Network Modeling Framework Research Articles

Motif discovery in hospital ward vital signs observation networks

Vital signs observations are regular measurements used by healthcare staff to track a patient’s overall health status on hospital wards. We look at the potential in re-purposing aggregated and anonymised hospital data sources surrounding vital signs recording to provide new insights into how care is managed and delivered on wards. In this paper, we conduct a retrospective longitudinal observational study of 770,720 individual vital signs recordings across 20 hospital wards in South Wales (UK) and present a network modelling framework to explore and extract behavioural patterns via analysis of the resulting network structures at a global and local level. Self-loop edges, dyad, triad, and tetrad subgraphs were extracted and evaluated against a null model to determine individual statistical significance, and then combined into ward-level feature vectors to provide the means for determining notable behaviours across wards. Modelling data as a static network, by aggregating all vital sign observation data points, resulted in high uniformity but with the loss of important information which was better captured when modelling the static-temporal network, highlighting time’s crucial role as a network element. Wards mostly followed expected patterns, with chains or stand-alone supplementary observations by clinical staff. However, observation sequences that deviate from this are revealed in five identified motif subgraphs and 6 anti-motif subgraphs. External ward characteristics also showed minimal impact on the relative abundance of subgraphs, indicating a ‘superfamily’ phenomena that has been similarly seen in complex networks in other domains. Overall, the results show that network modelling effectively captured and exposed behaviours within vital signs observation data, and demonstrated uniformity across hospital wards in managing this practice.

Efficient multi-fidelity reduced-order modeling for nonlinear flutter prediction

Flutter prediction is an important part of aircraft design. However, high-fidelity predictions for transonic flutter are difficult to make because of the associated computational costs. This paper proposes a multi-fidelity reduced-order modeling (MFROM) framework for flutter predictions to achieve high-fidelity simulations with limited computational costs. The high-fidelity data were obtained from a Navier–Stokes-equation-based solver, while the low-fidelity data were taken from an Euler-equation-based flow solver. By employing a multi-fidelity neural network trained with two types of data, this methodology enables nonlinear predictions for transonic results. To demonstrate the multi-fidelity process, a widely used pitching and plunging airfoil case is considered. Verification of the approach was performed by comparing with results from time-domain aeroelastic solvers. The results showed that the proposed multi-fidelity neural network modeling framework could realize online predictions of unsteady aerodynamic forces and flutter responses across multiple Mach numbers. Compared with the typical multi-fidelity method, the proposed neural network has a higher accuracy and a stronger generalization capability. Finally, the potential of the method to reduce the computational effort of high-fidelity aeroelastic analysis was demonstrated.

A Multi-Omics, Machine Learning-Aware, Genome-Wide Metabolic Model of Bacillus Subtilis Refines the Gene Expression and Cell Growth Prediction.

Given the extensive heterogeneity and variability, understanding cellular functions and regulatory mechanisms through the analysis of multi-omics datasets becomes extremely challenging. Here, a comprehensive modeling framework of multi-omics machine learning and metabolic network models are proposed that covers various cellular biological processes across multiple scales. This model on an extensive normalized compendium of Bacillus subtilis is validated, which encompasses gene expression data from environmental perturbations, transcriptional regulation, signal transduction, protein translation, and growth measurements. Comparison with high-throughput experimental data shows that EM_iBsu1209-ME, constructed on this basis, can accurately predict the expression of 605 genes and the synthesis of 23 metabolites under different conditions. This study paves the way for the construction of comprehensive biological databases and high-performance multi-omics metabolic models to achieve accurate predictive analysis in exploring complex mechanisms of cell genotypes and phenotypes.

Marginal models with individual-specific effects for the analysis of longitudinal bipartite networks

AbstractA new modeling framework for bipartite social networks arising from a sequence of partially time-ordered relational events is proposed. We directly model the joint distribution of the binary variables indicating if each single actor is involved or not in an event. The adopted parametrization is based on first- and second-order effects, formulated as in marginal models for categorical data and free higher order effects. In particular, second-order effects are log-odds ratios with meaningful interpretation from the social perspective in terms of tendency to cooperate, in contrast to first-order effects interpreted in terms of tendency of each single actor to participate in an event. These effects are parametrized on the basis of the event times, so that suitable latent trajectories of individual behaviors may be represented. Inference is based on a composite likelihood function, maximized by an algorithm with numerical complexity proportional to the square of the number of units in the network. A classification composite likelihood is used to cluster the actors, simplifying the interpretation of the data structure. The proposed approach is illustrated on simulated data and on a dataset of scientific articles published in four top statistical journals from 2003 to 2012.

A framework for detecting fighting behavior based on key points of human skeletal posture

Detecting fights from videos and images in public surveillance places is an important task to limit violent criminal behavior. Real-time detection of violent behavior can effectively ensure the personal safety of pedestrians and further maintain public social stability. Therefore, in this paper, we aim to detect real-time violent behavior in videos. We propose a novel neural network model framework based on human pose key points, called Real-Time Pose Net (RTPNet). Utilize the pose extractor (YOLO-Pose) to extract human skeleton features, and classify video level violent behavior based on the 2DCNN model (ACTION-Net). Utilize appearance features and inter frame correlation to accurately detect fighting behavior. We have also proposed a new image dataset called VIMD (Violence Image Dataset), which includes images of fighting behavior collected online and captured independently. After training on the dataset, the network can effectively identify skeletal features from videos and locate fighting movements. The dataset is available on GitHub (https://github.com/ChinaZhangPeng/Violence-Image-Dataset). We also conducted experiments on four datasets, including Hockey-Fight, RWF-2000, Surveillance Camera Fight, and AVD dataset. These experimental results showed that RTPNet outperformed the most advanced methods in the past, achieving an accuracy of 99.4% on the Hockey-Fight dataset, 93.3% on the RWF-2000 dataset, and 93.4% on the Surveillance Camera Fight dataset, 99.3% on the AVD dataset. And with speeds capable of reaching 33fps, state-of-the-art results are achieved with faster speeds. In addition, RTPNet can also have good detection performance in violent behavior in complex backgrounds.

(Invited) Engineering Porous Electrodes for Redox Flow Batteries – Leveraging Computational and Synthetic Platforms

Porous electrodes are critical components in redox flow batteries where they are responsible for multiple critical functions in the cell, such as providing surfaces for electrochemical reactions, an open structure for fluid flow and mass transport, and the conductive scaffold for electronic and thermal conduction. However, our current portfolio of materials is limited to carbon-fiber-based electrodes which are assembled using various mechanical methods (e.g., paper making, weaving, hydro-entangling) resulting in distinct structures such as papers, cloths, and felts (1). These fabrication methods involve multiple complex subprocessing steps impacting the final product cost and limiting the achievable range of electrode properties, which ultimately restricts device performance (2). Thus, there is a need to develop new material sets with precise control over microstructure and surface composition while employing synthetic methods that are compatible with large-scale manufacturing.In this lecture, I will discuss our efforts to design tailored electrodes for redox flow batteries, which hold promise for long-duration stationary energy storage (3). First, I will introduce a pore network modeling framework that captures electrochemical performance in complex microstructures with low computational requirements (4), and then show the coupling of this framework with a genetic algorithm to enable bottom-up electrode design (5). In our most recent effort, we are building topology optimization models of porous electrodes using multi-scale modeling of redox flow batteries, coupling computational macro- and meso-scale models of transport processes constructed using finite element and finite volume methods. We then couple these approaches with upscaling strategies that enable manufacturability of the structures. Second, I will describe a new approach to synthesizing porous electrodes based on non-solvent induced phase separation (NIPS) (6). NIPS is a simple and versatile fabrication method that can be used to synthesize morphologically-diverse electrode microstructures (e.g., isoporous, pore-size gradients, and multimodal porosity) by tuning simple parameters in the polymer solution, such as polymer concentration, solvent type, and temperature. Using a suite of microscopic, spectroscopic, and electrochemical diagnostic tools, I will present synthesis-property-performance relationships of the NIPS electrodes (7). Third and finally, I will highlight our efforts to functionalize porous electrode interfaces with organic molecules to impart targeted functional properties including hydrophilicity, kinetic activity (8), and selectivity (e.g. H2 evolution suppression). References A. Forner-Cuenca, F. R. Brushett, Curr. Opin. Electrochem. 18, 113–122 (2019).C. Minke, U. Kunz, T. Turek, J. Power Sources. 342, 116–124 (2017).M. L. Perry, K. E. Rodby, F. R. Brushett, ACS Energy Lett. 7, 659–667 (2022).M. van der Heijden, R. van Gorp, M. A. Sadeghi, J. Gostick, A. Forner-Cuenca, J. Electrochem. Soc. 169, 040505 (2022).R. van Gorp, M. van der Heijden, M. A. Sadeghi, J. Gostick, A. Forner-Cuenca, Chem. Eng. J. 455, 139947 (2023).C. T. Wan et al., Adv. Mater. 33, 2006716 (2021).R. R. Jacquemond et al., Cell Reports Phys. Sci. 3, 100943 (2022).E. B. Boz, P. Boillat, A. Forner-Cuenca, ACS Appl. Mater. Interfaces. 14, 41883–41895 (2022).

Topological network and fuzzy AHP modeling framework for the suitability analysis of evacuation shelters: A case study in Japan

Evacuation shelters play a critical role in disasters, providing a safe environment for evacuees to remain safe from threats or disasters. In this study, we analyze the suitability of evacuation shelters using a topological network and fuzzy analytic hierarchy process (AHP) modeling framework with GIS applications, in which potential shelters are assessed according to their accessibility, reachability, and critical criteria within an evacuation network. The accessibility and reachability of evacuation shelters are identified using network analysis, whereas the critical criteria for evacuation shelters are deduced from the referenced requirements. The accessibility, reachability, and criteria of shelters are then weighted with fuzzy AHP analysis, and the suitability of potential shelters for use as evacuation shelters are then analyzed. An illustrative case study was conducted to illustrate the modeling framework; an evacuation network with 2797 nodes was constructed in the study area. Of the 2797 nodes, nine were identified as suitable for use as evacuation shelters. These nine included four accredited designated evacuation shelters and another five potential shelters as alternative shelters. The modeling framework in this study provides a systematic approach to identifying the characteristics of potential shelters, which is expected to be useful in decision making for evacuation planning, disaster preparedness and response, and postdisaster activities in recovery and restoration.

Characterizing behavioural differentiation in gene regulatory networks with representation graphs.

We introduce the formal notion of representation graphs, encapsulating the state space structure of gene regulatory network models in a compact and concise form that highlights the most significant features of stable states and differentiation processes leading to distinct stability regions. The concept has been developed in the context of a hybrid system-based gene network modelling framework; however, we anticipate that it can also be adapted to other approaches of modelling gene networks in discrete terms. We describe a practical algorithm for representation graph computation as well as two case studies demonstrating their real-world application and utility. The first case study presents models for three phage viruses. It shows that the process of differentiation into lytic and lysogenic behavioural states for all these models is described by the same representation graph despite the distinctive underlying mechanisms for differentiation. The second case study shows the advantages of our approach for modelling the process of myeloid cell differentiation from a common progenitor into different cell types. Both case studies also demonstrate the potential of the representation graph approach for deriving and validating hypotheses about regulatory interactions that must be satisfied for biologically viable behaviours.

Evaluation of best management practices for mitigating harmful algal blooms risk in an agricultural lake basin using a watershed model integrated with Bayesian Network approach

Lake eutrophication caused by nitrogen and phosphorus has led to frequent harmful algal blooms (HABs), especially under the unknown challenges of climate change, which have seriously damaged human life and property. In this study, a coupled SWAT-Bayesian Network (SWAT-BN) model framework was constructed to elucidate the mechanisms between non-point source nitrogen pollution in agricultural lake watersheds and algal activities. A typical agricultural shallow lake basin, the Taihu Basin (TB), China, was chosen in this study, aiming to investigate the effectiveness of best management practices (BMPs) in controlling HABs risks in TB. By modeling total nitrogen concentration of Taihu Lake from 2007 to 2022 with four BMPs (filter strips, grassed waterway, fertilizer application reduction and no-till agriculture), the results indicated that fertilizer application reduction proved to be the most effective BMP with 0.130 of Harmful Algal Blooms Probability Reduction (HABs-PR) when reducing 40% of fertilizer, followed by filter strips with 0.01 of HABs-PR when 4815ha of filter strips were conducted, while grassed waterway and no-till agriculture showed no significant effect on preventing HABs. Furthermore, the combined practice between 40% fertilizer application reduction and 4815ha filter strips construction showed synergistic effects with HABs-PR increasing to 0.171. Precipitation and temperature data were distorted to model scenarios of extreme events. As a result, the combined approach outperformed any single BMP in terms of robustness under extreme climates. This research provides a watershed-level perspective on HABs risks mitigation and highlights the strategies to address HABs under the influence of climate change.

ViTR-SP: A CT-based vision transformer model for prediction of pneumonitis in patients with non-small cell lung cancer who received thoracic radiotherapy and immunotherapy.

e20034 Background: The combination of thoracic radiotherapy (TRT) and immunotherapy accompanied by the accumulation of pulmonary toxicity. This study aimed to construct a deep radiomics model based on sequential computed tomography (CT) images for radioimmune-associated pneumonitis prediction in non-small cell lung cancer (NSCLC) patients treated with TRT and immune checkpoint inhibitors (ICIs). Methods: Three models were constructed based on a vision transformer (ViT) network using sequential chest CT images, tumor-mask and lung-mask as inputs, respectively. NSCLC patients treated with TRT and ICIs at Shandong Cancer Hospital and Qilu Hospital were included for model training and external validation, respectively. The image-level model was trained based on ViT to identify predictive images, and the image-level probability distribution were performed. Then, a ViT-Recursive Neural Network (ViT-RNN) model framework was constructed at the patient-level to integrate the image-level information to derive the final prediction, which was evaluated in an independent validation cohort. Results: A total of 230 NSCLC patients from Shandong Cancer Hospital were randomly assigned in a 7:3 ratio for model training and validation, respectively. Another 42 patients from Qilu Hospital were used as an independent external validation cohort. The ViT-RNN model for symptomatic pneumonitis (ViTR-SP) based on sequential CT images achieved better area under curve (AUC), sensitivity, and specificity of 0.725, 0.750, and 0.886 than lung-mask based model (0.575, 0.625, and 0.705) and tumor-mask based model (0.658, 0.667, and 0.773) in the internal validation set. In the external validation set, the proposed model achieved the AUC, sensitivity, and specificity of 0.654, 0.818, 0.727, while the lung-mask and tumor-mask based models had the AUC, sensitivity, and specificity of 0.540, 0.545, 0.677, and 0.630, 0.545, 0.742, respectively. The performance of the model was improved after integrating clinical features, with internal and external validation AUC of 0.782 and 0.672, respectively. Conclusions: In conclusion, the clinical combined ViTR-SP neural network risk prediction model based on sequential chest CT images can be used for predicting the risk of developing radioimmune-associated pneumonitis in NSCLC patients.

BN-BacArena: Bayesian network extension of BacArena for the dynamic simulation of microbial communities.

Simulating gut microbial dynamics is extremely challenging. Several computational tools, notably the widely used BacArena, enable modeling of dynamic changes in the microbial environment. These methods, however, do not comprehensively account for microbe-microbe stimulant or inhibitory effects or for nutrient-microbe inhibitory effects, typically observed in different compounds present in the daily diet. Here, we present BN-BacArena, an extension of BacArenaconsisting on the incorporation within the native computational framework of a Bayesian network model that accounts for microbe-microbe and nutrient-microbe interactions. Using in vitro experiments, 16S rRNA gene sequencing data and nutritional composition of 55 foods, the output Bayesian network showed 23 significant nutrient-bacteria interactions, suggesting the importance of compounds such as polyols, ascorbic acid, polyphenols and other phytochemicals, and 40 bacteria-bacteria significant relationships. With test data, BN-BacArena demonstrates a statistically significant improvement over BacArena to predict the time-dependent relative abundance of bacterial species involved in the gut microbiota upon different nutritional interventions. As a result, BN-BacArena opens new avenues for the dynamic modeling and simulation of the human gut microbiota metabolism. MATLAB and R code are available in https://github.com/PlanesLab/BN-BacArena.

Towards Modeling Uncertainties of Self-Explaining Neural Networks via Conformal Prediction

Despite the recent progress in deep neural networks (DNNs), it remains challenging to explain the predictions made by DNNs. Existing explanation methods for DNNs mainly focus on post-hoc explanations where another explanatory model is employed to provide explanations. The fact that post-hoc methods can fail to reveal the actual original reasoning process of DNNs raises the need to build DNNs with built-in interpretability. Motivated by this, many self-explaining neural networks have been proposed to generate not only accurate predictions but also clear and intuitive insights into why a particular decision was made. However, existing self-explaining networks are limited in providing distribution-free uncertainty quantification for the two simultaneously generated prediction outcomes (i.e., a sample's final prediction and its corresponding explanations for interpreting that prediction). Importantly, they also fail to establish a connection between the confidence values assigned to the generated explanations in the interpretation layer and those allocated to the final predictions in the ultimate prediction layer. To tackle the aforementioned challenges, in this paper, we design a novel uncertainty modeling framework for self-explaining networks, which not only demonstrates strong distribution-free uncertainty modeling performance for the generated explanations in the interpretation layer but also excels in producing efficient and effective prediction sets for the final predictions based on the informative high-level basis explanations. We perform the theoretical analysis for the proposed framework. Extensive experimental evaluation demonstrates the effectiveness of the proposed uncertainty framework.

DiseaseNet: a transfer learning approach to noncommunicable disease classification

As noncommunicable diseases (NCDs) pose a significant global health burden, identifying effective diagnostic and predictive markers for these diseases is of paramount importance. Epigenetic modifications, such as DNA methylation, have emerged as potential indicators for NCDs. These have previously been exploited in other contexts within the framework of neural network models that capture complex relationships within the data. Applications of neural networks have led to significant breakthroughs in various biological or biomedical fields but these have not yet been effectively applied to NCD modeling. This is, in part, due to limited datasets that are not amenable to building of robust neural network models. In this work, we leveraged a neural network trained on one class of NCDs, cancer, as the basis for a transfer learning approach to non-cancer NCD modeling. Our results demonstrate promising performance of the model in predicting three NCDs, namely, arthritis, asthma, and schizophrenia, for the respective blood samples, with an overall accuracy (f-measure) of 94.5%. Furthermore, a concept based explanation method called Testing with Concept Activation Vectors (TCAV) was used to investigate the importance of the sample sources and understand how future training datasets for multiple NCD models may be improved. Our findings highlight the effectiveness of transfer learning in developing accurate diagnostic and predictive models for NCDs.

A neural-network-based model of radio-frequency hollow cathode discharge characterized using particle-in-cell/Monte Carlo collision simulation

An understanding of the plasma dynamics of radio-frequency (RF) hollow cathode discharges (HCDs) at low to moderate pressures is important due to their wide range of applications. A HCD consists of a hollow cylindrical cavity in the RF-powered cathode separated from a grounded electrode by a dielectric. In RF HCDs, RF sheath heating can play a significant role in plasma production in addition to secondary electrons. In this study, a single hollow cathode hole is modeled using the particle-in-cell/Monte Carlo collision (PIC-MCC) technique at low pressure, where kinetic effects are important. Characterization of a single hollow cathode using PIC-MCC simulation is, however, computationally expensive. For improved computational efficiency, a neural network modeling framework has been developed using the temporal variations of applied RF voltages as input and the electrode current as output. A space-filling design for computational experiments is used, where the variables include the RF voltage at the fundamental frequency, RF voltage at the second harmonic, and their phase difference. The predictions of the electrode current using the trained neural network model compare well with the results of the PIC/MCC simulations, but at a significantly lower computational cost. The neural network model predicts the current very well inside the training domain, and reasonably well even outside the training domain considered in this study.

Event-Triggered Relearning Modeling Method for Stochastic System with Non-Stationary Variable Operating Conditions

This study presents a novel event-triggered relearning framework for neural network modeling, designed to improve prediction precision in dynamic stochastic complex industrial systems under non-stationary and variable conditions. Firstly, a sliding window algorithm combined with entropy is applied to divide the input and output datasets across different operational conditions, establishing clear data boundaries. Following this, the prediction errors derived from the neural network under different operational states are harnessed to define a set of event-triggered relearning criteria. Once these conditions are triggered, the relevant dataset is used to recalibrate the model to the specific operational condition and predict the data under this operating condition. When the predicted data fall within the training input range of a pre-trained model, we switch to that model for immediate prediction. Compared with the conventional BP neural network model and random vector functional-link network, the proposed model can produce a better estimation accuracy and reduce computation costs. Finally, the effectiveness of our proposed method is validated through numerical simulation tests using nonlinear Hammerstein models with Gaussian noise, reflecting complex stochastic industrial processes.

A Network Modeling Systematics for Transition Paths Toward Climate Neutral Gas Networks—NeMoSys ['nεmɒsis] –

Meaningful network modeling requires high spatial and temporal resolution and thus adds further complexity to the already complex energy systems analyses; unsecured and varying framework conditions, as well as their changes over time, have to be taken into consideration on the backdrop of path‐dependencies for network development over time. This perspectives article provides a systematic framework for gas network modeling, starting with a morphology of modeling dimensions and elements and the hierarchy of influencing factors. It gives an overview on the necessary data and attributes, development steps, and assessment criteria; it showcases solutions for data management, fluid dynamic network simulation, and results assessment that have been applied and developed in the course of the TransHyDE project and other related research projects. Framework and solutions are merged into a consolidated network modeling systematics that shall be further developed and made available to researchers as well as to network planners for practical application and governmental bodies for assessments. The systematics presented here will be further developed, applied, and operationalized in a database and tool compilation with open‐access for collaboration and further development.

A versatile optimization framework for porous electrode design.

Porous electrodes are performance-defining components in electrochemical devices, such as redox flow batteries, as they govern the electrochemical performance and pumping demands of the reactor. Yet, conventional porous electrodes used in redox flow batteries are not tailored to sustain convection-enhanced electrochemical reactions. Thus, there is a need for electrode optimization to enhance the system performance. In this work, we present an optimization framework to carry out the bottom-up design of porous electrodes by coupling a genetic algorithm with a pore network modeling framework. We introduce geometrical versatility by adding a pore merging and splitting function, study the impact of various optimization parameters, geometrical definitions, and objective functions, and incorporate conventional electrode and flow field designs. Moreover, we show the need for optimizing geometries for specific reactor architectures and operating conditions to design next-generation electrodes, by analyzing the genetic algorithm optimization for initial starting geometries with diverse morphologies (cubic and a tomography-extracted commercial electrode), flow field designs (flow-through and interdigitated), and redox chemistries (VO2+/VO2 + and TEMPO/TEMPO+). We found that for kinetically sluggish electrolytes with high ionic conductivity, electrodes with numerous small pores and high internal surface area provide enhanced performance, whereas for kinetically facile electrolytes with low ionic conductivity, low through-plane tortuosity and high hydraulic conductance are desired. The computational tool developed in this work can further expanded to the design of high-performance electrode materials for a broad range of operating conditions, electrolyte chemistries, reactor designs, and electrochemical technologies.

Performance analysis of distance-based resource allocation in UAV-Assisted vehicular ad hoc networks

Intelligent Transportation Systems (ITSs) have undergone a revolution due to the widespread use of vehicular networks, which have made it possible for cars and infrastructure to communicate effectively. The performance Analysis of a Road Side Unit (RSU) and an Unmanned Aerial Vehicle (UAV) in a realistic urban setting is modeled using a vehicular network modeling framework. According to our proposed protocol, when a new vehicle enters a vehicular network at any random position. Then the proposed protocol calculates the distance between the vehicle and RSU and between the vehicle and UAV. Based on the distance calculated, the position of the random vehicle concerning the UAV and RSU is determined. To reduce interference, those vehicles that are nearer to the UAV are assigned to UAV, and those that are nearer to RSU are assigned to RSU. For Successful transmission, the received power of the vehicle is compared with the pre-determined Signal-to-Interference plus Noise Ratio (SINR) threshold. The UAV may offer better throughput than the RSU in a vehicular network simulation due to its higher altitude, transmit power, antenna gain, frequency, and coverage range. We can control the distribution of vehicles between the two coverage regions by changing the UAV_distance_threshold while keeping the RSU_distance_threshold constant. This creates a dynamic system where increasing the UAV_distance_threshold results in a higher proportion of vehicles being served by UAVs and vice versa. Hence our proposed setup can optimize the positioning and configuration of communication infrastructure while considering the coverage and performance of RSUs and UAVs.

A pore-scale numerical framework for solute transport and dispersion in porous media

Pore-scale modeling plays a crucial role in understanding and upscaling solute transport behavior in porous media. Direct simulation offers the highest fidelity in resolving pore-scale fluid flow and mass transfer, nevertheless, with unacceptable computational costs for practical applications. Pore network models (PNM), on the other hand, provide an efficient alternative but with reduced accuracy in representing transport dynamics. In this study, we propose a new framework of pore network models that ensures both accuracy and efficiency, which reproduces the pore-scale shear dispersion effect by utilizing the pore-scale shear dispersion coefficient to calculate the diffusive mass exchange rate between network elements. The coefficient is determined based on the extension of Taylor-Aris theory that expands the classical Taylor-Aris theory to encompass the pre-asymptotic regime. Additionally, the framework adopts a physically representative pore network, where the conduit length and volume of network elements are determined based on local resistance equivalence. After verifying its accuracy and reliability, we conduct a series of numerical cases on tube networks, sphere packs, and sand packs. The breakthrough curves and concentration profiles obtained from our new model show good agreement with direct simulation results and experimental data, while the simulation outcomes of the models that rely on the molecular diffusion coefficient or Taylor dispersion coefficient for local mass exchange may exhibit significant errors. We finally demonstrate that the significance of pore-scale shear dispersion on solute transport weakens for tube networks with increasing degrees of geometrical disorder. The proposed model provides an accurate and efficient numerical framework for the study of solute transport and dispersion in porous media.

EA-Net: Research on skin lesion segmentation method based on U-Net

Accurate segmentation of skin lesions is a challenging task because the task is highly influenced by factors such as location, shape and scale. In recent years, Convolutional Neural Networks (CNNs) have achieved advanced performance in automated medical image segmentation. However, existing CNNs have problems such as inability to highlight relevant features and preserve local features, which limit their application in clinical decision-making. This paper proposes a CNN with an added attention mechanism (EA-Net) for more accurate medical image segmentation.EA-Net is based on the U-Net network model framework. Specifically, we added a pixel-level attention module (PA) to the encoder section to preserve the local features of the image during downsampling, making the feature maps input to the decoder more relevant to the ground-truth. At the same time, we added a spatial multi-scale attention module (SA) after the decoding process to increase the spatial weight of the feature maps that are more relevant to the ground-truth, thereby reducing the gap between the output results and the ground-truth. We conducted extensive segmentation experiments on skin lesion images from the ISIC 2017 and ISIC 2018 datasets. The results demonstrate that, when compared to U-Net, our proposed EA-Net achieves an average Dice score improvement of 1.94% and 5.38% for skin lesion tissue segmentation on the ISIC 2017 and ISIC 2018 datasets, respectively. The IoU also increases by 2.69% and 8.31%, and the ASSD decreases by 0.3783 pix and 0.5432 pix, indicating superior segmentation performance. EA-Net can achieve better segmentation results when the original image of skin lesions has an obscure boundary and the segmentation area contains interference factors, which proves that the addition of attention mechanism in the encoder and the application of comprehensive attention mechanism can improve the performance of neural network in the field of skin lesions image segmentation.