Stack Architecture Research Articles

Quality Assurance for Flexible Stack Assembly of Lithium‐Ion Cells

The development and scale‐up of lithium‐ion battery (LIB) production for a sustainable energy supply is advancing very rapidly and in versatile directions. Manufacturing processes and production steps are constantly developed and optimized to improve production efficiency. To integrate new machinery into a production line while conforming to the DIN EN ISO 9001 standards of certification, it is necessary to define the procedures for assuring product quality. Herein, a quality assurance concept is designed for an innovative flexible stacking process currently under development. Critical sources of errors are identified by employing the failure process matrix approach. On this basis, it is described how these errors are either avoided by suitable design of the machine components and its control or by which inspection technology is integrated to detect defective intermediate products. In the resulting quality assurance concept, a digital twin of the stacking machine is employed to constantly surveil process and product properties and also to facilitate the change of the produced products on the machine. Evaluating the quality management already in the design phase of the machine saves resources and enables a deeper understanding of the stacking process.

The Functional Machine Calculus

This paper presents the Functional Machine Calculus (FMC) as a simple model of higher-order computation with "reader/writer" effects: higher-order mutable store, input/output, and probabilistic and non-deterministic computation. The FMC derives from the lambda-calculus by taking the standard operational perspective of a call-by-name stack machine as primary, and introducing two natural generalizations. One, "locations", introduces multiple stacks, which each may represent an effect and so enable effect operators to be encoded into the abstraction and application constructs of the calculus. The second, "sequencing", is known from kappa-calculus and concatenative programming languages, and introduces the imperative notions of "skip" and "sequence". This enables the encoding of reduction strategies, including call-by-value lambda-calculus and monadic constructs. The encoding of effects into generalized abstraction and application means that standard results from the lambda-calculus may carry over to effects. The main result is confluence, which is possible because encoded effects reduce algebraically rather than operationally. Reduction generates the familiar algebraic laws for state, and unlike in the monadic setting, reader/writer effects combine seamlessly. A system of simple types confers termination of the machine.

Deep stacked pinball transfer matrix machine with its application in roller bearing fault diagnosis

At present, the matrix-based classification methods have played an extremely important role to mechanical fault diagnosis. However, the traditional matrix-based classification methods are mostly shallow classifiers, which are difficult to obtain the deep-seated sensitive features of vibration signals. To learn and extract low-rank structure information, a new deep stacked pinball transfer matrix machine (DSPTMM) is proposed to enhance the performance of traditional shallow matrix-based classifiers, which takes the stacking generalization principle as the main idea. In DSPTMM, a pinball transfer module (PTM) is constructed as the basic module of deep stacking network, in which the pinball loss is used to achieve an enhanced noise robustness. Specifically, PTM performs the function to obtain the weak prediction of the previous module, and the original results of previous weak prediction are modified by random projection and then will be input as a new feature set in the next module. Therefore, DSPTMM has a better classification performance and can modeled without enough annotation samples. Extensive experiments are carried out on two different datasets of roller bearing, and the results of experiment show that the proposed DSPTMM can use limited samples to establish the accurate model.

Accurate Dissolved Oxygen Prediction for Aquaculture Using Stacked Ensemble Machine Learning Model

Accurate Dissolved Oxygen Prediction for Aquaculture Using Stacked Ensemble Machine Learning Model

Deep stacked least square support matrix machine with adaptive multi-layer transfer for EEG classification

The recent extraordinary success of deep neural networks for electroencephalogram (EEG) decoding can be mainly attributed to the availability of large-scale labeled datasets. However, it requires prolonged EEG calibration, which is especially inconvenient for people with disabilities. Training deep neural networks with insufficient labeled EEG data will make the model suffer from limited generalization capability. In view of this, a new deep stacked network (DSN), called deep stacked transfer least square support matrix machine (DST-LSSMM), is proposed to alleviate the above issue. Following the philosophy of stacked generalization, the proposed DST-LSSMM utilizes least square support matrix machine (LSSMM) as its basic building unit and the random projection of previous layers as its stacking element. The adopted matrix classifier LSSMM directly models the matrix-form data, which is able to exploit the structural information between the columns or rows in EEG feature matrices to further improve the generalization capability of the model. Besides, we design an adaptive multi-layer model knowledge transfer learning scheme to guarantee consistency across the interrelated layers. The proposed adaptive multi-layer transfer scheme allows the use of model knowledge of previous related layers to facilitate the model construction at higher layers, consequently enhancing the generalization capability of the deep stacked model. DST-LSSMM is carried out in an efficient feed-forward way without parameter pre-training and fine-tuning, which can simplify the model optimization process. We extensively evaluate our method on three EEG datasets. Experimental results demonstrate that our method achieves promising performances on EEG classification.

Multivariate Time Series Sensor Feature Forecasting Using Deep Bidirectional LSTM

Multivariate Time series data forecasting (MTSF) is the assignment of forecasting future estimates of a particular series employing historic data. Lately, this work has enticed the focus of machine and deep learning researchers to tackle the complex and time consuming aspects of conventional forecasting techniques. Along with increasing access to historic documented data and the dire need of carrying out precise time-series feature forecasting. In this paper, we put forward a deep-learning(DL) technique proficient to tackle the setback of conventional forecasting techniques and display precise forecasting. The proposed technique is a Stacked Bidirectional long-short term memory architecture, which is an improvised version of existing Bidirectional LSTM in which multiple Bidirectional LSTM blocks are stacked, such that each layer contains multiple cells. In the direction of fair assessment, the functioning of the proposed technique is compared with existing LSTMs. Using varied evaluation measures, the obtained results show that the proposed Deep Bidirectional long-short term memory model exceeds standard approaches.

CNT-MXene ultralight membranes: fabrication, surface nano/microstructure, 2D-3D stacking architecture, ion-transport mechanism, and potential application as interlayers for Li-O2 batteries.

Multiwalled carbon nanotubes (MWCNTs) have shown effectiveness in improving the suitability of MXenes for energy-related applications. However, the ability of individually dispersed MWCNTs to control the structure of MXene-based macrostructures is unclear. Here, the correlation among composition, surface nano- and microstructure, MXenes' stacking order, structural swelling, and Li-ion transport mechanisms and properties in individually dispersed MWCNT-Ti3C2 films was investigated. The compact surface microstructure of MXene film, characterized by prominent wrinkles, is dramatically changed as MWCNTs occupy MXene/MXene edge interfaces. The 2D stacking order is preserved up to 30 wt% MWCNTs despite a significant swelling of ∼400%. Such alignment is completely disrupted at 40 wt%, and a more pronounced surface opening and internal expansion of ∼770% are realized. Both 30 wt% and 40 wt% membranes show stable cycling performance under a significantly higher current density due to faster transport channels. Notably, for the 3D membrane, the overpotential during repeated Li deposition/dissolution reactions is further reduced by ∼50%. Ion-transport mechanisms in the absence and presence of MWCNTs are discussed. Furthermore, ultralight yet continuous hybrid films comprising up to ∼0.027 mg cm-2 Ti3C2 can be prepared using aqueous colloidal dispersions and vacuum filtration for specific applications. The potential application of such ultralight membranes as interlayers for Li-O2 batteries is briefly examined.

Blockchain for the asset management industry

The financial asset management industry is constantly looking for ways to reduce the cost of running funds by simplifying their processes and making it easier to comply with regulatory requirements. Blockchain provides a feasible solution to these requirements. This paper explores the operational and implementation aspects of the business-grade version of the enterprise blockchain technology that benefits all stakeholders in the asset management industry using a theory-driven approach. Implementation technology, procedure, regulation, and social dimensions for practical implementation are discussed. Enterprise Ethereum Alliance's (EEA) Ethereum Architecture Stack describes the architectural framework in the technical implementation. This paper explains how EEA features give the asset management ecosystem a value proposition from a practical implementation perspective. Hence, stakeholders (regulators, asset management companies, registrars) and financial intermediaries get directions on migrating to the blockchain network.

Long-term reconstruction of satellite-based precipitation, soil moisture, and snow water equivalent in China

Abstract. A long-term high-resolution national dataset of precipitation (P), soil moisture (SM), and snow water equivalent (SWE) is necessary for predicting floods and droughts and assessing the impacts of climate change on streamflow in China. Current long-term daily or sub-daily datasets of P, SM, and SWE are limited by a coarse spatial resolution or the lack of local correction. Although SM and SWE data derived from hydrological simulations at a national scale have fine spatial resolutions and take advantage of local forcing data, hydrological models are not directly calibrated with SM and SWE data. In this study, we produced a daily 0.1∘ dataset of P, SM, and SWE in 1981–2017 across China, using global background data and local on-site data as forcing input and satellite-based data as reconstruction benchmarks. Global 0.1∘ and local 0.25∘P data in 1981–2017 are merged to reconstruct the historical P of the 0.1∘ China Merged Precipitation Analysis (CMPA) available in 2008–2017 using a stacking machine learning model. The reconstructed P data are used to drive the HBV hydrological model to simulate SM and SWE data in 1981–2017. The SM simulation is calibrated by Soil Moisture Active Passive Level 4 (SMAP-L4) data. The SWE simulation is calibrated by the national satellite-based snow depth dataset in China (Che and Dai, 2015) and the Moderate Resolution Imaging Spectroradiometer (MODIS) snow cover data. Cross-validated by the spatial and temporal splitting of the CMPA data, the median Kling–Gupta efficiency (KGE) of the reconstructed P is 0.68 for all grids at a daily scale. The median KGE of SM in calibration is 0.61 for all grids at a daily scale. For grids in two snow-rich regions, the median KGEs of SWE in calibration are 0.55 and −2.41 in the Songhua and Liaohe basins and the northwest continental basin respectively at a daily scale. Generally, the reconstruction dataset performs better in southern and eastern China than in northern and western China for P and SM and performs better in northeast China than in other regions for SWE. As the first long-term 0.1∘ daily dataset of P, SM, and SWE that combines information from local observations and satellite-based data benchmarks, this reconstruction product is valuable for future national investigations of hydrological processes.

High-Surface-Area Metalloporphyrin-Based Porous Ionic Polymers by the Direct Condensation Strategy for Enhanced CO2 Capture and Catalytic Conversion into Cyclic Carbonates.

Metalloporphyrin-based porous organic polymers (POPs) that behave as advanced biomimetic nanoreactors have drawn continuous attention for heterogeneous CO2 catalysis in the past decades. Inspired by the double activation model of epoxides, the design and synthesis of metalloporphyrin-based porous ionic polymers (PIPs) are considered as one of the most promising approaches for converting CO2 to cyclic carbonates under cocatalyst- and solvent-free conditions. To overcome the obstacle of poor reaction activity of ionic monomers or highly irregular stacking architecture, in this paper, we have proposed and demonstrated a modular bottom-up approach for constructing a series of high-surface-area metalloporphyrin-based PIPs in high yields by the direct condensation strategy, thus boosting the close contact of multiple active sites and achieving the enhanced CO2 capture and catalytic conversion into cyclic carbonates with high turnover frequencies under mild conditions. These recyclable aluminum-porphyrin-based PIPs are featured with high surface areas, prominent CO2 adsorptive capacities, rigid porphyrin skeletons, and flexible ionic pendants, as well as the matched amounts and spatial positions of metal centers and ionic sites, in which is demonstrated to be one of the quite competitive catalysts. Therefore, this strategy of introducing ionic components into the porphyrin frameworks as flexible side chains rather than main chains and adjusting the reactivity ratios of comonomers by structure-oriented methods, provides feasible guidance for the multifunctionalization of metalloporphyrin-based POPs, thereby increasing the accessibility of multiple active sites and improving their synergistic catalytic behavior.

Constructing high-efficiency orange-red thermally activated delayed fluorescence emitters by three-dimension molecular engineering

Preparing high-efficiency solution-processable orange-red thermally activated delayed fluorescence (TADF) emitters remains challenging. Herein, we design a series of emitters consisting of trinaphtho[3,3,3]propellane (TNP) core derivatized with different TADF units. Benefiting from the unique hexagonal stacking architecture of TNPs, TADF units are thus kept in the cavities between two TNPs, which decrease concentration quenching and annihilation of long-lived triplet excitons. According to the molecular engineering of TADF and host units, the excited states can further be regulated to effectively enhance spin-orbit coupling (SOC) processes. We observe a high-efficiency orange-red emission at 604 nm in one instance with high SOC value of 0.862 cm−1 and high photoluminescence quantum yield of 70.9%. Solution-processable organic light-emitting diodes exhibit a maximum external quantum efficiency of 24.74%. This study provides a universal strategy for designing high-performance TADF emitters through molecular packing and excited state regulation.

RegCPython: A Register-based Python Interpreter for Better Performance

Interpreters are widely used in the implementation of many programming languages, such as Python, Perl, and Java. Even though various JIT compilers emerge in an endless stream, interpretation efficiency still plays a critical role in program performance. Does a stack-based interpreter or a register-based interpreter perform better? The pros and cons of the pair of architectures have long been discussed. The stack architecture is attractive for its concise model and compact bytecode, but our study finds that the register-based interpreter can also be implemented easily and that its bytecode size only grows by a small margin. Moreover, the latter turns out to be appreciably faster. Specifically, we implemented an open source Python interpreter named RegCPython based on CPython v3.10.1. The former is register based, while the latter is stack based. Without changes in syntax, Application Programming Interface, and Application Binary Interface, RegCPython is excellently compatible with CPython, as it does not break existing syntax or interfaces. It achieves a speedup of 1.287 on the most favorable benchmark and 0.977 even on the most unfavorable benchmark. For all Python-intensive benchmarks, the average speedup reaches 1.120 on x86 and 1.130 on ARM. Our evaluation work, which also serves as an empirical study, provides a detailed performance survey of both interpreters on modern hardware. It points out that the register-based interpreters are more efficient mainly due to the elimination of machine instructions needed, while changes in branch mispredictions and cache misses have a limited impact on performance. Additionally, it confirms that the register-based implementation is also satisfactory in terms of memory footprint, compilation cost, and implementation complexity.

AlScN-Based Dual-Mode Devices With Temperature Compensated Strategy and Process Optimization

MEMS resonators with interdigital transducer (IDT) based on aluminum nitride (AlN) thin-film (a few micrometers thickness) materials gradually arouse researchers’ interest due to their ability to excite multiple modes of acoustic waves. However, the low electromechanical coupling coefficient is still one of the main reasons that limit the AlN-based resonators. In this article, we demonstrate a design journey for 10% scandium-doped AlN-based dual-mode acoustic wave resonators. Also, we investigated the influence of stack architectures with and without introduced AlN interlays, as well as the influence of stacked film thicknesses, on the performance of resonators. Simulation results show that changes in the SiO2 film thickness have different effects on the TCF of the two acoustic wave modes, providing a way for sensors to decouple two parameters that jointly affect the frequency drift. Moreover, measured results show that the introduced AlN interlayer (AlN-IL) induces a better crystalline orientation of the scandium-doped AlN (AlScN) thin film and a higher coupling coefficient of the leaky longitudinal (LL) mode ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${K}^{{2}}$ </tex-math></inline-formula> : from 4.19% to 4.96%), which make the resonators have great potential in surface acoustic wave (SAW)-based applications such as sensors and high-frequency applications.

Semi-Supervised Intrusion Detection Based on Stacking and Feature-Engineering to Handle Data Imbalance

Objectives: To design an architecture that can effectively handle the imbalance levels and complexities in the network data to provide qualitative predictions. Methods: Experiments were performed with KDD CUP 99 dataset, NSL- KDD dataset and UNSW- NB15 dataset. Comparisons were performed with SAVAERDNN model. Oversampling technique is used for data balancing, and the stacking architecture handles the issue of overtraining introduced due to oversampling. Findings: The proposed Stacking and Feature engineeringbased Semi-supervised (SFS) model presents a combined architecture that integrates data balancing, feature engineering and a stacking-based prediction model that balances data to reduce imbalance, reduces the data size, and also provides highly effective predictions. Results: indicate 2% increase in accuracy levels on the UNSW-NB15 dataset and 10% increase in accuracy levels in the NSL-KDD dataset. Novelty: The architecture has been designed in a domainspecific manner. Multiple intrusion detection datasets, each with different levels of imbalance, have been used to depict the generic nature of the SFS model. Keywords: Intrusion Detection; Data Imbalance; Stacking; Feature Engineering; Oversampling; Semi Supervised Learning

The molecular mechanisms of plasticity in crystal forms of theophylline

Plastic and elastic behaviors of organic crystals have profound influence on the processability of pharmaceutical substances. Analogous to metals, the identifications of molecular slip planes in organic crystals are regarded as a strategy for harnessing plasticity. In this work, we experimentally characterized the form II anhydrous theophylline (THPa) and its monohydrate (THPm) for their distinct plastic and elastic behaviors. Extensive DFT calculations were performed to model the effects of increasing lattice strains on molecular packing. We discovered that the energy barrier associated with the strain-induced molecular rearrangement would link to the plasticity of THPa, and possibly other simple aromatic compounds. Meanwhile, water molecules in THPm disrupt the stacking architecture from THPm and effectively undermine the general mechanism for plasticity. Hydrate formation would therefore be an alternative strategy to engineer the mechanical property of organic crystalline materials.

PainRhythms: Machine learning prediction of chronic pain from circadian dysregulation using actigraph data — a preliminary study

Chronic pain is currently diagnosed using verbal self-reports, which present a challenge for patients with cognitive or physiological disorders. Prior work has explored machine learning prediction of pain from clinical data, which requires active user involvement and does not capture their behavior in natural settings. Passive objective assessment is desirable. Circadian Rhythms, including sleep-wake cycles, are biological processes that reoccur every 24 h and can be derived from physiological data such as heart rate, activity, and sleep, gathered using widely-owned smart wearables. This study investigated the feasibility of using machine learning and rest-activity circadian rhythm features to predict patients’ pain, including pain intensity, its interference with the patient’s life (dysregulation), and their difficulty in performing physical functions using passively gathered actigraphy data. To predict pain on day N , actigraphy data collected over that day were analyzed. Three sets of feature were extracted: (1) Activity (total sedentary bouts/time/breaks, % in sedentary/light/ moderate activity), (2) Sleep (sleep efficiency/latency, wake after sleep onset), and (3) Rest Activity Rhythm (mesor, acrophase, Intradaily Variability (IV)). These features were then classified using various machine learning algorithms. Our proposed PainRhythms approach achieved an average AUC-ROC of 0.97 with a stacking machine learning classifier for predicting pain, 0.67 and 0.62 with logistic regression for pain intensity and interference, and 0.56 with gradient boosting for physical function. We found that chronic pain predictions were more accurate using rest-activity rhythm features than sleep or activity features. Of all the rhythmic features, Intradaily Variability (IV) was the most predictive feature, with elevated values in pain associated with disturbed sleep. PainRhythms provides preliminary evidence that rest-activity rhythms can effectively detect subjects with chronic pain. In future work, we aim to gather more data and confirm our preliminary findings on a large, class-balanced and diverse dataset. • PainRhythms provides preliminary evidence that rest-activity rhythms can effectively detect subjects with chronic pain. • Chronic pain predictions were more accurate using rest-activity rhythms than sleep or activity. • Intradaily Variability (IV) was the most predictive feature, with elevated values in pain associated with disturbed sleep. • Sleep disturbance was more associated with pain and pain intensity. • Besides sleep disturbance, features encompassing the subject’s physical activity and sleep quality also influenced pain interference and physical function.

Constructing 1 T-2 H TaS2 nanosheets with architecture and defect engineering for enhance hydrogen evolution reaction

Tailoring tantalum disulfide (TaS2) electrocatalysts with tunable stacking architecture and phase structure is of great importance for advancement of hydrogen evolution reaction (HER). In this work, stacking architecture- and phase- modulated tantalum disulfide nanosheets are reported through a facile liquid phase synthesis strategy. Accordion-like nanosheets, vertical nanosheets and ultrathin nanosheets are named according to their obvious stacking architecture, and 1 T-2 H mixed phase structures containing different degrees of 2 H phases are realized. High-resolution transmission electron microscopy shows that there are obvious lattice defects in the nanosheets. Tantalum disulfide nanosheets synergistically integrate multiple advantages of nanostructure engineering, phase engineering and defect engineering. In the HER process, the ultrathin tantalum disulfide nanosheet structure is conducive to the in-plane electron transfer, and its close contact with the electrode surface makes the charge transfer faster and has the best HER activity. The overpotential of the ultrathin nanosheet electrocatalyst at 10 mA cm−2 is 280 mV. After 10,000 CV cycles and over 40 h of i-t test, the catalyst still has high stability in acidic media. Theoretical calculations show that the addition of 2 H phase structure and rich defects improves the intrinsic activity and active sites, and jointly promotes the improvement of HER performance. Our work integrates a variety of optimal regulation methods into TaS2 nanosheets, providing a new idea for the construction of advanced electrocatalysts for water electrolysis.

Combining mid infrared spectroscopy with stacked generalisation machine learning for prediction of key soil properties

AbstractAccurate assessment of key soil attributes such as soil organic carbon (OC), available phosphorus (P), and available potassium (K) using mid‐infrared spectroscopy (MIRS) is essential for better soil management in precision agriculture. However, the calibration of the portable version of MIRS is more challenging than the benchmark technologies, hence, demanding more efficient modelling methods to provide accurate outcomes. This research aims to use the stacked generalisation machine learning (SG–ML) framework, combining support vector machine (SVM), gradient boosted regression (GBR), and random forest (RF), using linear ridge regression as a meta learner, for predicting OC, P, and K using MIR spectra of 375 soil samples collected from four farms (Flanders, Belgium). The performance of the SG–ML models was compared with the multilayer perceptron (MLP) deep learning (DL) method. Results showed the superiority of the SG–ML method over the corresponding single ML and DL models. The predictive performance of SG–ML using the validation set was excellent for the three soil attributes, with coefficient of determination (R2) and root mean square error (RMSE) values of 0.88% and 0.10%, 0.85 and 4.53 mg 100 g−1, and 0.84 and 3.87 mg 100 g−1 for OC, K, and P, respectively. The performance of DL models were good for OC (R2 = 0.65, and RMSE = 0.17%), poor for K (R2 = 0.58 and RMSE = 7.59 mg 100 g−1), and very poor for P (R2 = 0.46, and RMSE = 6.57 mg 100 g−1). The SG–ML reduced the prediction RMSE by 10% to 31%, compared with the single ML (SVM, RF, and GBR) models. In summary, the proposed stacking method is a powerful modelling tool for the accurate prediction of key soil attributes using portable MIRS.Highlights Machine learning (ML) and deep learning (DL) models were developed based on mid‐infrared soil spectra. The stacked generalisation ML (SG–ML) was compared with ML and DL for predicting OC, P, and K. SG–ML models outperformed ML (GBR, RF, and SVM) and deep learning (DL) models. The SG–ML models significantly decreased RMSE by 10% to 31% compared with the single ML models.

A Machine Learning Model for Adsorption Energies of Chemical Species Applied to CO2 Electroreduction

Machine learning methods are applied to obtain adsorption energies of different chemical species on (100), (111), and (211) FCC surfaces of several transition metals and Pb. Based on information available in databases containing adsorption energies obtained via first-principles calculations, we implemented MLPRegressor, XGBRegressor, Support Vector Regressor, and Stacking machine learning models. The fourth model is created from the combination of the previous three through a Stacking technique. In a broader context, our results showed the robustness of machine learning models and the ability of these methods to speed up the screening materials to specific goals, at a low computational cost. We emphasize the ability of our models to predict the adsorption energy for different systems. Due to their generality of them, we were able to make ion predictions on metallic surfaces, taking into account the influence of different functionals. This capability is of special significance due to the difficulty of calculating the correct energy for charged systems by traditional atomistic simulations. From then on, we made predictions for important chemical species in the CO2 electroreduction process, such as the radical anion CO2 −•, an important intermediary for obtaining new products in view of a negative carbon footprint.

Circuit-Level Memory Technologies and Applications based on 2D Materials.

Memory technologies and applications implemented fully or partially using emerging 2D materials have attracted increasing interest in the research community in recent years. Their unique characteristics provide new possibilities for highly integrated circuits with superior performances and low power consumption, as well as special functionalities. Here, an overview of progress in 2D-material-based memory technologies and applications on the circuit level is presented. In the material growth and fabrication aspects, the advantages and disadvantages of various methods for producing large-scale 2D memory devices are discussed. Reports on 2D-material-based integrated memory circuits, from conventional dynamic random-access memory, static random-access memory, and flash memory arrays, to emerging memristive crossbar structures, all the way to 3D monolithic stacking architecture, are systematically reviewed. Comparisons between experimental implementations and theoretical estimations for different integration architectures are given in terms of the critical parameters in 2D memory devices. Attempts to use 2D memory arrays for in-memory computing applications, mostly on logic-in-memory and neuromorphic computing, are summarized here. Finally, challenges that impede the large-scale applications of 2D-material-based memory are reviewed, and perspectives on possible approaches toward a more reliable system-level fabrication are also given, hopefully shedding some light on future research.