Dimensional Network Research Articles

Sea surface heat flux helps predicting thermocline in the South China Sea

In this study, a deep learning model called Four Dimensional Residual Network (4D-ResNet) was proposed, which can capture both temporal and spatial information. Temperatures at various depths were predicted for the next 40 days using the last month's sea surface variables, and a spatio-temporal prediction of the thermocline was achieved. In addition to the satellite-observed sea surface parameters: sea surface temperature (SST), sea level anomaly (SLA), and sea surface wind (SSW), net heat flux (Qnet) was also included in the model input. Qnet can alter the density of the upper water, resulting in convection or improved stratification stability. The results indicate that the additional input of Qnet improves the model's accuracy, especially at the depth of the thermocline, where the RMSE reduced by up to 13.7%. The 4D-ResNet model has much lower estimation error compared to other models and successfully captures the seasonal characteristics of the thermocline.

A Cd-based crystalline network material: catalytic properties and post-synthetic metal-ion metathesis with enhanced stability and gas sorption behaviour.

This study presents the synthesis of a Cd(II) based hydrophobic three dimensional crystalline network material (CNM), [Cd3(L)2(LH)2(bpe)2], {L = {4,4'-(hexafluroisopropylidine)bis(benzoate)} and 1,2-di(4-pyridyl) ethylene (bpe)}, 1(Cd), by employing the slow-diffusion method. The three-dimensional structure of 1(Cd) was determined by single crystal X-ray diffraction and characterized by powder X-ray diffraction (PXRD), FT-IR spectroscopy and thermogravimetric analysis (TGA). Subsequently, post-synthetic modification of 1(Cd) with Cu(II) at room temperature led to the formation of isostructural 1(Cu) with partial substitution. This transformation, unattainable through de novo synthesis, was monitored using energy dispersive X-ray analysis (EDX), PXRD, FT-IR spectroscopy, and through visual observation confirming a single crystal to single crystal metal exchange. The modified material, 1(Cu), exhibited red-shifted emission with enhanced thermal stability and a tenfold increase in N2 uptake. Furthermore, the catalytic potential of 1(Cd) in aza-Michael addition reactions of α,β-unsaturated olefins to nucleophilic aromatic/aliphatic amines was demonstrated successfully under ambient conditions. This approach employed a heterogeneous and acid-base free methodology, showcasing the versatility and effectiveness of 1(Cd) as a catalyst.

A predictive model for electrical conductivity of polymer carbon nanofiber composites considering nanofiber/interphase network and tunneling dimensions

A predictive model for electrical conductivity of polymer carbon nanofiber composites considering nanofiber/interphase network and tunneling dimensions

MoS2 photonic synaptic device development for non-invasive detection of Parkinson’s disease

Abstract The brain-inspired computation using synaptic devices has opened a new avenue for neuromorphic intelligent systems. This paper presents the development of a MoS2 based photonic synaptic devices using wafer-scalable microfabrication for large scale production. The synaptic device was tested using a 450 nm wavelength having a power density of 17 mW cm−2. The test results show that developed devices has enhanced characteristics such as paired pulse facilitation (PPF) index of 307% and long-term retention of 1000 s. The optically stimulated synaptic current was ∼2.5 times higher than the electrically stimulated synaptic current. The potentiation and depression data of the MoS2 device act as the synaptic interface in the artificial intelligence model. The detection of Parkinson’s disease (PD) using voice spectrograms where MoS2 behaves as photonic synapses in two dimensional convolutional neural networks algorithm provided a promising accuracy of 96%. This neuromorphic computing platform for the detection of PD is a next-generation revolution for efficient and smart healthcare solutions.

The role of disulfide bonds in L-arginine ameliorating the quality of low-salt sturgeon surimi gels induced by microwave: Increasing the diameter and fractal dimension of network

The role of disulfide bonds in L-arginine ameliorating the quality of low-salt sturgeon surimi gels induced by microwave: Increasing the diameter and fractal dimension of network

Network Dimension Theory and Its Application to Miniemulsion Vinyl/Divinyl Copolymerization

Network Dimension Theory and Its Application to Miniemulsion Vinyl/Divinyl Copolymerization

Identification as a Sword in the War for Talent: Understanding Organizational Identification Through a Prospective Employee Lens

ABSTRACTIn the intense “war for talent”, companies increasingly leverage employer branding. Previous research cites the relevance of identification for current employees (internal employer branding); our research instead addresses the relevance of identification for attracting prospective employees (external employer branding). Drawing on social identity and social comparison theory, we consider if and how prospective employees' sense of identification with the company brand, the work community, and the CEO as a human brand influences their intentions to apply. This study features a mixed‐method two‐study design. A qualitative study identifies both internal (team) and external (network) dimensions of identification with the work community. The quantitative study (including 507 job seekers) further suggests that identification with the work community exerts the strongest impact on intentions to apply. The effects of both identification with the work community and with the human brand on intentions to apply are mediated by identification with the company brand. Therefore, companies can enhance the effectiveness of their external employer branding by leveraging all three identification constructs.

Iterative neural networks for improving memory capacity

In recent years, the problem of the multistability of neural networks has been studied extensively. From the research results obtained, the number of stable equilibrium points depends only on a power form of the network dimension. However, in practical applications, the number of stable equilibrium points needed is often not expressed in power form. Therefore, can we determine an appropriate activation function so that the neural network has exactly the required number of stable equilibrium points? This paper provides a new way to study this problem by means of an iteration method. The necessary activation function is constructed by an appropriate iteration method, and the neural network model is established. Based on the mathematical theories of matrix analysis and functional analysis and on the inequality method, the number and distribution of the network equilibrium points are determined by dividing the state space reasonably, and some multistability criteria that are related to the number of iterations and are independent of the network dimension are established.

A new luminescent material based on bismuth(III): Synthesis, structural characterization, DFT calculations, Hirshfeld surface, thermal behavior, vibrational and optical properties

Bismuth-halide-based hybrid materials are desirable in luminescent applications due to low toxicity and chemical stability. (C8H12N)4Bi2Cl10, was elaborated by the slow evaporation technique at room temperature. Single-crystal X-ray diffraction analysis indicates that the compound belongs to the monoclinic crystal system with the centrosymmetric space group P21/c. The formula unit comprises four protonated organic cations (C8H12N)4+, one [Bi2Cl10]4− dimer. The organic layers are inserted between the inorganic ones and connected with N-H…Cl and C-H…Cl hydrogen bonds to build a three dimensional network. The infrared IR and Raman studies which were recorded at room temperature in the 500–4000 cm−1 and 50–4000 cm−1 frequency regions, respectively, confirmed the existence of vibrational modes that correspond to the organic and inorganic groups. The optimized molecular structure and vibrational frequencies were calculated by the Density Functional Theory (DFT) method using the B3LYP level employing level employing a LANL2DZbasis set. It shows a good agreement between the calculated and the experimental vibrational frequencies. Hirshfeld surface analysis of close intermolecular interactions in this compound enables the identification and examination of molecular shapes. The crystal exhibits thermal stability up to 160 °C using the Thermogravimetric analysis TGA. The optical properties were characterized experimentally by UV–visible absorption studies and photoluminescence measurements. The Photoluminescence spectrum shows a green luminescence peak which is attributed to excitonic emissions within the chlorobismuthate octahedron. HOMO-LUMO orbital energies were studied by using DFT calculations.

Dry-Deposited Self-Standing Composite Electrodes for EV Battery Applications

Novel type of composite materials consisting of powdered materials embedded into a 3-dimensional network of pristine individual or a few-bundled single wall carbon nanotubes (SWNTs) has been developed [1]. A wide selection of powdered materials can be used, and SWNT content can be varied in a broad range. The resulted composite materials typically are self-standing, mechanically robust and with electrical conductivity 101-105 S/m, depending on the material composition and density. We have been using this technology to create composites of Li-ion battery materials for self-standing, binder and collector free battery electrodes. Application of these electrodes allows to increase battery energy density up to 40% and increase specific energy up to 70% in relatively small flexible pouch cells [1]. Since we can produce electrodes with very high thicknesses and high loadings of the active material, with minimal amount of electrochemically inactive components, we are studying applicability of such electrodes for large EV-sized batteries and for both high energy and high-power applications typical for EV/EVTOL batteries, and what parameters of the electrode material is necessary for such applications. Normally, EV batteries operate at 0.1C to 0.5C, to allow 10 to 2 hours of driving. However, they need to be able to operate at higher C-rates (2C to 10C typical) for short periods of time, when more power is needed for passing, hill climbing, etc. Battery performance for a particular battery chemistry depends on multiple interdependent parameters: electrode loading, electrode electronic conductivity, electrode density/porosity, electrode thickness, electrolyte ionic conductivity and many others. To achieve high energy density of a battery cell its electrodes are compressed as much as possible, and the amount of the electrolyte is minimized. However, this often leads to insufficient ionic conductivity of the electrodes, and, therefore, poor power performance (high C-rate performance). Parameters of electrodes of EV/EVTOL batteries need to be optimized for these applications. Here we concentrate on dependence of the high C-rate performance of the cells on the electrode density/porosity and electrode thickness with fixed cathode areal loading and conductivity (proportional to the CNT percentage). Cells with electrodes with the density in the 1.5-2 g/cm3 range demonstrate the best C-rate performance in the 3C to 30C range. O.A. Kuznetsov, S. Mohanty, E. Pigos, G. Chen, W. Cai, A.R. Harutyunyan, High energy density flexible and ecofriendly lithium-ion smart battery. Energy Storage Materials, 2023 (54) 266-275, ISSN 2405-8297, doi.org/10.1016/j.ensm.2022.10.023

VC dimension of Graph Neural Networks with Pfaffian activation functions

Graph Neural Networks (GNNs) have emerged in recent years as a powerful tool to learn tasks across a wide range of graph domains in a data-driven fashion. Based on a message passing mechanism, GNNs have gained increasing popularity due to their intuitive formulation, closely linked to the Weisfeiler–Lehman (WL) test for graph isomorphism, to which they were demonstrated to be equivalent (Morris et al., 2019 and Xu et al., 2019). From a theoretical point of view, GNNs have been shown to be universal approximators, and their generalization capability — related to the Vapnik Chervonekis (VC) dimension (Scarselli et al., 2018) — has recently been investigated for GNNs with piecewise polynomial activation functions (Morris et al., 2023). The aim of our work is to extend this analysis on the VC dimension of GNNs to other commonly used activation functions, such as the sigmoid and hyperbolic tangent, using the framework of Pfaffian function theory. Bounds are provided with respect to the architecture parameters (depth, number of neurons, input size) as well as with respect to the number of colors resulting from the 1–WL test applied on the graph domain. The theoretical analysis is supported by a preliminary experimental study.

Pectin Hydrogels as Structural Platform for Antibacterial Drug Delivery.

Hydrogels are hydrophilic 3-dimensional networks characterized by the retention of a large amount of water. Because of their water component, hydrogels are a promising method for targeted drug delivery. The water component, or "free volume", is a potential vehicle for protein drugs. A particularly intriguing hydrogel is pectin. In addition to a generous free volume, pectin has structural characteristics that facilitate hydrogel binding to the glycocalyceal surface of visceral organs. To test drug function and pectin integrity after loading, we compared pectin films from four distinct plant sources: lemon, potato, soybean, and sugar beet. The pectin films were tested for their micromechanical properties and intrinsic antibacterial activity. Lemon pectin films demonstrated the greatest cohesion at 30% water content. Moreover, modest growth inhibition was observed with lemon pectin (p < 0.05). No effective inhibition was observed with soybean, potato, or sugar beet films (p > 0.05). In contrast, lemon pectin films embedded with carbenicillin, chloramphenicol, or kanamycin demonstrated significant bacterial growth inhibition (p < 0.05). The antibacterial activity was similar when the antibiotics were embedded in inert filter disks or pectin disks (p > 0.05). We conclude that lemon pectin films represent a promising structural platform for antibacterial drug delivery.

Dynamic evolution analysis and parameter optimization design of data-driven network infectious disease model

Background and Objective:globalization and population mobility have increased the spread of infectious diseases and challenged public health security. This paper proposes a complex network epidemic model with nonlinear incidence rate and quadratic transmission. The Turing pattern, sensitivity analysis and parameter identification of the epidemic model under different network structures are studied; Methods:this paper discusses the Turing pattern of the model under different network structures, and identifies the key parameters of the model through sensitivity analysis. The influence of network dimension on the spread of infectious diseases on random networks is also explored, and the problems of minimum path and minimum cover set of random networks are further discussed. We also carry out parameter identification experiments, adopt gradient descent algorithm to realize heterogeneous spatial fitting pattern of red blood cell plasma and simulate the transmission path of COVID-19 through Markov chain Monte Carlo fitting experiment, verifying the effectiveness of the model; Results:the necessary conditions for Turing instability on homogeneous and heterogeneous networks are found. On the heterogeneous lattice network, we observe the special patterns of equal density population. Sensitivity analysis shows that the higher the infection rate, the more infected people. On random networks, the higher the dimension, the better the effect of suppressing the spread of infectious diseases. Through comparison experiment, it is found that gradient descent algorithm has the best performance in parameter identification experiments. Red blood cell plasma fitting experiment reveals the spatial density distribution of infection rate; Conclusions:this study provides theoretical support for the prevention and control of infectious diseases, and the complex network model can simulate the transmission process of infectious diseases more accurately. Sensitivity analysis and parameter identification experiments reveal the key influencing factors of propagation and the role of network structure. The effectiveness of the model is supported by actual data, which is helpful for the government health departments to formulate scientific prevention and control strategies.

Analysis of Cardiovascular Disease Classification Through Deep Learning Approach

Multiple cardiovascular disease classification from Electrocardiogram (ECG) signal is necessary for efficient and fast remedial treatment of the patient. This paper presents a method to classify multiple heart diseases using one dimensional deep convolutional neural network (CNN) where a modified ECG signal is given as an input signal to the network. Each ECG signal is first decomposed through Empirical Mode Decomposition (EMD) and higher order Intrinsic Mode Functions (IMFs) is combined to form a modified ECG signal. It is believed that the use of EMD would provide a broader range of information and can provide denoising performance. This processed signal is fed into the CNN architecture that classifies the record according to cardiovascular diseases using soft max regress or at the end of the network. It is observed that the CNN architecture learns the inherent features of the modified ECG signal better in comparison with the raw ECG signal. The method is applied on three publicly available ECG databases and it is found to be superior to other approaches in terms of classification accuracy. In MIT-BIH, St. Petersburg, PTB databases the proposed method achieves maximum accuracy of 0.9770, 0.9971 and 0.9871 respectively.

Abstract 4114276: Identification of the heterogeneity in the risk of heart failure in those with left bundle branch block pattern using artificial intelligence

Introduction: The association between left bundle branch block (LBBB) and heart failure is well-established. However, the lack of a practical risk assessment tool for patients with LBBB poses a clinical challenge to the management despite the heterogeneity of the LBBB population. Hypothesis: Artificial intelligence (AI) can extract features to stratify the risk of heart failure (HF) within the heterogeneity of LBBB. Methods: All 12-lead ECGs diagnosed with LBBB recorded from 1/1/2015 to 12/31/2019 were identified in our institution. ECG data were analyzed as 2D matrices with a shape of 12×2500. The Uniform Manifold Approximation and Projection (UMAP) was used to visualize the heterogeneity of LBBB ECG. Additionally, a 2-dimensional convolutional neural network model was trained to detect LBBB ECGs associated with past HF diagnosis. The model was then applied to ECGs from an external cohort of patients with LBBB but without a history of HF. Cumulative incidences of HF admission were compared by stratifying according to tertiles of the model output (low, intermediate and high AI score groups) Results: We identified 15,124 LBBB ECGs in our institution (training dataset) and 2,463 individuals with LBBB ECGs in the external cohort (external test dataset). The UMAP projection of the ECGs revealed distinct heterogeneity in the data ( Fig A ). This heterogeneity was not fully explained by patient demographics, ECG features, or institutions. When the external cohort was divided into 3 groups according to the tertile of AI prediction, patients with high AI scores had a higher risk of HF admission (high vs low AI score group: hazard ratio (HR), 2.10; 95% confidence interval (CI), 1.66-2.65; intermediate vs low AI score group: HR, 1.39; 95%CI, 1.08-1.78: log-rank p &lt;0.0001; Fig 1B ). This finding was unchanged after adjusting for multiple clinical characteristics and ECG parameters, including QRS duration. Conclusion: We observed significant heterogeneity in ECGs presenting LBBB, and the AI model excellently stratified the risk of heart failure admission in patients with LBBB from a single recording of 12-lead ECG, suggesting great potential for clinical utility in managing patients with LBBB.

Precisely tuning ε′-negative and ε′-near-zero responses by synergistic effect of graphene and carbon black in ternary metacomposites

Understanding how to precisely regulate the magnitude and dispersion characteristics of radio-frequency (RF) ε’-negative and ε’-near-zero (ENZ) responses presents a challenge. This challenge significantly impacts the design of versatile electronic devices. In this study, we introduce a synergistic strategy utilizing graphene (GR) and carbon black (CB) in ternary metacomposites for tunable ε’-negative and ENZ responses. Due to the randomly constructed 3-dimensional (3D) conductive GR-CB networks in a CaCu3Ti4O12 matrix, we observed two types of ε’-negative response mechanisms: electric dipole resonance and low-frequency plasma oscillation, which dominate at different frequency bands. Consequently, the weakly ε’-negative values (0 < |ε’| 〈1000) and frequency dispersion were successfully adjusted due to the interplay between these two mechanisms. The ENZ frequencies were also tuned over ∼580 MHz, 225 MHz, ∼189 MHz and ∼ 188 MHz with variable GR-to-CB ratios. Furthermore, we studied the conduction behavior, loss mechanism, and electrical characteristics of the ε’-negative metacomposites to shed light on the relationship between microstructural changes and dielectric performance.

Experimental and Computational Investigations of Supramolecular Interactions in Cd(II)–Organic Hydrate: Catalytic Applications in the Knoevenagel Condensation Reaction under Solvent‐Free Conditions

AbstractA dinuclear cadmium(II) complex, [Cd2(Hhqc)2(H2O)4] (H3hqc=4,8‐dihydroxyquinoline‐2‐carboxylic acid) was synthesized and characterized by single crystal X‐ray diffraction studies, elemental analyses and thermo‐gravimetric studies. The metal centers are found to be six coordinated with distorted pentagonal pyramidal geometry around them. The complex shows few significant intermolecular hydrogen bonding interactions to produce a three dimensional supramolecular network comprising of T8(2) water tapes. Besides these non‐covalent interactions few other supramolecular interactions are also involved within the molecular dimer like π⋅⋅⋅π stacking, Cd⋅⋅⋅π etc. They have been analyzed using DFT calculations, MEP surface analyses and NCI plot index analyses. The detailed study highlights the complex interplay of π‐stacking, H‐bonding, and spodium interactions in stabilizing the dimers of complex, which providing precious insights into their energetic and structural characteristics. Further more, the catalytic efficiency of the synthesized complex has been investigated as an effective heterogeneous catalyst in the Knoevenagel condensation reactionunder solvent‐free conditions.

A time‐varying and frequency‐dependent network of conventional and organic milk markets in France and Germany

AbstractThis work investigates the linkages among conventional and organic milk markets in France and Germany using the flexible TVP‐VAR connectedness approach. Price connectedness tends to be stronger in the spatial than in the quality dimension of the market network. The bulk of the adjustments to incoming price innovations are completed beyond the 3‐month frequency range. Organic milk markets are more likely to be net‐transmitters of shocks to conventional markets. Total connectedness has not been increasing. The strength of price spillovers between market pairs varies widely with time; their pattern does not, thus pointing to a fairly stable internal network structure.

Associations of Retinal Vessel Geometry and Optical Coherence Tomography Angiography Metrics With Choroidal Metrics in Diabetic Retinopathy.

To elucidate the mechanism underlying changes in choroidal metrics (choroidal thickness [CT], choroidal vascularity index [CVI], and choriocapillaris [CC] flow deficit [FD]) observed in diabetic retinopathy (DR) and examine the association of choroidal metrics with both retinal vessel geometry and optical coherence tomography angiography (OCTA) metrics. Overall, 133 eyes of 133 patients were analyzed retrospectively. Retinal vessel geometry parameters were assessed using semiautomated software. The OCTA metrics and CT were calculated using automated algorithms provided by the manufacturer, whereas the CVI and CC-FD were calculated using ImageJ software from the binarized choroid B-scan image and the CC slab provided by the manufacturer, respectively. To assess the associations among choroidal metrics, retinal vessel geometry, and OCTA metrics, multivariable regression analyses were performed while controlling for clinical features and DR severity. In the multivariable linear regression analysis, CT (β = -399.84; P = 0.014) and CVI (β = -2.34; P = 0.021) showed significant associations with the arteriole-venule ratio, which is a ratio of central retinal arteriolar equivalent caliber with respect to central retinal venular equivalent caliber. The CC-FD showed a significant association with the fractal dimension of retinal arteriolar network (β = -2.90; P = 0.040). In contrast, the OCTA metrics showed no significant association with the choroidal metrics. The CT, CVI, and CC-FD in patients with DR were associated with retinal arteriolar geometry parameters rather than OCTA metrics, which indicates an association between choroidal changes and hemodynamic alterations in retinal arterioles and venules.

HNet: A deep learning based hybrid network for speaker dependent visual speech recognition

Visual Speech Recognition (VSR) is a popular area in computer vision research, attracting interest for its ability to precisely analyze lip motion and seamlessly convert them into textual representation. VSR systems leverage visual features to augment the understanding of automated speech and predict text. VSR finds various applications, including enhancing speech recognition in scenarios with degraded acoustic signals, aiding individuals with hearing impairments, bolstering security by reducing reliance on text based passwords, facilitating biometric authentication for liveness detection, and enabling underwater communications. Despite the various techniques proposed for improving the resilience and precision of automatic speech recognition, VSR still has challenges like homophones of words, gradient descent issues with varying sequence lengths, and lip reading demands accounting for short and long range correlations between consecutive video frames. We have proposed a hybrid network (HNet) with multilayered three dimensional dilated convolution neural network (3D-CNN). The spatio-temporal feature extraction process will be facilitated by dilated 3D-CNN. HNet integrates two bidirectional recurrent neural networks (BiGRU and BiLSTM) to process the feature sequences bidirectionally to establish the temporal relationship. The fusion of BiGRU-BiLSTM capabilities allows the model to process feature sequences more comprehensively and effectively. The proposed work focuses on face based biometric authentication for liveness detection using the VSR model to boost security against face spoofing. The existing face based biometric systems are widely used for individual authentication and verification but are still vulnerable to 3D masks and adversarial attacks. The VSR system can be added to existing face based verification systems as a second level authentication technique to identify a person with liveness. The working ideology of the VSR system will be based on the challenge response technique, where a person has to pronounce the passcode silently displayed on the screen. The VSR model assesses its effectiveness using word error rate (WER), which matches the pronounced passcode to the one presented on the screen. Overall, the proposed work aims to enhance the accuracy of VSR so that it can be combined with existing face based authentication systems. The proposed system outperforms the existing VSR system and obtained 1.3% WER. The significance of the proposed hybrid model is that it efficiently captures temporal dependencies, enhancing context embedding, improving robustness to input variability, reducing information loss, and enhancing performance and accuracy in modeling and analyzing passcode pronunciation patterns.