Analyzing COVID-19 Medical Papers Using Artificial Intelligence: Insights for Researchers and Medical Professionals

Since the beginning of the COVID-19 pandemic almost two years ago, there have been more than 700,000 scientific papers published on the subject. An individual researcher cannot possibly get acquainted with such a huge text corpus and, therefore, some help from artificial intelligence (AI) is highly needed. We propose the AI-based tool to help researchers navigate the medical papers collections in a meaningful way and extract some knowledge from scientific COVID-19 papers. The main idea of our approach is to get as much semi-structured information from text corpus as possible, using named entity recognition (NER) with a model called PubMedBERT and Text Analytics for Health service, then store the data into NoSQL database for further fast processing and insights generation. Additionally, the contexts in which the entities were used (neutral or negative) are determined. Application of NLP and text-based emotion detection (TBED) methods to COVID-19 text corpus allows us to gain insights on important issues of diagnosis and treatment (such as changes in medical treatment over time, joint treatment strategies using several medications, and the connection between signs and symptoms of coronavirus, etc.).

UAV-BASED 3D VIRTUAL TOUR CREATION

Lately, improvements in game engines have increased the interest in virtual reality (VR) technologies, that engages users with an artificial environment, and have led to the adoption of VR systems to display geospatial data. Because of the ongoing COVID-19 pandemic, and thus the necessity to stay at home, VR tours became very popular. In this paper, we tried to create a three-dimensional (3D) virtual tour for Gebze Technical University (GTU) Southern Campus by transferring high-resolution unmanned air vehicle (UAV) data into a virtual domain. UAV data is preferred in various applications because of its high spatial resolution, low cost and fast processing time. In this application, the study area was captured from different modes and altitudes of UAV flights with a minimum ground sampling distance (GSD) of 2.18 cm using a 20 MP digital camera. The UAV data was processed in Structure from Motion (SfM) based photogrammetric evaluation software Agisoft Metashape and high-quality 3D textured mesh models were generated. Image orientation was completed using an optimal number of ground control points (GCPs), and the geometric accuracy was calculated as ±8 mm (~0.4 pixels). To create the VR tour, UAV-based mesh models were transferred into the Unity game engine and optimization processes were carried out by applying occlusion culling and space subdivision algorithms. To improve the visualization, 3D object models such as trees, lighting poles and arbours were positioned on VR. Finally, textual metadata about buildings and a player with a first-person camera were added for an informative VR experience.

Fast Processing RNA-Seq on Multicore Processor

RNA Sequencing (RNA-Seq) is the sequencing and analysis of transcriptomes. The main purpose of RNA-Seq analysis is to find out the presence and quantity of RNA in an experimental sample under a specific condition. Essentially, RNA raw sequence data was massive. It can be as big as hundreds of Gigabytes (GB). This massive data always makes the processing time become longer and take several days. A multicore processor can speed up a program by separating the tasks and running the tasks’ errands concurrently. Hence, a multicore processor will be a suitable choice to overcome this problem. Therefore, this study aims to use an Intel multicore processor to improve the RNA-Seq speed and analyze RNA-Seq analysis's performance with a multiprocessor. This study only processed RNA-Seq from quality control analysis until sorted the BAM (Binary Alignment/Map) file content. Three different sizes of RNA paired end has been used to make the comparison. The final experiment results showed that the implementation of RNA-Seq on an Intel multicore processor could achieve a higher speedup. The total processing time of RNA-Seq with the largest size of RNA raw sequence data (66.3 Megabytes) decreased from 317.638 seconds to 211.916 seconds. The reduced processing time was 105 seconds and near to 2 minutes. Furthermore, for the smallest RNA raw sequence data size, the total processing time decreased from 212.380 seconds to 163.961 seconds which reduced 48 seconds.

Efficient attention-based deep encoder and decoder for automatic crack segmentation

Recently, crack segmentation studies have been investigated using deep convolutional neural networks. However, significant deficiencies remain in the preparation of ground truth data, consideration of complex scenes, development of an object-specific network for crack segmentation, and use of an evaluation method, among other issues. In this paper, a novel semantic transformer representation network (STRNet) is developed for crack segmentation at the pixel level in complex scenes in a real-time manner. STRNet is composed of a squeeze and excitation attention-based encoder, a multi head attention-based decoder, coarse upsampling, a focal-Tversky loss function, and a learnable swish activation function to design the network concisely by keeping its fast-processing speed. A method for evaluating the level of complexity of image scenes was also proposed. The proposed network is trained with 1203 images with further extensive synthesis-based augmentation, and it is investigated with 545 testing images (1280 × 720, 1024 × 512); it achieves 91.7%, 92.7%, 92.2%, and 92.6% in terms of precision, recall, F1 score, and mIoU (mean intersection over union), respectively. Its performance is compared with those of recently developed advanced networks (Attention U-net, CrackSegNet, Deeplab V3+, FPHBN, and Unet++), with STRNet showing the best performance in the evaluation metrics-it achieves the fastest processing at 49.2 frames per second.

Prediction of COVID-19 pandemic dynamics in Russia based on simple mathematical models of epidemics

Introduction: The COVID-19 pandemic which began in 2020 and has taken more than five million lives has become a threat to the very existence of mankind. Therefore, predicting the spread of COVID-19 in each individual country is a very urgent task. The complexity of its solution is due to the requirement for fast processing of large amounts of data and the fact that the data are mostly inaccurate and do not have the statistical properties necessary for the successful application of statistical methods. Therefore, it seems important to develop simple forecasting methods based on classical simple models of epidemiology which are only weakly sensitive to data inaccuracies. It is also important to demonstrate the feasibility of the approach in relation to the incidence data in Russia. Purpose: Obtaining forecast data based on classical simple models of epidemics, namely SIR and SEIR. Methods: For discrete versions of SIR and SEIR models, it is proposed to estimate the parameters of the models using a reduced version of the least squares method, and apply a scenario approach to the forecasting. The simplicity and a small number of parameters are the advantages of SIR and SEIR models, which is very important in the context of a lack of numerical input data and structural incompleteness of the models. Results: A forecast of the spread of COVID-19 in Russia has been built based on published data on the incidence from March 10 to April 20, 2020, and then, selectively, according to October 2020 data and October 2021 data. The results of the comparison between SIR and SEIR forecasts are presented. The same method was used to construct and present forecasts based on morbidity data in the fall of 2020 and in the fall of 2021 for Russia and for St. Petersburg. To set the parameters of the models which are difficult to determine from the official data, a scenario approach is used: the dynamics of the epidemic is analyzed for several possible values of the parameters. Practical relevance: The results obtained show that the proposed method predicts well the time of the onset of the peak incidence, despite the inaccuracy of the initial data.

Adaptive Language Processing Unit for Malaysian Sign Language Synthesizer

Language Processing Unit (LPU) is a system built to process text-based data to comply with the rules of sign language grammar. This system was developed as an important part of the sign language synthesizer system. Sign language (SL) uses different grammatical rules from the spoken/verbal language, which only involves the important words that Hearing/Impaired Speech people can understand. Therefore, it needs word classification by LPU to determine grammatically processed sentences for the sign language synthesizer. However, the existing language processing unit in SL synthesizers suffers time lagging and complexity problems, resulting in high processing time. The two features, i.e., the computational time and success rate, become trade-offs which means the processing time becomes longer to achieve a higher success rate. This paper proposes an adaptive Language Processing Unit (LPU) that allows processing the words from spoken words to Malaysian SL grammatical rule that results in relatively fast processing time and a good success rate. It involves n-grams, NLP, and Hidden Markov Models (HMM)/Bayesian Networks as the classifier to process the text-based input. As a result, the proposed LPU system has successfully provided an efficient (fast) processing time and a good success rate compared to LPU with other edit distances (Mahalanobis, Levensthein, and Soundex). The system has been tested on 130 text-input sentences with several words ranging from 3 to 10 words. Results showed that the proposed LPU could achieve around 1.497ms processing time with an average success rate of 84.23% for a maximum of ten-word sentences.

Sustainability of an Integrated E-Payment System in Developing Countries

Over the recent years, there has been tremendous ease in monetary transactions; all thanks to a convenient payment platform. The Integrated e-payment system (IEPS) facilitates financial transactions electronically.This study aims to examine the factorial models necessitating the continuous use of the integrated e-payment system in the light of the Unified Theory of Acceptance and Use of Technology (UTAUT) with two additional variables; processing time and processing charges proposed. The study data were collected from 285 valid respondents in Ghana through the random sampling approach. The hypothesis verification was conducted via the SmartPLS. The result of the structural equation modelling indicate that processing charges, processing time, and social influence are the critical influencers of IEPS continuous usage intention. In addition, it also implies that financial institutions and banks liaise with designers to, perhaps, consider consumers’ perspectives in their design; such that will bring about convenience, fast processing and minimal cost in order to foster continuous use.

Informal Lending Practices and Profitability of Rice Retailers in Borongan City

This correlational study explored the relationship between informal lender’s practices and profitability of fifty-three (53) rice retailers in city of Borongan using a researcher-made-three part questionnaire on the business profile, level of practices among informal lenders and weekly income of the respondents. Data were sourced out from rice retailers who lent sum of money from “5-6” informal lenders, and were utilized using descriptive and inferential analysis test at .05 level of significance. Findings revealed that Majority of the respondents had been in the middle years of retailing business, used an average initial capital sourced out from their personal savings and are earning an average weekly income. The respondents’ perceived that informal lenders assess clients’ financial needs and willingness to pay without requiring collateral from them and that rice-retailers reloan due to fast processing of funds and the great customer service. Using regression analysis, year in business was found to be a significant predictor of increase in the profitability of rice retailing services. While, no significant relationship was seen between the rice retailers’ weekly net income and the level of practice among informal lenders. The researcher recommends that government and non-government agencies to provide intervention that will improve the microbusiness enterprise in the market using the help of informal and formal lending institutions and that wider scope of the study should be made to include other small business sectors that are affected by informal lenders.

Rodent Area Prostriata Converges Multimodal Hierarchical Inputs and Projects to the Structures Important for Visuomotor Behaviors

Area prostriata is a limbic structure critical to fast processing of moving stimuli in far peripheral visual field. Neural substrates underlying this function remain to be discovered. Using both retrograde and anterograde tracing methods, the present study reveals that the prostriata in rat and mouse receives inputs from multimodal hierarchical cortical areas such as primary, secondary, and association visual and auditory cortices and subcortical regions such as the anterior and midline thalamic nuclei and claustrum. Surprisingly, the prostriata also receives strong afferents directly from the rostral part of the dorsal lateral geniculate nucleus. This shortcut pathway probably serves as one of the shortest circuits for fast processing of the peripheral vision and unconscious blindsight since it bypasses the primary visual cortex. The outputs of the prostriata mainly target the presubiculum (including postsubiculum), pulvinar, ventral lateral geniculate nucleus, lateral dorsal thalamic nucleus, and zona incerta as well as the pontine and pretectal nuclei, most of which are heavily involved in subcortical visuomotor functions. Taken together, these results suggest that the prostriata is poised to quickly receive and analyze peripheral visual and other related information and timely initiates and modulates adaptive visuomotor behaviors, particularly in response to unexpected quickly looming threats.