Computer Memory Research Articles

Nested Neural Networks: A Novel Approach to Flexible and Deep Learning Architectures

In this paper, we introduce a novel neural network architecture termed Nested Neural Networks (NNN), which incorporates nested layers within a more complex overarching structure. The architecture leverages the concept of internal and external conditions, akin to linguistic structures like Future Perfect and Future Perfect Progressive in English grammar, to create a flexible and deep model. This architecture allows the network to handle complex data patterns with enhanced computational efficiency and memory savings. We demonstrate the effectiveness of NNNs through experiments on benchmark datasets, showcasing their ability to outperform traditional models in terms of accuracy, training efficiency, and adaptability.

Convergence Acceleration of Favre-Averaged Non-Linear Harmonic Method

Abstract This paper develops a numerical procedure to accelerate the convergence of the Favre-averaged non-linear harmonic (FNLH) method. The scheme provides a unified mathematical framework for solving the sparse linear systems formed by the mean flow and the time-linearized harmonic flows of FNLH in an explicit or implicit fashion. The approach explores the similarity of the sparse linear systems of FNLH and leads to a memory-efficient procedure, so that its memory consumption does not depend on the number of harmonics to compute. The proposed method has been implemented in the industrial computational fluid dynamics solver Hydra. Three test cases are used to conduct a comparative study of explicit and implicit schemes in terms of convergence, computational efficiency, and memory consumption. Comparisons show that the implicit scheme yields better convergence than the explicit scheme and is also roughly 7–10 times more computationally efficient than the explicit scheme with four levels of multigrid. Furthermore, the implicit scheme consumes only approximately 50% of the memory required by the explicit scheme with four levels of multigrid. Compared with the full-annulus unsteady Reynolds-averaged Navier–Stokes simulations, the implicit scheme produces comparable results to URANS with computational time and memory consumption that are two orders of magnitude smaller.

A state-space relational event modeling approach for learning dynamic social interaction behavior

Relational event models (REMs) are the primary choice for the analysis of relational-event network data. However, the standard REM assumes static parameters, which hinders the modeling of time-varying dynamics. This assumption might be too restrictive in real-life scenarios, making a model that allows for time-varying parameters more valuable. We introduce a state-space extension of the relational event model as a way to tackle this problem. The model has three main attributes. First, it provides a statistical framework of the temporal change of the parameters. Second, it enables the forecasting of future parameter values (which can be utilized to simulate new networks that can account for temporal dynamics in out-of-sample predictions). Third, it requires smaller data structures to be loaded into computer memory compared to the standard REM; this makes the model easily scalable to large networks. We conduct empirical analyses on bike-sharing data, corporate communications, and interactions among socio-political actors to illustrate model usage and applicability.

Custom RISC-V architecture incorporating memristive in-memory computing

Due to the rise in data-intensive applications, the von Neumann bottleneck is increasingly restricting modern computer architectures, resulting to latency and energy consumption. Addressing this challenge necessitates a CMOS-compatible solution with high energy efficiency and significant parallelism. Utilizing resistive switching components within a 1T1R crossbar array and the application of Stanford RRAM model, this paper suggests an original method for in-memory computing. Moreover, this work shows a new way to advance the popular RISC-V architecture by including memristive crossbar array. It does this by adding a custom instruction set, special hardware blocks, and the Scouting Logic Scheme. These modifications serve both as a comprehensive testbed for the memory system and a proof of concept for the future integration of memristors in computing architectures. The proposed design undergoes extensive testing and power analysis to validate its functionality and performance under various conditions. The results demonstrate significant improvements in computational efficiency and energy savings, highlighting the potential of memristor-based in-memory computing systems to overcome current architectural limitations.

A linear compensation method for inference accuracy improvement of memristive in-memory computing

Memristive computing system (MCS), with the feature of in-memory computing capability, for artificial neural networks (ANNs) deployment showing low power and massive parallelism, is a promising alternative for traditional Von-Neumann architecture computing system. However, because of the various non-idealities of both peripheral circuits and memristor array, the performance of the practical MCS tends to be significantly reduced. In this work, a linear compensation method (LCM) is proposed for the performance improvement of MCS under the effect of non-idealities. By considering the effects of various non-ideal states in the MCS as a whole, the output error of the MCS under different conditions is investigated. Then, a mathematic model for the output error is established based on the experimental data. Furthermore, the MCS is researched at the physical circuit level as well, in order to analyze the specific way in which the non-idealities affect the output current. Finally, based on the established mathematical model, the LCM output current is compensated in real time to improve the system performance. The effectiveness of LCM is verified and showing outstanding performance in the residual neural network-34 network architecture, which is easily affected by the non-idealities in hardware. The proposed LCM can be naturally integrated into the operation processes of MCS, paving the way for optimizing the deployment on generic ANN hardware based on the memristor.

Enhancing Security in IoT Devices with a Comprehensive Analysis of Lightweight Cryptographic Algorithms

A lot of gadgets connected to the Internet of Things (IoT) have changed many fields, from healthcare and smart houses to smart cities and commercial robotics. But this fast growth has also brought about big security problems, mostly because IoT devices don't have a lot of resources, which makes it hard to use standard cryptographic methods. The reason of this ponder is to grant a intensive examination of lightweight cryptographic strategies that can make IoT gadgets more secure without abating them down. Since IoT gadgets have uncommon security needs, lightweight cryptography has ended up one of the foremost critical ways to secure them. These methods are made to supply solid assurance whereas utilizing as few computer assets, power, and memory as conceivable. We see at well-known illustrations from each gather, like Display and Driven (square ciphers), Trivium and Grain (stream ciphers), and PHOTON and SPONGENT (hash capacities), to see how well they work in IoT settings. To donate a full picture of the contrasts, we carefully see at a few speed measures, such as running time, vitality utilize, memory utilization, and security against cryptographic dangers. The appraisal is based on both a common see at the subject and real-world tests on common IoT frameworks. We moreover conversation almost the trade-offs that come with choosing the correct cryptographic strategies to meet security needs whereas moreover taking under consideration the limits of IoT gadgets. Our inquire about appears that lightweight cryptographic calculations have a part of benefits, but the correct calculation must be chosen based on the IoT deployment's one of a kind utilize case and threat demonstrate. For example, applications that require a part of speed might advantage from lightweight stream ciphers, while applications that require to form beyond any doubt the security of the data might select lightweight hash capacities. The think about moreover talks approximately unused thoughts and bearings for lightweight cryptography for IoT, such as utilizing post-quantum cryptographic primitives to create IoT security future-proof. We see into the plausibility of blended strategies that use more than one lightweight procedure to progress security without utilizing as well numerous assets. Within the conclusion, this in-depth ponder appears how vital lightweight cryptographic strategies are for progressing IoT security. This paper aims to help researchers and practitioners set up good security measures for IoT devices by giving a thorough look at how well they work and what kinds of uses they are best for. This will make the IoT environment safer and more reliable.

A 28-nm 9T SRAM-based CIM macro with capacitance weighting module and redundant array-assisted ADC

In the emerging field of Computing-in-Memory (CIM), this study introduces a 28-nm CMOS-based Static Random Access Memory (SRAM) CIM macro capable of various computational modes, potentially offering a solution to the Von Neumann bottleneck. Beyond traditional SRAM read and write operations, to enhance the flexibility of the CIM macro, a 9T cell is proposed for performing AND, OR, and XNOR operations; a new capacitive weighting module is introduced to reduce the area of conventional ladder capacitors; and a redundant array-assisted Analog-to-Digital Converter (ADC) is proposed to improve linearity during ADC quantization. The proposed architecture can achieve multi-bit multiplication and accumulation (MAC), OR accumulation (ORA), and XNOR accumulation (XNORA). Simulated using a 28-nm CMOS process, the architecture demonstrated a minor standard deviation in BL voltage of 16.27 mV at the SS process corner, as evidenced by Monte Carlo simulation. At the TT process corner, the energy expenditure for MAC, XNORA, and ORA operations was 5.76, 5.85, and 5.82 fJ/op, respectively.

Organic Ferroelectrics for Regulation of Electronic and Ionic Transport Toward Neuromorphic Applications

AbstractOrganic ferroelectrics (OFEs) have been of significant research interest not only for nonvolatile memory applications but also for their unique material characteristics such as mechanical softness, biocompatibility, facile processibility, and chemically tailorable functionalities that inorganic counterparts are hard to achieve. Despite these promising merits, the utilization of OFEs has mainly focused on simply demonstrating flexible nonvolatile memories wherein modulation of electronic conductance is of interest. Recent studies indicate that the applicability of OFEs can be further extensive, particularly when combined with electronic, ionic, and mixed electronic‐ionic conducting media. Herein, we discuss that OFEs can be employed for the regulation of electronic as well as ionic charges, and lead to unique device behaviors. First, we comprehensively introduce organic ferroelectric materials classified with their structures and compositions. Next, we discuss recent studies where organic ferroelectricity has been incorporated with electronic, ionic, or mixed transport system to resolve issues in devices and endow multifunctionality, which are promising for neuromorphic computing and sensory memory systems. Finally, insight into the research direction of OFEs is provided, and what hurdles shall be overcome for real‐world applications.

Synthesis and Anisotropic Memristive Behavior of Borophene Nanosheets.

Neuromorphic computing, marked by its parallel computational abilities and low power usage, has become pivotal in advancing artificial intelligence. However, the advancement of neuromorphic computing has faced significant obstacles due to the performance limitations of traditional memory devices struggling with high power consumption and limited reliability. Two-dimensional (2D) materials have been extensively investigated as high-performance memristive materials, but they are often restricted by fixed memristive properties, which complicate circuit design and limit flexibility. Here, we report that multilayer borophene nanosheets represent a breakthrough material, displaying anisotropic variable memristive properties. The nanosheets, comprising semiconductor α'-4H-borophene sheets and metal β12-borophene sheets, have been synthesized on aluminum foil surface through chemical vapor deposition (CVD) method. The multilayer borophene nanosheets exhibit volatile memory behavior in the vertical direction and non-volatile memory behavior in the planar direction. This innovative class of 2D nanosheets not only overcomes the limitations of conventional memory devices but also expands the potential applications of borophene-based memories in information storage and in-memory computing.

A comprehensive review of model compression techniques in machine learning

This paper critically examines model compression techniques within the machine learning (ML) domain, emphasizing their role in enhancing model efficiency for deployment in resource-constrained environments, such as mobile devices, edge computing, and Internet of Things (IoT) systems. By systematically exploring compression techniques and lightweight design architectures, it is provided a comprehensive understanding of their operational contexts and effectiveness. The synthesis of these strategies reveals a dynamic interplay between model performance and computational demand, highlighting the balance required for optimal application. As machine learning (ML) models grow increasingly complex and data-intensive, the demand for computational resources and memory has surged accordingly. This escalation presents significant challenges for the deployment of artificial intelligence (AI) systems in real-world applications, particularly where hardware capabilities are limited. Therefore, model compression techniques are not merely advantageous but essential for ensuring that these models can be utilized across various domains, maintaining high performance without prohibitive resource requirements. Furthermore, this review underscores the importance of model compression in sustainable artificial intelligence (AI) development. The introduction of hybrid methods, which combine multiple compression techniques, promises to deliver superior performance and efficiency. Additionally, the development of intelligent frameworks capable of selecting the most appropriate compression strategy based on specific application needs is crucial for advancing the field. The practical examples and engineering applications discussed demonstrate the real-world impact of these techniques. By optimizing the balance between model complexity and computational efficiency, model compression ensures that the advancements in AI technology remain sustainable and widely applicable. This comprehensive review thus contributes to the academic discourse and guides innovative solutions for efficient and responsible machine learning practices, paving the way for future advancements in the field.Graphical abstract

Improving model robustness to weight noise via consistency regularization

Abstract As an emerging computing architecture, the computing-in-memory (CIM) exhibits significant potential for energy efficiency and computing power in artificial intelligence applications. However, the intrinsic non-idealities of CIM devices, manifesting as random interference on the weights of neural network, may significantly impact the inference accuracy. In this paper, we propose a novel training algorithm designed to mitigate the impact of weight noise. The algorithm strategically minimizes cross-entropy loss while concurrently refining the feature representations in intermediate layers to emulate those of an ideal, noise-free network. This dual-objective approach not only preserves the accuracy of the neural network but also enhances its robustness against noise-induced degradation. Empirical validation across several benchmark datasets confirms that our algorithm sets a new benchmark for accuracy in CIM-enabled neural network applications. Compared to the most commonly used forward noise training methods, our approach yields approximately a 2% accuracy boost on the ResNet32 model with the CIFAR-10 dataset and a weight noise scale of 0.2, and achieves a minimum performance gain of 1% on ResNet18 with the ImageNet dataset under the same noise quantization conditions.

When in-memory computing meets spiking neural networks—A perspective on device-circuit-system-and-algorithm co-design

This review explores the intersection of bio-plausible artificial intelligence in the form of spiking neural networks (SNNs) with the analog in-memory computing (IMC) domain, highlighting their collective potential for low-power edge computing environments. Through detailed investigation at the device, circuit, and system levels, we highlight the pivotal synergies between SNNs and IMC architectures. Additionally, we emphasize the critical need for comprehensive system-level analyses, considering the inter-dependencies among algorithms, devices, circuit, and system parameters, crucial for optimal performance. An in-depth analysis leads to the identification of key system-level bottlenecks arising from device limitations, which can be addressed using SNN-specific algorithm–hardware co-design techniques. This review underscores the imperative for holistic device to system design-space co-exploration, highlighting the critical aspects of hardware and algorithm research endeavors for low-power neuromorphic solutions.

Some Operations on Neutrosophic Hypersoft Matrices and Their Applications

This paper aims to extend the concept of Neutrosophic Hypersoft Matrix (NHSM) theory. NHSM is the matrix representation of a Neutrosophic Hypersoft Set (NHSS), where NHSS is the combination of a Neutrosophic set and a Hypersoft set. An NHSS can be stored in computer memory using the matrix notion, which is very useful and applicable. Based on NHSM, we provide some new notions (operations) such as NHS-sub-matrix, Equal NHSM, Null NHSM, Universal NHSM, Complement NHSM, NH-choice matrix (NHCM), product of NHCM and combined NHCM along with examples and characterizations. Additionally, we develop an NHSM algorithm using a value matrix, grace matrix, and mean matrix to solve a decision-making problem based on NHSSs. Finally, the NHSM algorithm was used to select the critically suffering chronic kidney disease (CKD) patient.

ADVANCING MEMORY DENSITY: A NOVEL DESIGN FOR MULTIPLE-BIT-PER-CELL PHASE CHANGE MEMORY

Multiple-bit-per-cell phase-change memory (MPCM) has emerged as a promising solution to address the escalating demands for high-density, low-power, and fast-access memory in modern computing and data storage systems. This paper presents a novel device design aimed at enabling multiple bits per cell in phase-change memory, thereby significantly enhancing memory density while maintaining performance and reliability. Leveraging innovative material compositions and advanced fabrication techniques, the proposed design demonstrates the potential to push the boundaries of memory capacity, efficiency, and scalability. Through comprehensive simulation analysis and performance evaluations, we showcase the feasibility and advantages of the new device design, highlighting its potential to revolutionize memory architectures and meet the evolving needs of next-generation computing systems.

Komodo Dragon Mlipir Algorithm-based CNN Model for Detection of Illegal Tree Cutting in Smart IoT Forest Area

Introduction: Trees and woods are vital to preventing climate change and protecting our planet. Sadly, they are constantly being destroyed due to human activities like deforestation, fires, etc. Method: This research presents and examines an outline for using audio event categorisation to automatically detect unlawful tree-cutting activity in forests. To monitor large swaths of forest, the research team proposes using ultra-low-power, minor devices incorporating edgecomputing microcontrollers and long-range wireless communication. An efficient and accurate audio classification solution based on multi-layer perceptron (MLP) and modified convolutional neural networks (M-CNN) is projected and tailored for cutting. The Komodo Dragon Mlipir Algorithm (KDMA) is used to pick the best weight for the CNN. Result: Compared to earlier efforts, the suggested system uses a computing technique to recognise deforestation-related hazards. Various preprocessing methods have been evaluated, with special attention paid to the trade-off between classification precision and computer resources, memory, and power use. Conclusion: Additionally, there have been long-range communication trials performed in natural settings. The experimental consequences demonstrate that the suggested method can notice and apprise tree-cutting occurrences through smart IoT for efficient and lucrative forest nursing.

Dual in-memory computing of matrix-vector multiplication for accelerating neural networks

Dual in-memory computing of matrix-vector multiplication for accelerating neural networks

Thermally induced oscillatory rarefied gas flow inside a rectangular cavity

Thermally induced oscillatory rarefied gas flow inside a two-dimensional rectangular cavity is investigated based on the hybrid macro-/mesoscopic scheme. The effects of the Knudsen (Kn) numbers and the oscillation frequency of lid temperature on the flow parameters are analyzed. The Shakhov model equation is solved numerically based on the mesoscopic approach in the near-wall region, and the macroscopic approach is adopted in the bulk flow region to reduce the computational cost. To close the numerical iteration procedure, the velocity distribution functions serving as the pseudo boundary between macroscopic and mesoscopic methods are reconstructed using the high-order Hermite polynomials. Numerical simulations demonstrate that the temperature profile at the central vertical of the cavity predicted by the hybrid method is in good agreement with results from the mesoscopic method, with a maximum error of 0.23%. In addition, the computational memory cost can be saved up to about 69.91%. The hybrid approach is able to capture the nonlinear phenomenon in the thermally induced oscillatory rarefied gas flow under high Kn numbers, where the horizontal velocity no longer obeys the law of periodic oscillating cosine function, and the rise time of the horizontal velocity is much longer than the fall time. The thickness of the viscous penetration layer and the disturbed region increases as the Kn number increases and decreases as the Strouhal number increases.

Deep-learning optimization using the gradient of a custom objective function: A full-waveform inversion example study on the convolutional objective function

The integration of conventional high-performance full-waveform inversion (FWI) algorithms with deep-learning frameworks is an innovative and promising research direction with the potential to enhance and broaden the application prospects of this field significantly. Automatic differentiation with backpropagation techniques can derive the gradients of model parameters in an equivalent manner to that of adjoint methods; however, it requires a substantial amount of computer memory, particularly when considering the time-step layers. In addition, some excellent objective functions suitable for FWI are not available in deep-learning frameworks. In comparison, the adjoint method, based on effective boundary storage technology, offers greater practicality for calculating the gradients of various objective functions. Therefore, this paper develops a novel approach toward FWI that integrates deep-learning optimization with high-performance gradient computation. In particular, our method inputs model parameter gradients from a custom objective function into the deep-learning framework. Herein, we use the acoustic equation with variable density as an example to demonstrate how a convolutional objective function, along with its corresponding velocity and density gradients, can be used for optimized inversion, multiscale inversion, and deep network parameterization-based multiscale inversion within the deep-learning framework. This approach provides a paradigm for deep-learning-optimized FWI, which we apply to synthetic and field data scenarios.

Numerical investigation of the mesh-free collocation approach for solving third kind VIEs with nonlinear vanishing delays

The principal purpose of this investigation is to present an approximate algorithm for solving third-kind Volterra integral equation with nonlinear vanishing delays using mesh-free techniques based on a radial basis function collocation scheme. This technique does not require background approximation cells, and the algorithm is efficient, has good stability and does not require much computer memory. A description of the approach for the proposed equations is provided. The error estimation of the proposed approach is also provided. The accuracy and reliability of the new scheme are demonstrated through numerical examples. Some of these examples have been compared with analytical solutions and moving least squares methods. The numerical results validate their agreement with theoretical error estimates.

Fast computation of the spectral correlation via frequency-averaging

Spectral Correlation (SC) has become a critical analytical instrument for assessing the second-order cyclostationarity (CS) in signals, especially for mechanical vibration signals. Nonetheless, the computational demand for computing SC is considerable. Built upon the Smoothed Cyclic Periodogram (SCP), a consistent estimator named Fast Smoothed Cyclic Periodogram (FastSCP) is proposed to address this issue. FastSCP represents the computational process of SCP through the convolution of spectral components and incorporates the Overlap-Save algorithm for accelerating. Comparison results with other fast SC estimators, including Fast Spectral Correlation (FastSC) and Faster Spectral Correlation (FasterSC), indicate that the proposed method is competitive in terms of both computational cost and memory usage, particularly when dealing with wide cyclic frequency ranges.