balance between positive and negative regulators of transcription operating through the Ydr1-Bur6 repressor complex.

The heterotetrameric Dr1-DRAP1 transcriptional repressor complex was functionally dissected. Dr1 was found to contain two domains required for repression of transcription. The tethering domain interacts with the TATA box binding protein and directs the repressor complex to the promoter. This tethering domain can be replaced by a domain conferring sequence-specific recognition to the repressor complex. In the absence of the tethering domain, Dr1 interacts with its corepressor DRAP1, but this interaction is not functional. The enhancement of Dr1-mediated repression of transcription by DRAP1 requires the tethering domain. The second domain of Dr1 is the repression domain, which is glutamine-alanine rich. A 65-amino-acid polypeptide containing the repression domain fused to the Ga14 DNA binding domain repressed transcription when directed to TATA-containing and TATA-less promoters. This repression domain was also found to functionally and directly interact with the TATA box binding protein.

With an incubation time of about 5 days, early diagnosis of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is critical to control the spread of the coronavirus disease 2019 (COVID-19) that killed more than 3 million people in its first 1.5 years. Here, we report on the modification of the dopant density and the phononic energy of antibody-coupled graphene when it interfaces with SARS-CoV-2 spike protein. This graphene chemeo-phononic system was able to detect SARS-CoV-2 spike protein at the limit of detection of ∼3.75 and ∼1 fg/mL in artificial saliva and phosphate-buffered saline, respectively. It also exhibited selectivity over proteins in saliva and MERS-CoV spike protein. Since the change in graphene phononics is monitored instead of the phononic signature of the analyte, this optical platform can be replicated for other COVID variants and specific-binding-based biodetection applications.

• Al 2 O 3 /TiO x based memristor devices for physical unclonable function. • Analyzing conduction mechanisms and origin of device-to-device variation for both states. • Randomness improvement by crossing operation with switching box. • Evaluating the performance of physical unclonable function based on memristor crossbar array with different response-extraction method. In this study, we evaluate the performance of a physical unclonable function (PUF) using Al 2 O 3 /TiO x based memristors. Through a conduction mechanism analysis, it is confirmed that the distribution of high-resistance state (HRS) is wider than that of low-resistance state (LRS) due to the direct-tunneling gap. Furthermore, cycle-to-cycle variation and random telegraph noise (RTN) characteristics which can affect the reliability of the PUF are also analyzed. Since the switching characteristics of the devices are less affected by temperature in the LRS thanks to ohmic conduction, the device state of a whole array is determined as the LRS for a better reliability. In addition, three kinds of response extraction methods are compared by evaluating the diffuseness and uniqueness with 2 × 2 switch block for the improved randomness. Finally, the reliability of the PUF is verified considering the measured conductance dependency on temperature and noise effect, and the bit error rates are compared between two states. [ABSTRACT FROM AUTHOR]

On-chip learning in neuromorphic systems, wherein both training and inference are performed on memristive synaptic devices, has been actively studied recently. However, on-chip learning is often affected by the weight-update linearity of memristive synaptic devices. Herein, we fabricated a Pt/Al 2 O 3 /TiO x /Ti/Pt stacked memristor device with excellent switching and reliability characteristics. Its weight-update linearity was analyzed via nonlinear A fitting through an on-chip simulation of the modified National Institute of Standards and Technology (MNIST) dataset. We confirmed the excellent recognition accuracy and low-fluctuation characteristics of the proposed model based on its similar characteristics to software learning. We obtained the perfect linear model and two types of nonlinear model characteristics of the memristor through incremental step pulse programming and performed an on-chip simulation. In addition, the characteristics of the measured cycle-to-cycle variation were reflected in the on-chip learning and were analyzed. We expect the low-fluctuation nonlinear model developed herein to be useful for on-chip learning owing to its excellent learning characteristics. • TiO x /Al 2 O 3 -based memristor for on-chip training of neuromorphic system • Analyzing the effect of nonlinear weight-update in memristive devices • Evaluating the performance of pattern recognition depending on linearity of weight-update • Low-fluctuation nonlinear model achieved by engineering training pulse [ABSTRACT FROM AUTHOR]

Recently, hardware-based neural network using memristive devices, so called neuromorphic system, has been extensively studied. Especially, on-chip (in situ) learning methods where training occurs inside hardware structure itself have been proposed and optimized based on memristor crossbar arrays regarding the linearity of weight-update characteristics. In this study, we analyze the effect of conductance overlap region of memristor on the recognition accuracy for on-chip learning simulation. The effect of conductance overlap region on recognition accuracy for modified national institute of standards and technology (MNIST) dataset is studied with an identical potentiation/depression pulse applied to Pt/Al 2 O 3 /TiO x /Ti/Pt stacked memristor. The overlap range can be varied by different pulse amplitude, and the training characteristics of memristive neural network is significantly dependent on the weight-update overlap region. • Ti/Pt/Al 2 O 3 /TiO x /Ti/Pt stacked memristor devices for neuromorphic system • Evaluating reliability including endurance and retention characteristics • Learning characteristics depending on pulse amplitude regarding linearity and weight overlap region • Analyzing effect of learning characteristics on recognition accuracy for on-chip training [ABSTRACT FROM AUTHOR]