Review of neuromorphic computing based on NAND flash memory.

The proliferation of data has facilitated global accessibility, which demands escalating amounts of power for data storage and processing purposes. In recent years, there has been a rise in research in the field of neuromorphic electronics, which draws inspiration from biological neurons and synapses. These electronics possess the ability to perform in-memory computing, which helps alleviate the limitations imposed by the 'von Neumann bottleneck' that exists between the memory and processor in the traditional von Neumann architecture. By leveraging their multi-bit non-volatility, characteristics that mimic biology, and Kirchhoff's law, neuromorphic electronics offer a promising solution to reduce the power consumption in processing vector-matrix multiplication tasks. Among all the existing nonvolatile memory technologies, NAND flash memory is one of the most competitive integrated solutions for the storage of large volumes of data. This work provides a comprehensive overview of the recent developments in neuromorphic computing based on NAND flash memory. Neuromorphic architectures using NAND flash memory for off-chip learning are presented with various quantization levels of input and weight. Next, neuromorphic architectures for on-chip learning are presented using standard backpropagation and feedback alignment algorithms. The array architecture, operation scheme, and electrical characteristics of NAND flash memory are discussed with a focus on the use of NAND flash memory in various neural network structures. Furthermore, the discrepancy of array architecture between on-chip learning and off-chip learning is addressed. This review article provides a foundation for understanding the neuromorphic computing based on the NAND flash memory and methods to utilize it based on application requirements.

Toward Ideal Low-Frequency Noise in Monolayer CVD MoS2 FETs: Influence of van der Waals Junctions and Sulfur Vacancy Management.

The pursuit of sub-1-nm field-effect transistor (FET) channels within 3D semiconducting crystals faces challenges due to diminished gate electrostatics and increased charge carrier scattering. 2D semiconductors, exemplified by transition metal dichalcogenides, provide a promising alternative. However, the non-idealities, such as excess low-frequency noise (LFN) in 2D FETs, present substantial hurdles to their realization and commercialization. In this study, ideal LFN characteristics in monolayer MoS2 FETs are attained by engineering the metal-2D semiconductor contact and the subgap density of states (DOS). By probing non-ideal contact resistance effects using CuS and Au electrodes, it is uncovered that excess contact noise in the high drain current (ID) region can be substantially reduced by forming a van der Waals junction with CuS electrodes. Furthermore, thermal annealing effectively mitigates sulfur vacancy-induced subgap density of states (DOS), diminishing excess noise in the low ID region. Through meticulous optimization of metal-2D semiconductor contacts and subgap DOS, alignment of 1/f noise with the pure carrier number fluctuation model is achieved, ultimately achieving the sought-after ideal LFN behavior in monolayer MoS2 FETs. This study underscores the necessity of refining excess noise, heralding improved performance and reliability of 2D electronic devices.

Massively parallel evaluation and computational prediction of the activities and specificities of 17 small Cas9s.

Recently, various small Cas9 orthologs and variants have been reported for use in in vivo delivery applications. Although small Cas9s are particularly suited for this purpose, selecting the most optimal small Cas9 for use at a specific target sequence continues to be challenging. Here, to this end, we have systematically compared the activities of 17 small Cas9s for thousands of target sequences. For each small Cas9, we have characterized the protospacer adjacent motif and determined optimal single guide RNA expression formats and scaffold sequence. High-throughput comparative analyses revealed distinct high- and low-activity groups of small Cas9s. We also developed DeepSmallCas9, a set of computational models predicting the activities of the small Cas9s at matched and mismatched target sequences. Together, this analysis and these computational models provide a useful guide for researchers to select the most suitable small Cas9 for specific applications.

Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic Devices.

Deep learning produces a remarkable performance in various applications such as image classification and speech recognition. However, state-of-the-art deep neural networks require a large number of weights and enormous computation power, which results in a bottleneck of efficiency for edge-device applications. To resolve these problems, deep spiking neural networks (DSNNs) have been proposed, given the specialized synapse and neuron hardware. In this work, the hardware neuromorphic system of DSNNs with gated Schottky diodes was investigated. Gated Schottky diodes have a near-linear conductance response, which can easily implement quantized weights in synaptic devices. Based on modeling of synaptic devices, two-layer fully connected neural networks are trained by off-chip learning. The adaptation of a neuron's threshold is proposed to reduce the accuracy degradation caused by the conversion from analog neural networks (ANNs) to event-driven DSNNs. Using left-justified rate coding as an input encoding method enables low-latency classification. The effect of device variation and noisy images to the classification accuracy is investigated. The time-to-first-spike (TTFS) scheme can significantly reduce power consumption by reducing the number of firing spikes compared to a max-firing scheme.

Predicting the efficiency of prime editing guide RNAs in human cells.

Prime editing enables the introduction of virtually any small-sized genetic change without requiring donor DNA or double-strand breaks. However, evaluation of prime editing efficiency requires time-consuming experiments, and the factors that affect efficiency have not been extensively investigated. In this study, we performed high-throughput evaluation of prime editor 2 (PE2) activities in human cells using 54,836 pairs of prime editing guide RNAs (pegRNAs) and their target sequences. The resulting data sets allowed us to identify factors affecting PE2 efficiency and to develop three computational models to predict pegRNA efficiency. For a given target sequence, the computational models predict efficiencies of pegRNAs with different lengths of primer binding sites and reverse transcriptase templates for edits of various types and positions. Testing the accuracy of the predictions using test data sets that were not used for training, we found Spearman's correlations between 0.47 and 0.81. Our computational models and information about factors affecting PE2 efficiency will facilitate practical application of prime editing.

Neuromorphic Computing Using NAND Flash Memory Architecture With Pulse Width Modulation Scheme.

A novel operation scheme is proposed for high-density and highly robust neuromorphic computing based on NAND flash memory architecture. Analog input is represented with time-encoded input pulse by pulse width modulation (PWM) circuit, and 4-bit synaptic weight is represented with adjustable conductance of NAND cells. Pulse width modulation scheme for analog input value and proposed operation scheme is suitably applicable to the conventional NAND flash architecture to implement a neuromorphic system without additional change of memory architecture. Saturated current-voltage characteristic of NAND cells eliminates the effect of serial resistance of adjacent cells where a pass bias is applied in a synaptic string and IR drop of metal wire resistance. Multiply-accumulate (MAC) operation of 4-bit weight and width-modulated input can be performed in a single input step without additional logic operation. Furthermore, the effect of quantization training (QT) on the classification accuracy is investigated compared with post-training quantization (PTQ) with 4-bit weight. Lastly, a sufficiently low current variance of NAND cells obtained by the read-verify-write (RVW) scheme achieves satisfying accuracies of 98.14 and 89.6% for the MNIST and CIFAR10 images, respectively.

On-Chip Training Spiking Neural Networks Using Approximated Backpropagation With Analog Synaptic Devices.

Hardware-based spiking neural networks (SNNs) inspired by a biological nervous system are regarded as an innovative computing system with very low power consumption and massively parallel operation. To train SNNs with supervision, we propose an efficient on-chip training scheme approximating backpropagation algorithm suitable for hardware implementation. We show that the accuracy of the proposed scheme for SNNs is close to that of conventional artificial neural networks (ANNs) by using the stochastic characteristics of neurons. In a hardware configuration, gated Schottky diodes (GSDs) are used as synaptic devices, which have a saturated current with respect to the input voltage. We design the SNN system by using the proposed on-chip training scheme with the GSDs, which can update their conductance in parallel to speed up the overall system. The performance of the on-chip training SNN system is validated through MNIST data set classification based on network size and total time step. The SNN systems achieve accuracy of 97.83% with 1 hidden layer and 98.44% with 4 hidden layers in fully connected neural networks. We then evaluate the effect of non-linearity and asymmetry of conductance response for long-term potentiation (LTP) and long-term depression (LTD) on the performance of the on-chip training SNN system. In addition, the impact of device variations on the performance of the on-chip training SNN system is evaluated.

Sequence-specific prediction of the efficiencies of adenine and cytosine base editors.

Base editors, including adenine base editors (ABEs)1 and cytosine base editors (CBEs)2,3, are widely used to induce point mutations. However, determining whether a specific nucleotide in its genomic context can be edited requires time-consuming experiments. Furthermore, when the editable window contains multiple target nucleotides, various genotypic products can be generated. To develop computational tools to predict base-editing efficiency and outcome product frequencies, we first evaluated the efficiencies of an ABE and a CBE and the outcome product frequencies at 13,504 and 14,157 target sequences, respectively, in human cells. We found that there were only modest asymmetric correlations between the activities of the base editors and Cas9 at the same targets. Using deep-learning-based computational modeling, we built tools to predict the efficiencies and outcome frequencies of ABE- and CBE-directed editing at any target sequence, with Pearson correlations ranging from 0.50 to 0.95. These tools and results will facilitate modeling and therapeutic correction of genetic diseases by base editing.

Pruning for Hardware-Based Deep Spiking Neural Networks Using Gated Schottky Diode as Synaptic Devices.

Deep learning represents state-of-the-art results in various machine learning tasks, but for applications that require real-time inference, the high computational cost of deep neural networks becomes a bottleneck for the efficiency. To overcome the high computational cost of deep neural networks, spiking neural networks (SNN) have been proposed. Herein, we propose a hardware implementation of the SNN with gated Schottky diodes as synaptic devices. In addition, we apply L1 regularization for connection pruning of the deep spiking neural networks using gated Schottky diodes as synap-tic devices. Applying L1 regularization eliminates the need for a re-training procedure because it prunes the weights based on the cost function. The compressed hardware-based SNN is energy efficient while achieving a classification accuracy of 97.85% which is comparable to 98.13% of the software deep neural networks (DNN).

Prediction of the sequence-specific cleavage activity of Cas9 variants.

Several Streptococcus pyogenes Cas9 (SpCas9) variants have been developed to improve an enzyme's specificity or to alter or broaden its protospacer-adjacent motif (PAM) compatibility, but selecting the optimal variant for a given target sequence and application remains difficult. To build computational models to predict the sequence-specific activity of 13 SpCas9 variants, we first assessed their cleavage efficiency at 26,891 target sequences. We found that, of the 256 possible four-nucleotide NNNN sequences, 156 can be used as a PAM by at least one of the SpCas9 variants. For the high-fidelity variants, overall activity could be ranked as SpCas9 ≥ Sniper-Cas9 > eSpCas9(1.1) > SpCas9-HF1 > HypaCas9 ≈ xCas9 >> evoCas9, whereas their overall specificities could be ranked as evoCas9 >> HypaCas9 ≥ SpCas9-HF1 ≈ eSpCas9(1.1) > xCas9 > Sniper-Cas9 > SpCas9. Using these data, we developed 16 deep-learning-based computational models that accurately predict the activity of these variants at any target sequence.

Effect of Word-Line Bias on Linearity of Multi-Level Conductance Steps for Multi-Layer Neural Networks Based on NAND Flash Cells.

NAND flash memory which is mature technology has great advantage in high density and great storage capacity per chip because cells are connected in series between a bit-line and a source-line. Therefore, NAND flash cell can be used as a synaptic device which is very useful for a high-density synaptic array. In this paper, the effect of the word-line bias on the linearity of multi-level conductance steps of the NAND flash cell is investigated. A 3-layer perceptron network (784×200×10) is trained by a suitable weight update method for NAND flash memory using MNIST data set. The linearity of multi-level conductance steps is improved as the word line bias increases from Vth -0.5 to Vth +1 at a fixed bit-line bias of 0.2 V. As a result, the learning accuracy is improved as the word-line bias increases from Vth -0.5 to Vth+1.

High-throughput analysis of the activities of xCas9, SpCas9-NG and SpCas9 at matched and mismatched target sequences in human cells.

The applications of clustered regularly interspaced short palindromic repeats (CRISPR)-based genome editing can be limited by a lack of compatible protospacer adjacent motifs (PAMs), insufficient on-target activity and off-target effects. Here, we report an extensive comparison of the PAM-sequence compatibilities and the on-target and off-target activities of Cas9 from Streptococcus pyogenes (SpCas9) and the SpCas9 variants xCas9 and SpCas9-NG (which are known to have broader PAM compatibility than SpCas9) at 26,478 lentivirally integrated target sequences and 78 endogenous target sites in human cells. We found that xCas9 has the lowest tolerance for mismatched target sequences and that SpCas9-NG has the broadest PAM compatibility. We also show, on the basis of newly identified non-NGG PAM sequences, that SpCas9-NG and SpCas9 can edit six previously unedited endogenous sites associated with genetic diseases. Moreover, we provide deep-learning models that predict the activities of xCas9 and SpCas9-NG at the target sequences. The resulting deeper understanding of the activities of xCas9, SpCas9-NG and SpCas9 in human cells should facilitate their use.

SpCas9 activity prediction by DeepSpCas9, a deep learning-based model with high generalization performance.

We evaluated SpCas9 activities at 12,832 target sequences using a high-throughput approach based on a human cell library containing single-guide RNA-encoding and target sequence pairs. Deep learning-based training on this large dataset of SpCas9-induced indel frequencies led to the development of a SpCas9 activity-predicting model named DeepSpCas9. When tested against independently generated datasets (our own and those published by other groups), DeepSpCas9 showed high generalization performance. DeepSpCas9 is available at http://deepcrispr.info/DeepSpCas9.

Gated Schottky Diode-Type Synaptic Device with a Field-Plate Structure to Reduce the Forward Current.

A gated Schottky diode with a field-plate structure is proposed and investigated as a new low-power synaptic device to suppress the forward current of the Schottky diode. In a hardware-based neural network, unwanted forward current can flow through gated Schottky diode-type synaptic devices during integration operations, possibly causing a malfunction of the neural network and increasing the power consumption. By adopting a field-plate structure, a virtual pn junction to suppress the forward current of the Schottky diode is formed in the poly-Si active layer. As a result, the unwanted forward current of the gated Schottky diode is successfully reduced to less than 1 pA/µm.

Emerging memory technologies for neuromorphic computing.

In this paper, we reviewed the recent trends on neuromorphic computing using emerging memory technologies. Two representative learning algorithms used to implement a hardware-based neural network are described as a bio-inspired learning algorithm and software-based learning algorithm, in particular back-propagation. The requirements of the synaptic device to apply each algorithm were analyzed. Then, we reviewed the research trends of synaptic devices to implement an artificial neural network.

Radial nerve palsy after arthroscopic anterior capsular release for degenerative elbow contracture.

Because the use of arthroscopy has increased recently for the treatment of elbow lesions, reports of complications have become more common. Nerve injury after arthroscopic anterior capsular release is an extremely rare complication, with 4 reported cases worldwide. We usually use a sharp-tipped electrocautery device with a 0.5-mm diameter during arthroscopic capsular release. In this case, because the former was not prepared, we used a ball-tipped electrocautery device with a 3-mm diameter. Herein, we experienced a case of radial nerve palsy after arthroscopic anterior capsular release using a ball-tipped electrocautery device on a degenerative elbow contracture. We supposed that the electrocautery device caused transiently thermal injury of the radial nerve despite proper portal entry site, intra-articular distension, and gentle arthroscopic manipulation. Elbow arthroscopy remains a technically difficult procedure with the potential for neurologic complications. To perform surgery safely, knowledge of the regional neuroanatomy and a thorough understanding of proper instrument usage are required.

Bone age delay patterns in Legg-Calvé-Perthes disease: an analysis using the Tanner and Whitehouse 3 method.

To study the bone age delay patterns in different stages of Perthes disease, 140 hand and corresponding hip radiographs in 83 patients were assessed. In the hand radiographs, the radius, ulna, metacarpals and phalanges (RUS) and carpal bone ages were calculated using the Tanner and Whitehouse 3 method and the Greulich and Pyle (G and P) bone age was assessed using the G and P atlas. From corresponding hip radiographs, the modified Elizabethtown stage was assessed. The RUS and carpal bone age as well as G and P bone age were found to lag behind the chronological age. The 95% confidence interval for the difference between RUS and G and P bone ages was 0.19 to 0.43 years and between carpal and G and P bone ages was -0.516 to -0.14 years, indicating a close agreement between the Tanner and Whitehouse 3 and G and P methods. The RUS bone age delay was maximum in stage Ia (2.00 +/- 1.08 years), whereas carpal delay was maximum in stage IIa (2.15 +/- 1.28 years). Bone maturation acceleration was observed in later stages of the disease as bone age tried to catch up with chronological age. Carpal delay was significantly greater than RUS delay from stage Ib to IIIb (P<0.05), but no significant difference was observed between carpal and RUS delays in stage IV (P=0.21), implying that bone maturation acceleration occurs in the RUS in the earlier stages, and carpal bone age tends to catch up with RUS bone age in the healed stage of the disease. The RUS and carpal bone age delays in stage I were significantly greater in severe (Catterall groups 3 and 4) disease than in mild (Catterall groups 1 and 2) disease. All patients in whom RUS or carpal bone age delay in stage I was greater than 2 years subsequently developed severe disease, indicating a positive correlation between bone age delay in stage I and subsequent extent of involvement of capital femoral epiphysis.

Genetic diversity of Acanthamoeba isolated from ocean sediments.

Genetic diversity of 18 Acanthamoeba isolates from ocean sediments was evaluated by comparing mitochondrial (mt) DNA RFLP, 18S rDNA sequences and by examining their cytopathic effects on human corneal epithelial cells versus reference strains. All isolates belonged to morphologic group II. Total of 16 restriction phenotypes of mtDNA from 18 isolates demonstrated the genetic diversity of Acanthamoeba in ocean sediments. Phylogenetic analysis using 18s rDNA sequences revealed that the 18 isolates were distinct from morphological groups I and III. Fifteen isolates showed close relatedness with 17 clinical isolates and A. castellanii Castellani and formed a lineage equivalent to T4 genotype of Byers group. Two reference strains from ocean sediment, A. hatchetti BH-2 and A. griffini S-7 clustered unequivocally with these 15 isolates. Diversity among isolates was also evident from their cytopathic effects on human corneal cells. This is the first time describing Acanthamoeba diversity in ocean sediments in Korea.

Antioxidant and cytotoxic activities of xanthones from Cudrania tricuspidata.

The new catecholic xanthone, 1,3,7-trihydroxy-4-(1,1-dimethyl-2-propenyl)-5,6-(2,2-dimethylchromeno)-xanthone (1), was isolated from the root bark of Cudrania tricuspidata together with seven known xanthones. The structures were fully characterized by analysis of physical and spectral (UV, IR, mass, and NMR) data. Relationships between the structural characteristics of xanthones and their antioxidant activities (DPPH, superoxide, and hydroxyl radical) were studied. Among the range of catecholic xanthones, 6,7-dihydroxyl xanthones (3-8) exhibited a strong scavenging effect on the DPPH radical. When one of the catecholic hydroxyl groups was protected as in compounds 1 and 2, DPPH radical scavenging activity was markedly decreased (IC(50)>200microM). DPPH activities were consistent with electrochemical response by cyclic voltammetry. Interestingly, compounds (1, 2) which had the weak activities on DPPH, exhibited both potent superoxide and hydroxyl radical scavenging activities. The strong activity on the hydroxyl radical of compounds (1, 2) could be rationalized by their chelating effect with iron (Fe(2+)) due to a redshift of its complex. The catecholic xanthones (3-8), being able to convert quinone methide intermediate, showed potent cytotoxicities against human cancer cell lines (HT-29, HL-60, SK-OV3, AGS, and A549). In particular, compounds 3, 6, and 7 had strong cytotoxic activities against AGS (LD(50)<5microM). DNA fragmentation patterns induced by catecholic xanthones revealed that tumor cell death was due to apoptosis.