The determination and classification of seafloor sediment types are crucial for the exploitation of marine resources, construction of marine engineering, and maintenance of marine ecological en-vironments. Automatic classification of seafloor sediment based on acoustic telemetry data is an important method to quickly understand the type of a large range of sediment. Currently, most studies on sediment classification are based on multi-beam backscattering intensity data, which is a relatively single data type. Besides, the low-dimensional details of standard U-Net gradually weaken in the propagation process, limiting the accuracy of sediment classification. Therefore, this study proposes an automatic classification method of seafloor sediment types based on an im-proved U-Net and K-means clustering algorithm, using multi-beam water depth, sub-bottom profile, and sample test data in the Northern Slope of the South China Sea. Six sediment types, including gravelly muddy sand, sand, silty sand, less muddy silt, muddy silt, and silty mud, were identified in the study area. Additionally, the study area was divided into four sedimentary en-vironment zones, including the shelf sedimentary area, the upper shelf slope break sedimentary area, the lower shelf slope break mixed sedimentary area, and the slope deposition area, based on the results of sediment classification and geological background. The sedimentary environment zones were found to be distributed along the trend of the shelf slope break line. The results of this study not only provide an important supplement to the existing classification methods of seafloor sediments but also contribute to the understanding of the sedimentary environment and process of the Northern Slope of the South China Sea.

United Nations Framework Convention on Climate Change (UNFCCC) has recently established the Reducing Emissions from Deforestation and forest Degradation (REDD+) program that requires countries to report their carbon emissions and sink estimates through national greenhouse gas inventories (NGHGI). Thus, developing automatic systems capable of estimating the carbon absorbed by forests without in-situ observation becomes essential. To support this important need, in this work we introduce ReUse, a simple but effective deep-learning approach to estimate the carbon absorbed by forest areas based on remote sensing. The novelty of the proposed method is in the use of the public above-ground biomass (AGB) data from the European Space Agency's Climate Change Initiative Biomass project as ground truth to estimate the carbon sequestration capacity of any portion of land on Earth using Sentinel-2 images and a pixel-wise Regressive UNet. The approach has been compared to two literature proposals, using a private dataset and human-engineered features. The results show a greater generalization ability of the proposed approach, with a decrease in Mean Absolute Error and Root Mean Squared Error, respectively, of 16.9 and 14.3 in the area of Vietnam and 4.7 and 5.1 in the area of Myanmar over the runner-up. Finally, as a case study, we reported an analysis made for the Astroni area, a nature reserve located near the metropolitan area of Naples in southern Italy, struck by a large fire, producing predictions consistent with values found by experts in the field. These results further support the use of such an approach for the early detection of AGB variations, both in urban and rural areas.

Already more than 140 countries consider or have pledged to reach net-zero emission targets by 2050 or earlier and the share of global emissions falling into an emission pricing scheme has steeply increased over the past three years. Even where there are no direct implications for industry (yet), there is a series of subtle pressure points driving an increasing number of companies across the globe to work towards climate neutrality and pledging ambitious emission reduction goals. This article sheds light on the pressure points, the subtle triggers, the underlying considerations as well as the hoped-for benefits for industrial companies from achieving net-zero emissions. The observations and ideas presented in this paper are derived from quantitative data obtained via the Energy Efficiency Index of German Industry (EEI) and qualitative data. Not only societal, work force, supply chain and investor expectations play a large role, but also many strategic considerations which have the potential to make the company more resilient and profitable, particularly in time of crisis. Those companies that do not move towards decarbonisation, on the other hand, may face a costly late-mover disadvantage. This piece uncovers subtle interconnections, helping stakeholders from industry and beyond to grasp opportunities and challenges ahead.

This note makes available five years of data gathered in a measurement site equipped with a micrometeorological station and two monitoring wells. Series of data of hydrological and atmospheric variables allow to estimate the flux of water across the atmosphere-land interface and to calculate the water budget, that are crucial topics in climate and environmental sciences. The measures of water-table began during 2017, one of the driest year of the whole instrumental period of climate history for the Central Mediterranean. Data from the micrometeorological station have been used to construct two more datasets of daily and monthly cumulative values of different terms of the surface water budget, from which the groundwater infiltration has been estimated. A significant decreasing trend characterizes both the data time series of piezometric heights and estimated infiltration in the considered period. All the data are collected in 2017-2022 datasets that are available in the supplementary material.

Considering increasingly ambitious pledges by countries, pressure from society, investors, and clients further up the supply chain, the question for companies is not so much whether to take decarbonisation action, but what action and by when. However, determining an ideal mix of measures to apply ‘decarbonisation efficiency’ requires more than knowledge of technically feasible measures and how to combine them to achieve the most economic outcome: In this paper, the author describes seven aspects which heavily influence the composition of an ‘ideal mix’ that executive leadership needs to take a (strategic) position on. These aspects consider underlying motivations and span across (socio-)economic, technical, regulatory, strategic, corporate culture and environmental factors and further underline the necessity of clarity of definitions. How these decisions influence the determination of the decarbonisation-efficient ideal mix of measures, both in principle but also in terms of specific impact is further explored by providing examples. What choices are taken by German manufacturers in several of the aspects is disclosed by insights from about 850 responses to the ‘Energy Efficiency Index of German Industry’. Knowledge of the status quo, and clarity in definitions, objectives, time frames and scope are key.

The notion of potential-growth indicator came to being in the field of matrix population models long ago, almost simultaneously with the pioneering Leslie model for age-structured population dynamics, albeit the term has been given and the theory developed only recent years. The indicator represents an explicit function, R(L), of matrix L elements and indicates the position of the spectral radius of L relative to 1 on the real axis, thus signifying the population growth, or decline, or stabilization. Some indicators turned out useful in theoretical layouts and practical applications prior to calculating the spectral radius itself. The most senior (1994) and popular indicator, R0(L), is known as the net reproductive rate, and we consider two more ones, R1(L) and RRT(A), developed later on. All the three are different in what concerns their simplicity and the level of generality, and we illustrate them with a case study of Calamagrostis epigeios, a long-rhizome perennial weed actively colonizing open spaces in the temperate zone. While the R0(L) and R1(L) fail respectively because of complexity and insufficient generality, the RRT(L) does succeed, justifying the merit of indication.

The NETTER-1 study has proven peptide receptor radionuclide therapy (PRRT) to be one of the most effective therapeutic options for metastatic neuroendocrine tumors (NETs), improving progression-free survival and overall survival. However, PRRT response assessment is challenging and no consensus on methods and timing has yet been reached among experts in the field. This issue is owing to the suboptimal sensitivity and specificity of clinical biomarkers, limitations of morphological response criteria in slowly growing tumors and necrotic changes after therapy, a lack of standardized parameters and timing of functional imaging, and the heterogeneity of PRRT protocols in the literature. The aim of this article is to review the most relevant current approaches for PRRT efficacy prediction and response assessment criteria in order to provide an overview of suitable tools for safe and efficacious PRRT.

The presented study demonstrates the bi-sensor approach suitable for rapid and precise up-to-date mapping of forest canopy gaps for the larger spatial extent. The approach makes use of UAV RGB images on smaller areas for highly precise forest canopy mask creation. Sentinel-2 was used as a scaling platform for transferring information from UAV to a wider spatial extent. The various approaches of the improvement of the predictive performance were examined: (I) the highest R2 of the single satellite index was up to 0.57, (II) the highest R2 using multiple features obtained from the single date, S-2 image was 0.624 and, (III) the highest R2 on the multi-temporal set of S-2 images, was 0.697. Satellite indices such as ARVI, IPVI, NDI45, PSSRa, MCARI, CI, RI, and NDTI were the dominant predictors in most of the ML algorithms. The more complex ML algorithms such as SVM, Random Forest, GBM, XGBoost, and Catboost that provided the best performance on the training set exhibited weaker generalization capabilities. Therefore, a simpler and more robust Elastic Net algorithm was chosen for the final map creation.

Cement manufacturing is one of the most energy-intensive industries in the world. Most of the cost of producing cement is accounted by fuel consumption and power expenditures. Thermal power plants are the major source of electricity in Pakistan. But they are not efficient and environmentally friendly. This study simulates four different models for five cement plants of Pakistan on Homer Pro software and compares the optimal solutions based on the net present cost (NPC), levelized cost of electricity (LCOE) and greenhouse gas (GHG) emissions. Model-1 consists of solar panels, electrolyzer, hydrogen tank, hydrogen generator and converter. Model-2 has only a diesel generator and acts as a base case in this study. Model-3 has solar panels and a battery-converter system. In Model-4, diesel generators, solar panels and converters are considered. Based on NPC, the most optimal model is Model-4, having a 0.249 $/KWh LCOE in islanded systems. The NPC and operating costs are US$540 million and US$ 32.5 million per year, respectively, with a 29.80% reduction in CO2 emissions when compared to the base case. Based on GHG emissions, Model-1 and Model-3 are the best models with 0% GHG emissions. Sensitivity analyses is also performed using the parameters of load, inflation rate and discounted rate. The results prove that the proposed hybrid micropower systems (HMS) can sustainably provide electricity for 24 hours a day to the sites under consideration with minimum objectives.

Amidst the burgeoning global concerns surrounding skin cancer, the quest for precise and efficient diagnostic methodologies has intensified. Conventional diagnostic approaches relying on subjective visual inspection by dermatologists often grapple with subjectivity and resource constraints. This predicament underscores the critical role of Artificial Intelligence (AI) in revolutionizing diagnostic paradigms by significantly enhancing accuracy and accessibility. This study embarks on an extensive exploration of skin cancer classification, specifically targeting the intricate task of 8-class skin cancer classification. Employing cutting-edge deep learning models on the ISIC2019 dataset, a comprehensive analysis was conducted utilizing a diverse array of pre-trained ImageNet architectures and Vision Transformer models. To counteract the inherent class imbalance in skin cancer datasets, a pioneering "Naturalize" augmentation technique was introduced, leading to the creation of indispensable tools—the Naturalized 2.4K ISIC2019 and groundbreaking Naturalized 7.2K ISIC2019 datasets—catalyzing advancements in classification accuracy. The pivotal role of AI in mitigating the risks of misdiagnosis and under-diagnosis in skin cancer becomes evident through this research. AI’s proficiency in analyzing vast datasets and discerning subtle patterns augments the diagnostic prowess of dermatologists significantly. Integration of AI-powered systems in skin cancer diagnosis holds immense promise in facilitating early detection and substantially improving patient outcomes. The meticulous evaluation employed quantitative measures such as confusion matrices, classification reports, and visual analyses using Score-CAM across diverse dataset variations. Astonishingly, the culmination of these endeavors culminated in an unprecedented achievement—100% average accuracy, precision, recall, and F1-Score—within the groundbreaking Naturalized 7.2K ISIC2019 dataset. This remarkable feat underscores the transformative potential of AI-driven methodologies, exemplifying their prowess in redefining the landscape of skin cancer diagnosis and management. In summary, this research represents a pivotal stride towards redefining dermatological diagnosis, showcasing the remarkable impact of AI-powered solutions in surmounting the challenges inherent in skin cancer diagnosis. The attainment of 100% across crucial metrics within the Naturalized 7.2K ISIC2019 dataset serves as a testament to the transformative capabilities of AI-driven approaches in reshaping the trajectory of skin cancer diagnosis and patient care.

Combustion parameters are the important aspects in the development and research of an internal combustion engines. Hence present work is an attempt to develop an artificial neural network (ANN) model for anticipating the combustion characteristics of computerized CI DI VCR, four strokes, single cylinder, and water cooled engine using mixture of two biodiesel and diesel blends as a fuel. The two biodiesel selected are Jatropha and Simarouba and mixed in the ratio of 75:25. Two biodiesel mixtures is used for the preparation of various blends with diesel. Experiments are carried out at different compression ratio, at different % load and blends B20, B40, B60and neat diesel (B00) fuel to accumulate the assorted combustion data. The input variables are % load, compression ratio, blends and crank angle. The outputs well thought-out are cylinder pressure (CP), net heat release rate (NHRR), cumulative heat release (CHR), rate of pressure rise (RPR), mass fraction burned (MFB) and maximum cylinder pressure. The assessment is focused on developing artificial neural network for the prediction and verifying the combustion characteristics of engine. MATLAB software is used for analysis of different combustion aspects of the engine. Multilayer neural network perceptions are used with Feed-Forward Back Propagation algorithm with performance function for mapping the input and output parameters. The experimental outcomes are compared with ANN estimated results for the analysis. The outcome from the study authenticate that experimental results are in match with ANN results. The accuracy of the ANN model shows the developed model can effectively predict the engine combustion characteristics with excellent regression coefficients ‘R’ and its values are close to unity and the mean square errors are extremely less. ANN model approach is competent towards predicting the combustion aspect of diesel engine with good accuracy

Machine Learning can be used for social good. In this paper, we discuss how it can be used to combat climate change and facilitate land management and farming in developing countries and in particular in Côte d’Ivoire. This paper explores models that improve land and water management and agricultural farming cultivation to contrast climate change. Côte d’Ivoire is the largest producer of cocoa beans (43%) in the world, but deforestation, lack of rainfall, drought, and climate change threaten crops and the already fragile economy of Ivorian farmers. It is important to combat climate change with methods and techniques that are affordable to the local farmers and also induce positive effects in production. We discuss the use of low-cost sensors to collect data on the soil and open data and open source software to develop AI tools. We show that using deep neural networks (YOLOv5m) is effective for detecting healthy plants and pods of cocoa from damaged ones only using mobile phone images. Focusing on a single land is not enough to combat climate change, which has different causes and involves also knowledge at a higher scale. We propose a new method of forecasting for the analysis of remote sensors. Remote sensor data come from GRACE NASA Mission and ERA5 produced by the Copernicus Climate Change Service at ECMWF. We implement a new deep neural network architecture named CIWA-net. It is based on a Fully Convolutional Neural Network (FCN) [ 1] and it is a U-net like architecture [2]. The aim of CIWA-net is to forecast Total Water Storage Anomalies (TWSA). We show the quality of our model with a comparison to a vanilla Convolutional Neural Network. CIWA-net could be used also for the detection of lands that interfere with agricultural work and yields, such as deserted areas, water-soaking soil areas, zones at risk of desertification, and poor land use. The employment of AI at the service of agriculture can decrease crop losses and waste, lower the inputs onto the soil of fertilizers, responsible for the increase of Greenhouse Gases. It could be useful to help the small farmers (at a local scale) and also the policy-makers and farmers’ cooperatives (at the regional scale) to take valid and coordinated countermeasures to improve the correct use of the lands, helping to contrast and adapt to climate change.

For its enthusiasts, the blockchain is an archivist's dream come true: an immutable historical record that is publicly accessible and immune from censorship. Proposals like Juried Protocol Galleries could enable bots to compute the provenance of digital artworks and their appearance in exhibitions and publications. Distributed file systems like IPFS claim to store creative works in a redundant, networked fashion outside the control of any single cloud provider. Projects designed to work with blockchains like Filecoin and Arweave propose a dedicated cryptocurrency as financial incentive to support the cost of governing and sustaining this "permaweb." As dreamy as this picture sounds, many of these promises depend on technologies that have yet to be sufficiently developed or adopted. Rather than forecast the future, we can separate the hype from the reality in the present day by examining a real-life example, namely the 2021 auction of Andy Warhol's digital art with NFTs. This essay will focus less on blockchain's general promise as a preservation medium and more on the particular case of the digital Warhols, which both in form and spirit would seem a perfect application of NFTs to preserve historically important works of digital art. Which promises of the crypto-dream of permanent access to digital heritage ring true for this case study, and which are overblown?

Conducting computational stress-strain analysis using finite element methods (FEM) is a common approach when dealing with the complex geometries of atherosclerosis, which is a leading cause of global mortality and complex cardiovascular disease. The considerable expense linked to FEM analysis encourages the substitution of FEM with a considerably faster data-driven machine learning (ML) approach. This study investigated the potential of end-to-end deep learning tools as a more effective substitute for FEM in predicting stress-strain fields within 2D cross sections of arterial wall. We first proposed a U-Net based fully convolutional neural network (CNN) to predict the von Mises stress and strain distribution based on the spatial arrangement of calcification within arterial wall cross-sections. Further, we developed a conditional generative adversarial network (cGAN) to enhance, particularly from the perceptual perspective, the prediction accuracy of stress and strain field maps for arterial walls with various calcification quantities and spatial configurations. On top of U-Net and cGAN, we also proposed their ensemble approaches to further improve the prediction accuracy of field maps. Our dataset, consisting of input and output images, was generated by implementing boundary conditions and extracting stress-strain field maps. The trained U-Net models can accurately predict von Mises stress and strain fields, with structural similarity index scores (SSIM) of 0.854 and 0.830 and mean squared errors of 0.017 and 0.018 for stress and strain, respectively, on a reserved test set. Meanwhile, the cGAN models in a combination of ensemble and transfer learning techniques demonstrate high accuracy in predicting von Mises stress and strain fields, as evidenced by SSIM scores of 0.890 for stress and 0.803 for strain. Additionally, mean squared errors of 0.008 for stress and 0.017 for strain further support the model's performance on a designated test set. Overall, this study developed a surrogate model for finite element analysis, which can accurately and efficiently predict stress-strain fields of arterial walls regardless of complex geometries and boundary conditions.

This article presents a unique set of voltage and current data from a public building and acquired using a hybrid measurement solution that combines Python and Grafana. The transversal purpose consists of contributing to the community with a vision of the quality of the supply more oriented to the monitoring of the state of the network, providing a more realistic vision, which allows a better understanding, and the adoption of the best decisions to achieve the efficient energy management and thus optimize the operation and maintenance of power systems. The work focuses on higher order statistical estimators that, combined with exploratory data analysis techniques, improve the characterization of the shape of the stress signal. These techniques and data, together with the acquisition and monitoring system, present a unique combination in the line of low-cost measurement solutions. It also incorporates the underlying benefit of the contribution to industrial benchmarking. The paper also includes a computational comparison between Python and LabVIEW to elicit the performance of the measurement solution.

There is an increased interest in the district-scale energy transition within interdisciplinary research community. Agent-based modelling presents a suitable approach to address variety of questions related to policies, technologies, processes, and the different stakeholder roles that can foster such transition. This state-of-the-art review focuses on the application of agent-based modelling for exploring policy interventions that facilitate the decarbonisation (i.e., energy transition) of districts and neighbourhoods while considering stakeholders’ social characteristics and interactions. We systematically select and analyse peer-reviewed literature and discuss the key modelling aspects, such as model purpose, agents and decision-making logic, spatial and temporal aspects, and empirical grounding. The analysis reveals that the most established agent-based models’ focus on innovation diffusion (e.g., adoption of solar panels) and dissemination of energy-saving behaviour among a group of buildings in urban areas. We see a considerable gap in exploring the decisions and interactions of agents other than residential households, such as commercial and even industrial energy consumers (and prosumers). Moreover, measures such as building retrofits and conversion to district energy systems involve many stakeholders and complex interactions between them that up to now have hardly been represented in the agent-based modelling environment.

Cement industry is one of the highest energy consuming industries. The quantity of fuel and energy needed accounts for most of the cost of cement manufacturing. Thermal power plants generate electricity but are harmful and ineffective by nature. As a backup mechanism to account for main grid failures, batteries can be utilized. In this paper the first ever investigation on battery’s depth of discharge (DOD) for four different kinds of battery technologies is carried out in the framework of cement industry. The intended battery technologies are the lead-acid battery (LA), lithium-ion battery (Li-ion), vana-dium redox battery (VR), and nickel–iron battery (Ni-Fe). Four hybrid energy generation models (HEGMs) for five cement plants of Pakistan using the HOMER pro software are proposed. Cement plants includes Askari Cement Plant, Wah (ACPW); Bestway Cement Plant, Kalar Kahar (BCPKK); Bestway Cement Plant, Farooqia (BCPF); Bestway Cement Plant, Hattar (BCPH); and DG Cement Plant, Chakwal (DGCPC). HEGM-1 comprises of a diesel generator (DGen), a photovoltaic system (PV), a converter, and a battery system. HEGM-2 comprises of a PV system, a converter, and a battery system. HEGM-3 is the grid-connected version of HEGM-1 and HEGM-4 is the grid-connected version of HEGM-2. A base-model consisting of grid only is used as a reference. A multi-criteria decision analysis (MCDA) is performed by formulating a cumulative objective function (COF) which includes net present cost (NPC), levelized cost of energy (LCOE), and greenhouse gas (GHG) emissions. The principal objective is the maximization of COF while simultaneously minimizing the objectives (NPC, LCOE and GHG emissions), based on optimal battery technology and DOD. The results reveal that VR is the most suitable battery technology with 10% DOD. It is achieved for DGCPC with HEGM-3 with 61.49% of NPC, 78.62% of LCOE and 84.00% of GHG emissions reduction as com-pared to the base model.

Manufacturing systems need to meet I4.0 guidelines to deal with uncertainty in scenarios of turbulent demand for products. The engineering concepts to define the service’s resources to manufacture the products will be more flexible, ensuring the possibility of re-planning in operation. These can follow the engineering paradigm based on capabilities. The virtualization of industry components and assets achieves the RAMI 4.0 guidelines and (I4.0C), which describes the Asset Administration Shell (AAS). However, AAS are passive components that provide information about I4.0 assets. The proposal of specific paradigms is exposed for managing these components, as is the case of multi-agent systems (MAS) that attribute intelligence to objects. The implementation of resource coalitions with evolutionary architectures (EAS) applies cooperation and capabilities’ association. Therefore, this work focuses on designing a method for modeling the asset administration shell (AAS) as virtual elements orchestrating intelligent agents (MAS) that attribute cooperation and negotiation through contracts to coalitions based on the engineering capabilities concept. The systematic method suggested in this work is partitioned for the composition of objects, AAS elements, and activities that guarantee the relationship between entities. Finally, Production Flow Schema (PFS) refinements are applied to generate the final Petri net models (PN) and validate them with Snoopy simulations. The results achieved demonstrate the validation of the procedure, eliminating interlocking and enabling liveliness to integrate elements behavior.

This study examines the potential of combining decision support approaches to identify optimal bundles of ecosystem services. A forested landscape, Zona de Intervenção Florestal of Paiva and Entre-Douro and Sousa (Portugal), is used to test and demonstrate this potential. The landscape extends over 14,000 ha, representing 1,976 stands. The property is fragmented into 376 holdings. The overall analysis was performed in three steps. First, we selected six alternative solutions (A to F) in a Pareto frontier generated by a multiple criteria method within a decision support system (SADfLOR) for subsequent analysis. Next, an aspatial strategic multi-criteria decision analysis (MCDA) analysis was performed with the Criterium DecisionPlus (CDP) component of another decision support system (EMDS) to assess the aggregate performance of solutions A to F for the entire forested landscape with respect to their utility for delivery of ecosystem services. For the CDP analysis, SADfLOR data inputs were grouped into two sets of primary criteria: Wood Harvested and Other Ecosystem Services. Finally, a spatial logic-based assessment of solutions A to F for individual stands of the study area was performed with the NetWeaver component of EMDS. The NetWeaver model was structurally and computationally equivalent to the CDP model, but the key NetWeaver metric is a measure of the strength of evidence that solutions for specific land stands were optimal for the unit. Solutions D and B performed best in the aspatial strategic MCDA analysis, and a composite of the maps generated by NetWeaver demonstrated the spatial basis for the performance of solutions D and B in individual land stands. We conclude with a discussion of how the combination of decision support approaches encapsulated in the two systems could be further automated.

Recent studies have proven that synthetic aperture radar (SAR) automatic target recognition (ATR) models based on deep neural networks (DNN) are vulnerable to adversarial examples. However, existing attacks are easily failed in the case where adversarial perturbations cannot be fully fed to victim models. We call this situation perturbation offset. Moreover, since background clutter takes up most of the areas in SAR images and has low relevance to recognition results, fooling models with global perturbations is quite inefficient. This paper proposes a semi-whitebox attack network, called Universal Local Adversarial Network (ULAN), to generate universal adversarial perturbations (UAP) for the target regions of SAR images. In the proposed network, we calculate the model’s attention heatmaps through layer-wise relevance propagation (LRP), which is used to locate the target regions of SAR images that have high relevance to recognition results. In particular, we utilize a generator based on the U-Net to learn the mapping from noise to UAPs and craft adversarial examples by adding the generated local perturbations to target regions. Experiments indicate that the proposed method fundamentally prevents perturbation offset and achieves comparable attack performance to conventional global UAPs by perturbing only a quarter or less of SAR image areas.

In this paper, we redefine the boundaries of blood cell classification, expanding from 5 to 11 distinct blood cell types using the challenging 11-class PBC dataset. This shift enables a deeper understanding of blood cell diversity, overcoming previous limitations in medical image analysis. Our approach combines state-of-the-art deep learning techniques, including pre-trained ConvNets, ViTb16 models, and custom CNN architectures. We employ transfer learning, fine-tuning, and ensemble strategies, such as CBAM and Averaging ensembles, to achieve unprecedented accuracy and interpretability. Our fully fine-tuned EfficientNetV2 B0 model sets a new standard, with a macro-average precision, recall, and F1-score of 91%, 90%, and 90%, respectively, and an average accuracy of 93%. This breakthrough underscores the transformative potential of 11-class blood cell classification for more precise medical diagnoses. Moreover, our groundbreaking "Naturalize" augmentation technique produces remarkable results. The 2K-PBC dataset generated with "Naturalize" boasts a macro-average precision, recall, and F1-score of 97%, along with an average accuracy of 96% when leveraging the fully fine-tuned EfficientNetV2 B0 model. This innovation not only elevates classification performance but also addresses data scarcity and bias in medical deep learning. Our research marks a paradigm shift in blood cell classification, enabling more nuanced and insightful medical analyses. The "Naturalize" technique’s impact extends beyond blood cell classification, emphasizing the vital role of diverse and comprehensive datasets in advancing healthcare applications through deep learning.

Supermarkets in Port Harcourt (PH) city, Nigeria, predominantly rely on diesel electricity generation due to grid instability, leading to high cost of electricity prices. Although solar photovoltaic (PV) systems have been proposed as an alternative, these supermarkets have yet to adopt them, mainly due to high investment costs and a lack of awareness of the long-term financial and environmental benefits. This paper examines the technical and economic practicality of a PV system for these supermarkets using the PVsyst software and a spreadsheet model. Solar resources showed that PH has a daily average solar radiation and temperature of 4.21 kWh/m2/day and 25.73℃, respectively. Market Square, the supermarket with the highest peak power demand of 59.8 kW, and a 561 kWh/day load profile, was chosen as a case study. A proposed PV system with a power capacity of 232 kW, battery storage capacity of 34,021 Ah, a charge controller size of 100 A/560V, and an inverter with a power rating of 60V/75 kW has been designed to meet the load demand. The economic analysis showed a $266,936 life cycle cost, $0.12 per kWh levelized cost of electricity (LCOE), 4-year simple payback time, and a 20.5% internal rate of return (IRR). The PV system is feasible due to its positive net present value (NPV) of $165,322 and carbon savings of 582 tCO2/year.

Deep Learning (DL) has made significant advances in computer vision with the advent of Vision Transformers (ViT). Unlike Convolutional Neural Networks (CNNs), ViTs use self-attention to extract both local and global features from image data, and then use residual connections to feed these features directly into a fully networked multilayer perceptron head. In hospitals, hematologists prepare peripheral blood smears (PBSs) and read them under a medical microscope to detect abnormalities in blood counts such as leukemia. However, this task is time-consuming and prone to human error. This study investigates the transfer learning process of Google ViT and ImageNet CNNs to automate the reading of PBSs. The study used two online PBS datasets, PBC and BCCD, and transferred them into balanced datasets to investigate the influence of data amount and noise immunity on both neural networks. The PBC results show that Google ViT is an excellent DL neural solution for data scarcity. The BCCD results show that Google ViT is superior to ImageNet CNNs in dealing with unclean, noisy image data because it is able to extract both global and local features and use residual connections, despite the additional time and computational overhead.

Forest carbon sequestration is a widely accepted natural climate solution, however, methods to determine net carbon offsets are limited to commercial carbon proxies and CO2 eddy covariance research. Non-CO2 greenhouse gases (GHG) (e.g., CH4, N2O) receive less attention in the context of forests, in part, due to emphasis on CO2 and the operational requirements and cost for three-gas eddy covariance platforms. In this study, Howland forest flux tower (CO2, CH4) and soil flux data (CO2, CH4, N2O), representing net emission reductions, are linked to their respective social costs to estimate commercial revenue if sold as a GHG social cost forest offset product (GHG-SCF). Estimated annual revenue for GHG-SCF products, applicable to realization of a Green New Deal, range from 120,000 covering the site area of 557 acres in 2021, to 12,000,000 for extrapolation to 40,000 acres in 2040, assuming a 3% discount rate. The Howland Forest CO2 flux record for two adjacent towers is compared to California Air Resources Board forest carbon proxy data for compliance sequestration offsets, the only project site where these approaches overlap. Overcrediting, incomplete carbon accounting with annual errors of up to 2,256%, inadequate third-party verification, and limited application to non-CO2 GHG’s are established. In contrast, direct measurement of one or more GHG’s offers new forest products and revenue incentives to restore and conserve forests worldwide.