Computational Network Research Articles

A hybrid CMOS-memristor based adaptable Bidirectional Associative Memory neural network for pattern recognition applications

Abstract This research presents a circuit-level hybrid CMOS memristor architecture for constructing Bidirectional Associative Memory (BAM). Initially, a synaptic circuit structure was built by employing a voltage threshold memristor in a crossbar architecture. This synaptic structure is adaptable and flexible for generating a wide range of synaptic weights. It is then deployed in the BAM network to perform an associative function. To aid in better name recall, this BAM network has been trained to associate Greek and mathematical symbols with their first letters in English, and vice versa. The designed circuit was validated using MATLAB and the EDA (Electronic Design Automation) Tool: Cadence Virtuoso. The addition of noise further evaluates the performance of the BAM network. When tested with noise levels of 10%, 20%, and 30%, the input patterns were retrieved at 100% in both directions. Furthermore, the proposed synaptic circuit is validated for variations in RON, ROFF and its performance is compared with other memristor models. It is also found that the average power consumption of the proposed synaptic circuit is 1.22mW. These results, which were experimentally confirmed, demonstrate the precision and noise isolation of the proposed BAM design. With appropriate tuning of memristor, the synaptic weights can be mapped easily with the memristor conductance value. This circuit can be effectively used in the field of image processing, neural network, and neuromorphic computation, which helps to associate and restore original or damaged binary images, showing strong robustness and accuracy.

Highly Stable Electronics Based on β-Ga2O3 for Advanced Memory Applications.

Wide-bandgap (WBG) semiconductors are at the forefront of driving innovations in electronic technology, perpetuating Moore's Law and opening up new avenues for electronic devices. Although β-Ga2O3 has attracted extensive research interest in advanced electronics, its high-temperature and high-speed volatile memory applications in harsh environment has been largely overlooked. Herein, a high-performance hexagonal boron nitride (h-BN)/β-Ga2O3 heterostructure junction field-effect transistor (HJFET) is fabricated, exhibiting an off-state current as low as ≈10 fA, a high on/off current ratio of ≈108, a low contact resistance of 5.6 Ω·mm, and an impressive field-effect electron mobility of 156 cm2 （Vs）-1. Notably, the current h-BN/β-Ga2O3 HJFET exhibits outstanding thermal reliability in the ultra-wide temperature range from 223 to 573 K, as well as long-term environmental stability in air, which confirms its inherent capability of operation in harsh environments. Moreover, the h-BN/β-Ga2O3 HJFET demonstrates successful applications for accelerator-in-memory computing fields, including dynamic random-access memory structure and neural network computations. These superior characteristics position β-Ga₂O₃-based electronics as highly promising for applications in extreme environments, with particular relevance to the automotive, aerospace, and sensor sectors.

Dendritic growth and synaptic organization from activity-independent cues and local activity-dependent plasticity.

Dendritic branching and synaptic organization shape single-neuron and network computations. How they emerge simultaneously during brain development as neurons become integrated into functional networks is still not mechanistically understood. Here, we propose a mechanistic model in which dendrite growth and the organization of synapses arise from the interaction of activity-independent cues from potential synaptic partners and local activity-dependent synaptic plasticity. Consistent with experiments, three phases of dendritic growth - overshoot, pruning, and stabilization - emerge naturally in the model. The model generates stellate-like dendritic morphologies that capture several morphological features of biological neurons under normal and perturbed learning rules, reflecting biological variability. Model-generated dendrites have approximately optimal wiring length consistent with experimental measurements. In addition to establishing dendritic morphologies, activity-dependent plasticity rules organize synapses into spatial clusters according to the correlated activity they experience. We demonstrate that a trade-off between activity-dependent and -independent factors influences dendritic growth and synaptic location throughout development, suggesting that early developmental variability can affect mature morphology and synaptic function. Therefore, a single mechanistic model can capture dendritic growth and account for the synaptic organization of correlated inputs during development. Our work suggests concrete mechanistic components underlying the emergence of dendritic morphologies and synaptic formation and removal in function and dysfunction, and provides experimentally testable predictions for the role of individual components.

Dendritic growth and synaptic organization from activity-independent cues and local activity-dependent plasticity

Dendritic branching and synaptic organization shape single-neuron and network computations. How they emerge simultaneously during brain development as neurons become integrated into functional networks is still not mechanistically understood. Here, we propose a mechanistic model in which dendrite growth and the organization of synapses arise from the interaction of activity-independent cues from potential synaptic partners and local activity-dependent synaptic plasticity. Consistent with experiments, three phases of dendritic growth – overshoot, pruning, and stabilization – emerge naturally in the model. The model generates stellate-like dendritic morphologies that capture several morphological features of biological neurons under normal and perturbed learning rules, reflecting biological variability. Model-generated dendrites have approximately optimal wiring length consistent with experimental measurements. In addition to establishing dendritic morphologies, activity-dependent plasticity rules organize synapses into spatial clusters according to the correlated activity they experience. We demonstrate that a trade-off between activity-dependent and -independent factors influences dendritic growth and synaptic location throughout development, suggesting that early developmental variability can affect mature morphology and synaptic function. Therefore, a single mechanistic model can capture dendritic growth and account for the synaptic organization of correlated inputs during development. Our work suggests concrete mechanistic components underlying the emergence of dendritic morphologies and synaptic formation and removal in function and dysfunction, and provides experimentally testable predictions for the role of individual components.

First-order Euler method is effective for computation of associative-memory network of Kuramoto oscillators

First-order Euler method is effective for computation of associative-memory network of Kuramoto oscillators

Solving linear algebraic equations with limited computational power and network bandwidth

Solving linear algebraic equations with limited computational power and network bandwidth

Programming universal unitary transformations on a general-purpose silicon photonic platform

General-purpose programmable photonic processors provide a versatile platform for integrating diverse functionalities on a single chip. Leveraging a two-dimensional hexagonal waveguide mesh of Mach–Zehnder interferometers, these systems have demonstrated significant potential in microwave photonic applications. Additionally, they are a promising platform for creating unitary linear transformations, which are key elements in quantum computing and photonic neural networks. However, a general procedure for implementing these transformations on such systems has not been established yet. This work demonstrates the programming of universal unitary transformations on a general-purpose programmable photonic circuit with a hexagonal topology. We detail the steps to split the light on-chip, demonstrate that an equivalent structure to the Mach–Zehnder interferometer with one internal and one external phase shifter can be built in the hexagonal mesh, and program both the triangular and rectangular architectures for matrix multiplication. We recalibrate the system to account for passive phase deviations. Experimental programming of 3 × 3 and 4 × 4 random unitary matrices yields fidelities >98% and bit precisions over five bits. To the best of our knowledge, this is the first time that random unitary matrices are demonstrated on a general-purpose photonic processor and pave the way for the implementation of programmable photonic circuits in optical computing and signal processing systems.

Three-dimensional design, simulation and optimization of a centrifugal compressor impeller with double-splitter blades.

After the emergence of turbo-machines in the industrial field, researchers have always been seeking to increase their efficiency. Centrifugal compressors are widely used in industries, hence understanding their flow and attempting to increase their efficiency has always been a focus. In recent years, using two splitter blades alongside the main blades has been one of the novel methods used to improve the performance of centrifugal compressors. In this study, after validating the NASA-CC3 centrifugal compressor, a centrifugal compressor design with two splitter blades was carried out. Then, by defining eight variables, an attempt was made to optimize the design. The genetic algorithm was used as the optimization algorithm in this process. Since the optimization process only using genetic algorithms is very time-consuming and has high computational costs, artificial neural networks were used to reduce costs. The objective function in this optimization process was to increase efficiency while maintaining the flow rate and pressure ratio at the design point. After the optimization process was completed, the performance curve of the optimized compressor was compared with the NASA-CC3 compressor at and outside the design point, and its advantages and disadvantages were stated. At the end of the above optimization process, the efficiency of the centrifugal compressor increased by 1.06% at the design point. To investigate the effect of the number of computational networks on the independence of results, further analysis is required.

Tracing humanity’s path: the DISPERSALS project and the origins of global migration

Tracing humanity’s path: the DISPERSALS project and the origins of global migration DISPERSALS will compare human occupation and ecology between central Mozambique and eastern and southern Africa using a multi-scale approach based on the study of regional diachronic cultural traits. It will reconstruct regional population patterns, followed by comparative quantitative population genetics combined with GIS computational network analyses. The results will then be integrated through agent-based modelling, based on the incremental creation, elimination, or reorientation of network links to simulate a quantitative framework to study the evolution of population dispersal across southern-eastern Africa. The project will be crucial in providing ground-breaking high-resolution archaeological, chronological and paleoenvironmental data. DISPERSALS will deliver a fundamental perspective on the key processes that triggered migrations and dispersals within Africa and out-of-Africa, which ultimately resulted in the human diaspora over the entire planet.

Uncovering The Pharmacological Mechanism of Ficus elastica as Anti-hyperlipidemia Candidate: LC-HRMS, Network Pharmacology, In vitro and In vivo Studies

Hyperlipidemia is a major risk factor for cardiovascular diseases. While conventional treatments exist, there is a growing interest in natural remedies with fewer side effects. Ficus elastica has promising medicinal properties, yet its potential as an anti-hyperlipidemic agent remains unexplored. This study aimed to investigate the anti-hyperlipidemic effects of F. elastica using an integrated approach of LC-HRMS-based chemical bioinformatics and in vitro/in vivo experimental validation. The anti-hyperlipidemic potential of F. elastica and its mechanism of action were screened using integrative computational network pharmacology followed by in vitro HMG-CoA reductase inhibition and in vivo lipid-lowering activity in a hyperlipidemia rat model. Network pharmacology analysis identified STAT3, HSP90AA1, and TLR4 as potential core targets involved in lipid and atherosclerosis-related KEGG pathways. Molecular docking simulations revealed high-affinity interactions between F. elastica compounds and the identified targets, notably compound 41 and compound 61. In vitro assay demonstrated that ethanolic extract of F. elastica inhibited HMG-CoA reductase with an IC50 of 297.73 µg/mL. In vivo experiment using a hyperlipidemic rat model showed significant reductions in total cholesterol, triglycerides, and increased HDL levels. The reduction of triglycerides and elevation of HDL level after F. elastica ethanolic extract supplementation is similar to the effect from supplementation of simvastatin. These findings suggest that F. elastica ethanolic extract possesses notable anti-hyperlipidemic properties, likely mediated through multiple molecular targets and pathways. The study highlights the potential of F. elastica ethanolic extract as a promising candidate for anti-hyperlipidemic therapy and underscores the efficacy of integrating computational and experimental approaches in natural product research.

Comparing Traversal Strategies: Depth-first Search vs. Breadth-first Search in Complex Networks

This article compares and contrasts two basic graph traversal algorithms that are commonly employed in computational problem-solving and network research. Common applications of these algorithms include pathfinding, optimisation of network flows, collaborative exploration, and classification tasks. To find out how well they function with different types of datasets, network topologies, and issue domains, researchers have systematically reviewed previous works. We measured the efficiency of each solution using performance indicators like execution time, memory utilisation, and path length. According to the results, one approach is more effective in memory-constrained settings and deep searches, while the other is better at discovering the shortest paths and providing comprehensive coverage. Furthermore, the paper emphasises the advantages of hybrid techniques, which merge the best features of both algorithms to provide better results in specific cases. This comparison helps fill gaps in our knowledge of graph-based problem-solving methods and sheds light on how to choose the best traversal algorithms for different types of applications.

A Hybrid Framework for Securing 5G-Enabled Healthcare Systems

The rapid adoption of 5G technology in healthcare introduces significant challenges regarding data privacy and security. This paper proposes a hybrid framework integrating blockchain, zero-trust architecture (ZTA), and AI-driven threat detection to address these challenges. Blockchain ensures secure, tamper-proof data storage, while ZTA strengthens access control by continuously verifying users and devices. AI contributes by providing real-time threat detection and dynamic response capabilities, making the system more resilient to evolving cyber risks. A systematic literature review was conducted to analyze existing frameworks and identify gaps in 5G healthcare security. The findings reveal that while individual technologies such as blockchain and ZTA are well-established, their integration into a cohesive framework remains underexplored. The proposed hybrid solution effectively mitigates the risks associated with 5G networks by offering a multi-layered security approach. This research contributes to the field by proposing a scalable, adaptable security model suitable for 5G-enabled healthcare systems. Future research should focus on empirical validation, scalability testing, and exploring lightweight alternatives to blockchain and AI for resource-constrained environments. Additionally, investigating the integration of emerging technologies like quantum computing and 6G networks will further enhance the framework’s security capabilities. This study provides a foundation for developing secure, privacy-preserving systems for healthcare in the 5G era.

Low-power artificial neuron networks with enhanced synaptic functionality using dual transistor and dual memristor.

Artificial neurons with bio-inspired firing patterns have the potential to significantly improve the performance of neural network computing. The most significant component of an artificial neuron circuit is a large amount of energy consumption. Recent literature has proposed memristors as a promising option for synaptic implementation. In contrast, implementing memristive circuitry through neuron hardware presents significant challenges and is a relevant research topic. This paper describes an efficient circuit-level mixed CMOS memristor artificial neuron network with a memristor synapse model. From this perspective, the paper describes the design of artificial neurons in standard CMOS technology with low power utilization. The neuron circuit response is a modified version of the Morris-Lecar theoretical model. The suggested circuit employs memristor-based artificial neurons with Dual Transistor and Dual Memristor (DTDM) synapse circuit. The proposed neuron network produces a high spiking frequency and low power consumption. According to our research, a memristor-based Morris Lecar (ML) neuron with a DTDM synapse circuit consumes 12.55 pW of power, the spiking frequency is 22.72 kHz, and 2.13 fJ of energy per spike. The simulations were carried out using the Spectre tool with 45 nm CMOS technology.

Research on Unmanned Aerial Vehicle Intelligent Maneuvering Method Based on Hierarchical Proximal Policy Optimization

Improving decision-making in the autonomous maneuvering of unmanned aerial vehicles (UAVs) is of great significance to improving flight safety, the mission execution rate, and environmental adaptability. The method of deep reinforcement learning makes the autonomous maneuvering decision of UAVs possible. However, the current algorithm is prone to low training efficiency and poor performance when dealing with complex continuous maneuvering problems. In order to further improve the autonomous maneuvering level of UAVs and explore safe and efficient maneuvering methods in complex environments, a maneuvering decision-making method based on hierarchical reinforcement learning and ‌Proximal Policy Optimization (PPO) is proposed in this paper. By introducing the idea of hierarchical reinforcement learning into the PPO algorithm, the complex problem of UAV maneuvering and obstacle avoidance is separated into high-level macro-maneuver guidance and low-level micro-action execution, greatly simplifying the task of addressing complex maneuvering decisions using a single-layer PPO. In addition, by designing static/dynamic threat zones and varying their quantity, size, and location, the complexity of the environment is enhanced, thereby improving the algorithm’s adaptability and robustness to different conditions. The experimental results indicate that when the number of threat targets is five, the success rate of the H-PPO algorithm for maneuvering to the designated target point is 80%, which is significantly higher than the 58% rate achieved by the original PPO algorithm. Additionally, both the average maneuvering distance and time are lower than those of the PPO, and the network computation time is only 1.64 s, which is shorter than the 2.46 s computation time of the PPO. Additionally, as the complexity of the environment increases, the H-PPO algorithm outperforms other compared networks, demonstrating the effectiveness of the algorithm constructed in this paper for guiding intelligent agents to autonomously maneuver and avoid obstacles in complex and time-varying environments. This provides a feasible technical approach and theoretical support for realizing autonomous maneuvering decisions in UAVs.

Unveiling the role of miRNAs in Diminished Ovarian Reserve: an in silico network approach

MicroRNAs (miRNAs) have acquired an increased recognition to unravel the complex molecular mechanisms underlying Diminished Ovarian Reserve (DOR), one of the main responsible for infertility. To investigate the impact of miRNA profiles in granulosa cells and follicular fluid, crucial players in follicle development, this study employed a computational network theory approach to reconstruct potential pathways regulated by miRNAs in granulosa cells and follicular fluid of women suffering from DOR. Available data from published research were collected to create the FGC_MiRNome_MC, a representation of miRNA target genes and their interactions. 365 hubs were identified within the network, representing potential key regulators, and 210 nodes that act as both hubs and bottlenecks (H&BN nodes), suggesting that they may control the information flow within the network. GO enrichment analysis of the 210 H&BN nodes revealed their involvement in fundamental cellular processes relevant to ovarian function. In particular, the cluster analysis identified several shared pathways between cluster 1 and cluster 2 involved in the RAS/MAPK pathway, which plays a critical role in cell proliferation, differentiation and survival. These findings suggest that miRNAs play a significant role in DOR and highlight the potential of the RAS/MAPK pathway as a target for further investigation. Additionally, the genes identified as both hubs and bottlenecks revealed interesting connections to reproductive health in KO mice models. This in silico approach provides valuable insights into potential biomarkers and therapeutic targets for age-related reproductive disorders.

Neural Network Analysis of Age-Related Changes in the Biosystem of the Intestinal Microbiota of Broiler Chickens

The study aims to evaluate age-related changes and mass accumulation activity in the biosystem of the intestinal microbiota of broiler chickens using fractal frequency-taxonomic profiles and neural network analysis. The molecular genetic terminal restriction fragment length polymorphism method and polymerase chain reaction were used to analyze the frequency-taxonomic profiles of the intestinal microbiota. These profiles were processed using the NONN computational neural network and trained to correctly convert digital profile data into age-related activity index (CSIfract) and mass accumulation activity index (CSImass). The visualization of the age dependence of mass accumulation activity was carried out using a two-dimensional Scale matrix. With the increasing age of broilers, a decrease was observed in the biochemical activity of their intestinal microbiota, associated with the accumulation of functional disorders in the microbial biosystem. Mass accumulative activity reached a maximum at 2-3 weeks of development and then gradually decreased. The quantitative composition of microorganisms changes with age: the number of lactobacilli and bacilli decreased, accompanied by increased clostridium and other microorganisms that increase the risk of diseases in populations. The study showed that age-related changes in broilers' intestinal microbiota significantly affected their biochemical activity and disease risk. To maintain optimal microbial activity, using probiotic supplements in feed is recommended. Keywords: Fractal profile, age-related activity and mass accumulation activity of the biosystem of microorganisms, computational neural network.

Noncollinear phase of the antiferromagnetic sawtooth chain

The antiferromagnetic sawtooth chain is a prototypical example of a frustrated spin system with vertex-sharing triangles, giving rise to complex quantum states. Depending on the interaction parameters, this system has three phases, of which the gapless noncollinear phase (for strongly coupled basal spins and loosely attached apical spins) has received little theoretical attention so far. In this work, we comprehensively investigate the properties of the noncollinear phase using large-scale tensor network computations which exploit the full SU(2) symmetry of the underlying Heisenberg model. We study the ground state both for finite systems using the density-matrix renormalization group (DMRG) as well as for infinite chains via the variational uniform matrix-product state (VUMPS) formalism. Finite temperatures and correlation functions are tackled via imaginary or real time evolutions, which we implement using the time-dependent variational principle (TDVP). We find that the noncollinear phase is characterized by a low-momentum peak and a diffuse tail for the apex-apex correlations. Deep into the phase, the pattern sharpens into a peak indicating a 90∘ spiral. The apical spins are soft and highly susceptible to external perturbations; they give rise to a large number of gapless magnetic states that are polarized by weak fields and cause a long low-temperature tail in the specific heat. The dynamic spin-structure factor exhibits additive contributions from a two-spinon continuum (excitations of the basal chain) and a gapless peak at k=π/2 (excitations of the apical spins). Small temperatures excite the gapless states and smear the spectral weight of the k=π/2 peak out into a homogeneous flat-band structure. Our results are relevant, e.g., for the material atacamite Cu2Cl(OH)3 in high magnetic fields. Published by the American Physical Society 2025

A NiAl-layered double hydroxides memristor with artificial synapse function and its Boolean logic applications.

In the era of artificial intelligence, there has been a rise in novel computing methods due to the increased demand for rapid and effective data processing. It is of great significance to develop memristor devices capable of emulating the computational neural network of the brain, especially in the realm of artificial intelligence applications. In this work, a memristor based on NiAl-layered double hydroxides is presented with excellent electrical performance, including analog resistive conversion characteristics and the effect of multi-level conductivity modulation. In addition, the device's conductance can be continuously adjusted by varying pulse width, interval, and amplitude. The successful replication of synaptic features has been achieved. In order to implement the functions of "NOT," "AND," and "OR," a logic gate is constructed using two synaptic devices. The confirmation of the potential use of synaptic devices in brain-like computing was demonstrated. In addition, it demonstrates the potential of these devices in supporting computing models beyond von Neumann architecture.

Biocomputing at the crossroad between emulating artificial intelligence and cellular supremacy.

Biocomputation aims to create sophisticated biological systems capable of addressing important problems in (bio)medicine with a machine-like precision. At present, computational gene networks engineered by single- or multi-layered assembly of DNA-, RNA- and protein-level gene switches have allowed bacterial or mammalian cells to perform various regulation logics of interest, including Boolean calculation or neural network-like computing. This review highlights the molecular building blocks, design principles, and computational tasks demonstrated by current biocomputers, before briefly discussing possible fields where biological computers may ultimately outcompete their electronic counterparts and achieve cellular supremacy.

Silicon photonic convolution operator exploiting on-chip nonlinear activation function.

Nonlinear activation functions (NAFs) are essential in artificial neural networks, enhancing learning capabilities by capturing complex input-output relationships. However, most NAF implementations rely on additional optoelectronic devices or digital computers, reducing the benefits of optical computing. To address this, we propose what we believe to be the first implementation of a nonlinear modulation process using an electro-optic IQ modulator on a silicon photonic convolution operator chip as a novel NAF. We validated this operator by constructing a convolutional neural network for radio machine learning classification, achieving 92.5% accuracy-an improvement of 27% over the case without a NAF. Compared with optoelectronic systems that rely on separate components, this fully integrated silicon photonic chip allows the NAF to execute nearly synchronously with the convolution operation, significantly lowering latency and reducing the complexity of the peripheral control system. This work paves the way for a large-scale on-chip optical neural network computation.