Reaching net-zero emissions by the mid of this century requires the implementation of massive carbon dioxide (CO₂) emissions reduction strategies along with other greenhouse gases at both global and country scale. Thus, carbon capture, storage, and utilization (CCSU) is one of the promising technologies in combination with renewable energy transition. Currently, CO₂ utilization has been extensively attracted by scientific community worldwide, since it can improve the economic viability of CCSU deployment via creating a market for the recovered CO₂ stream. In this study, a brief assessment and comparison of potential CO₂ utilization pathways in Uzbekistan including CO₂ to chemicals/fuel conversion, CO₂ bio-fixation/mineralization, and direct use of CO₂ such as for enhanced hydrocarbon recovery (EHR) have been conducted considering the CO₂ stationary sources and site-specific conditions of the country. Apart from that, possible challenges and opportunities for large scale CO₂ utilization routes have also been discussed. According to the assessment, there is a great potential for CO₂ direct use as a process boosting agent for EHR in more than 22 major natural gas, crude oil, and coal reservoirs. Moreover, methanol and urea production processes can also make a huge market demand for recovered CO₂ as long as the conventional CO₂ production processes are replaced by sustainable one’s.

With the recent COVID-19 challenges, a growing interest to develop an efficient, economical, and effective infectious waste segregation system has prompted both the health sector and the government. This study presented a Microcontroller-Based Automated Infectious Waste Segregation and Disinfection System in a selected public medical facility in Metro Manila, Philippines. The prototype system applying the machine learning principles is capable of identifying three kinds of waste materials classified as sharps, electronic, and pathological wastes as interpreted by the Phyton Image processing software. In addition, an added feature of UV light mechanism to address the bacterial presence of Staphylococcus aureus and Escherichia coli. was incorporated in the prototype to ensure disinfection. Results showed that the mean average precision (maP) of identification was 95.7, 79.9, and 94.5%, respectively. Moreover, it was found that there was a noticeable decrease in the bacterial count signifying the effectivity of the prototype and has promising potential for large-scale implementation.

Multispectral remote sensing images classification is a challenging task due to the complexity and variety of natural scenes. Deep Learning has revolutionized the field of remote sensing by achieving impressive performance. However, its effectiveness relies on the availability of large labeled datasets, which can be hard to obtain. Recently, Transfer Learning has been proposed to resolve this issue by leveraging pre-existing knowledge from pre trained models on millions of raw data. Some attempts of transferring knowledge from computer vision to remote sensing have achieved acceptable results. However, those models may miss the rich spectrum information necessary for precise results and are not well suited for multispectral imageries. Additionally, transfer learning in remote sensing applications is extremely dependent on the selection of data from the source model, which has a significant impact on the target model. This study investigates the importance of selecting the appropriate source dataset and task for transfer learning. To this end, pretrained model on Landsat imageries have been finetuned on sentinel-2 data. Separated pre-trained models on sentinel-2 imageries for palm/building classifications have been employed to explore the impact of the source classes on the performance of the target class. The results have been evaluated by metrics such as Precision, Kappa, Recall and F1-score to assess the model's performances where results from train/finetune on similar data source (sentinel-2) reaches higher accuracy compared to multisource training/finetuning. This study illustrates how selecting appropriate source datasets and tasks can directly influence the performance of the target model.

Ensuring access to safe drinking water is a critical concern, particularly in regions with limited resources. This study evaluates groundwater potability using a range of machine learning models, including logistic regression, K-Nearest Neighbors (KNN), Support Vector Classifier (SVC), and Random Forest, as well as deep learning models such as Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Feedforward Neural Networks (FNN), and Long Short-Term Memory (LSTM). We collected thirty groundwater samples from residential and industrial locations in Jaen, Kano State, Nigeria, focusing on nine crucial physicochemical parameters: electric conductivity, pH, total dissolved solids, calcium, magnesium, chloride, zinc, manganese, and copper. Machine learning models, such as logistic regression and random forest, achieved accuracy scores of 0.833. They were closely followed by deep learning models, such as ANN with an accuracy score of 0.833, and LSTM, which scored 0.666. KNN and SVC provided moderately accurate predictions, scoring 0.667, while CNN and FNN achieved lower scores of 0.333 and 0.5, respectively. This study represents a significant step toward ensuring safe drinking water for communities and preserving the sustainability of natural resources.

This paper proposes a compressed Convolutional Neural Network (CNN) model for rice yield detection using the weight pruning technique. The initial CNN model achieved an accuracy of 91% on a dataset comprising 3120 images of both yield and unyield rice crops. However, it had a large size of approximately 603MB, posing challenges in terms of deployment and storage. To address this issue, weight pruning was applied to compress the model. The compressed model achieved a significant reduction in size to 186MB, representing a reduction of approximately 69.15%, while maintaining a reasonable accuracy of 86%.

The experiment was conducted in three phases. First, a dataset of 3120 images of yield and unyield rice crops was collected from different farms in Kano metropolis of Nigeria and preprocessed by resizing them to 250x250 pixels. Secondly, a CNN model with 12 layers was designed and trained using the preprocessed dataset. The model achieved an accuracy of 86%. Finally, weight pruning was applied to the trained CNN model to reduce its size. The compressed model exhibited a size of 300MB and an accuracy of 86%.

The results of this study demonstrate the effectiveness of weight pruning as a viable technique for compressing CNN models without significantly compromising their accuracy. The compressed model, with its reduced size, is well-suited for deployment on resource-constrained devices for rice yield detection applications.

Road infrastructure is critical to a nation's prosperity and safety. However, the presence of potholes in road networks creates substantial issues to road users, resulting to an increase in accidents and costly vehicle damages. Real-time pothole anomaly detection systems have emerged as a possible solution to this problem, employing innovative technology for fast identification and notifications about pothole existence. In underdeveloped countries with limited road maintenance resources, such technologies have the potential to improve road safety while lowering maintenance costs. This research provides a comprehensive assessment of various implementation options for real-time pothole anomaly detection in developing countries. It investigates the many strategies and technologies that can be used in developing countries to detect anomalies in the road network. The utilisation of deep learning, computer vision, and lidar-based systems is highlighted in particular. Furthermore, the paper addresses the obstacles associated with the deployment of such systems and provides alternative solutions. Additionally, the paper compares the different alternatives, discussing their potential benefits and drawbacks. The findings of the literature analysis and practical evidence reveal that, while deep learning and computer vision-based algorithms produce the most accurate results, their application is limited due to computational and economical constraints in developing countries such as Nigeria. On the contrary, lidar-based solutions offer a realistic and cost-effective alternative to deep learning and computer vision-based systems. Thus, lidar-based pothole detection technologies can be used efficiently to achieve safer roadways in underdeveloped countries such as Nigeria.

To meet the market demand for high-quality products, researchers and manufacturers have invested in the development of accurate methods for estimating shelf-life. Tests that consider the simultaneous effects of different parameters on food degradation are useful tools in shelf-life studies, as these parameters can directly influence quality and safety. With this in mind, the objective of this review is to gather pertinent information from recent studies (2006-2022) pertaining to multivariate analysis applied in accelerated shelf-life tests in order to facilitate a comprehensive understanding. The review focuses on multivariate techniques commonly employed in accelerated shelf-life modeling, namely, principal components analysis, partial least squares regression, orthogonal projections to latent structures discriminant analysis, and hierarchical cluster analysis. Through an extensive literature review, the collected data represent the evolution of these methods, taking into account current trends, advances in food shelf-life techniques, and future perspectives. It was observed that the recent literature provides limited information on the determination of shelf-life under multiple accelerated factors. However, the studies analyzed showed that multivariate analysis can be a useful tool in the interpretation of quality characteristics and can accurately predict the shelf-life of foods compared to univariate kinetic procedures. Multivariate statistical methods addressed in this work are presented as a promising method for foods tested, being applied together with different chemometric techniques. This comprehensive review contributes to the body of knowledge surrounding accelerated shelf-life testing, offering valuable insights for researchers, manufacturers, and stakeholders in the food industry.

Artificial Intelligence (IA) is a broad term that usually refers to a diverse set of computational procedures that can mimic human decisions and/or processes so closely that they appear intelligent, being able, in example, to process with enormous quickly large volumes of data. AI is such a powerful tool, that organizations are increasingly using it in various areas, including Human Resources (HR) Management, especially in recruitment and selection functions. For instance, big data algorithms are highly instrumental in expanding the process of candidates searching. However, there are several key questions that remain open, such as, ethical issues and the reactions and attitudes towards the AI of its users (recruiters, selection managers and potential candidates), being needed a greater empirical and systematic review effort of the literature at this level. In this context, this paper discusses AI and its applications in HR recruitment and selection process, addressing the future trends and challenges defined in the extant literature.

The inadequate handling of winery wastewater (WW) has left a harmful impression on the sustainable production of wine ¹. Usually, for each litre of wine produced, between 0.2 – 4L of wastewater can be generated, being rich in biodegradable organic matter with typical chemical oxygen demand (COD) concentration varying between 0.3 – 49 g L^{-1 2–4}. Preventing the discharge of WW into surface waters is crucial to mitigate the risks of eutrophication. Biological treatment is widely recognized as a cost-effective approach to treat wastewater ⁵. However, the selection of a specific biological method for WW treatment depends on various factors, including the size and geographical location of the winery, the volume of wastewater generated, and its organic content ⁴. In this study, the effectiveness of using a sequencing batch reactor (SBR) as a biological technology for WW treatment was investigated. COD removal efficiencies ranging from 70 to 95% were obtained (without nutrient supplementation), for organic loading rates (OLR) below 3.4 g_COD L^-1 d^-1 and a hydraulic retention time of 21 h. The best sludge settling properties, (i.e., 78 mL g_TSS^-1 and 1.1 cm min^-1) were observed at an OLR of 6 g_COD L^-1 d^-1. Despite these good settling properties, only 70% COD removal efficiency was attained, suggesting the need for nutrient supplementation for higher OLR. In fact, the addition of nutrient allowed the improvement of both effluent quality and settling properties for similar OLR. Overall, these findings demonstrate the potential of SBR as a biological WW treatment process.

References

1. Ngwenya, N., Gaszynski, C. & Ikumi, D. A review of winery wastewater treatment: A focus on UASB biotechnology optimisation and recovery strategies. J. Environ. Chem. Eng. 10, 108172 (2022).

2. Amor, C., Marchão, L., Lucas, M. S. & Peres, J. A. Application of advanced oxidation processes for the treatment of recalcitrant agro-industrial wastewater: A review. Water (Switzerland) 11, (2019).

3. Welz, P. J., Holtman, G., Haldenwang, R. & Le Roes-Hill, M. Characterisation of winery wastewater from continuous flow settling basins and waste stabilisation ponds over the course of 1 year: Implications for biological wastewater treatment and land application. Water Sci. Technol. 74, 2036–2050 (2016).

4. Pirra, A. J. D. Caracterização e Tratamento de Efluentes Vinícolas da Região Demarcada do Douro. (PhD Thesis, University of Trás-os-Montes e Alto Douro, 2005).

5. Lemaire, R., Webb, R. I. & Yuan, Z. Micro-scale observations of the structure of aerobic microbial granules used for the treatment of nutrient-rich industrial wastewater. ISME J. 2, 528–541 (2008).

Self-compacting concrete (SCC) is an innovative building material developed to have special properties such as increased flowability, good segregation resistance and compaction without vibration etc. Despite the benefits of SCC over conventional concrete, there are very few methods reported in the literature that can predict the SCC compressive strength accurately. Thus, to foster the utilization of SCC in construction industry, an inventive machine learning technique named Multi Expression Programming (MEP) is employed to forecast the SCC 28-day compressive strength. A database consisting of 231 compressive strength results is constructed using extensive literature search. The resulting equation obtained by employing MEP algorithm relates the compressive strength of SCC with six most influential input parameters i.e., water-cement ratio, amount of fly ash and silica fume, quantities of fine and coarse aggregate and superplasticizer dosage. The database is split into training and validation datasets used for training and validation of the algorithm respectively. The accuracy of MEP algorithm is verified by means of four statistical error metrices: mean absolute error (MAE), root mean square error (RMSE), coefficient of correlation (R) and coefficient of determination ( ). The results revealed that the errors are within the prescribed limits for both training and validation sets and the developed equation have excellent generalization capacity. This is also verified from the scatter and series plots of the training and validation datasets. Thus, the developed equation can be used practically to forecast the strength of SCC containing fly ash and silica fume.