Latest developments of generative artificial intelligence and applications in ophthalmology.

The emergence of generative artificial intelligence (AI) has revolutionized various fields. In ophthalmology, generative AI has the potential to enhance efficiency, accuracy, personalization and innovation in clinical practice and medical research, through processing data, streamlining medical documentation, facilitating patient-doctor communication, aiding in clinical decision-making, and simulating clinical trials. This review focuses on the development and integration of generative AI models into clinical workflows and scientific research of ophthalmology. It outlines the need for development of a standard framework for comprehensive assessments, robust evidence, and exploration of the potential of multimodal capabilities and intelligent agents. Additionally, the review addresses the risks in AI model development and application in clinical service and research of ophthalmology, including data privacy, data bias, adaptation friction, over interdependence, and job replacement, based on which we summarized a risk management framework to mitigate these concerns. This review highlights the transformative potential of generative AI in enhancing patient care, improving operational efficiency in the clinical service and research in ophthalmology. It also advocates for a balanced approach to its adoption.

Construction of a Nomogram Model to Identify Atherosclerotic Cardiovascular Disease in Patients with Type 2 Diabetes Mellitus.

BACKGROUND: Approximately 50.0% of patients with type 2 diabetes mellitus (T2DM) experience macrovascular diseases, and nearly 80.0% of them succumb to macrovascular complications. Atherosclerotic cardiovascular disease (ASCVD) ranks among the most prevalent macrovascular complications in T2DM. In this study, we aim to develop a nomogram model for the early detection of ASCVD in T2DM patients, enabling us to provide valuable recommendations for the clinical prevention and management of macrovascular complications in this patient population. METHODS: This retrospective analysis encompassed 2620 T2DM patients admitted between June 2015 and June 2021. The cohort comprised 1270 T2DM patients with coexisting ASCVD (referred to as the "ASCVD group") and 1350 individuals who did not experience ASCVD (the "non-ASCVD group"). We conducted a comparative assessment of their baseline characteristics and clinical data. A nomogram model for the identification of ASCVD in T2DM patients was constructed utilizing Logistic regression analysis and the R package. The model's performance was evaluated through receiver operating characteristic (ROC) curve analysis and calibration curves. RESULTS: We developed a nomogram model for the identification of ASCVD in T2DM patients, incorporating ten variables: sex, age, hypertension, smoking history, low-density lipoprotein cholesterol/high-density lipoprotein cholesterol (LDL-C/HDL-C) ratio, alanine transaminase (ALT), adenosine deaminase (ADA), postprandial 2-hour C-peptide, monocyte count (MONO), and eosinophil count (EOS). ROC curves demonstrated that the area under the curve (AUC) of the nomogram model for identifying ASCVD in T2DM patients was 0.673 for the training dataset (with a cut-off value of 0.473, specificity of 0.629, and sensitivity of 0.637) and 0.655 for the validation dataset (with a cut-off value of 0.460, specificity of 0.605, and sensitivity of 0.675). The calibration curve indicated a substantial agreement between the predicted and observed cases of ASCVD in the training dataset and an acceptable level of agreement in the validation dataset. CONCLUSIONS: The nomogram model effectively identifies ASCVD in T2DM patients, which can be instrumental in pinpointing the high-risk population for ASCVD among T2DM patients and facilitating timely clinical management.

Masked Graph Neural Networks for Unsupervised Anomaly Detection in Multivariate Time Series.

Anomaly detection has been widely used in grid operation and maintenance, machine fault detection, and so on. In these applications, the multivariate time-series data from multiple sensors with latent relationships are always high-dimensional, which makes multivariate time-series anomaly detection particularly challenging. In existing unsupervised anomaly detection methods for multivariate time series, it is difficult to capture the complex associations among multiple sensors. Graph neural networks (GNNs) can model complex relations in the form of a graph, but the observed time-series data from multiple sensors lack explicit graph structures. GNNs cannot automatically learn the complex correlations in the multivariate time-series data or make good use of the latent relationships among time-series data. In this paper, we propose a new method-masked graph neural networks for unsupervised anomaly detection (MGUAD). MGUAD can learn the structure of the unobserved causality among sensors to detect anomalies. To robustly learn the temporal context from adjacent time points of time-series data from the same sensor, MGUAD randomly masks some points of the time-series data from the sensor and reconstructs the masked time points. Similarly, to robustly learn the graph-level context from adjacent nodes or edges in the relation graph of multivariate time series, MGUAD masks some nodes or edges in the graph under the framework of a GNN. Comprehensive experiments are conducted on three public datasets. According to the experimental findings, MGUAD outperforms state-of-the-art anomaly detection methods.

China's investments in renewable energy through the belt and road initiative stimulated local economy and employment: A case study of Pakistan.

Since China's announcement of the Belt and Road Initiative (BRI) in 2015, much focus has been drawn on the environmental impacts of China's energy investments in the countries along the BRI. The economic and social impacts of these investments, which are also important for the wellbeing for local people, left largely uninvestigated. In this paper, we used China's renewable energy investments in Pakistan as a case study to investigate the contributions of these investments on local economy and employment. Through IO table analysis, we found that the 28 renewable power plant projects invested by China till now potentially provided 8905 jobs and generated around USD 39.8 million production values in related sectors in Pakistan, including USD 30.7 million from wind power plants development and 9.1 million from solar. When Chinese companies act as engineers and constructors, the increase of production value in relevant sectors in Pakistan (USD 6.05 million per 100 MW) are higher than wind power plant projects with other magnitude of engagement (3.82 million as a fully sponsor, 4.19 million as only finance supporter and 2.29 as equipment provider). Wind power plants will create more jobs and increase more production values than solar power plants. This study identifies the economic and social benefits of BRI renewable energy investments from China and the driving mechanism, thus providing basis for promoting renewable energy investments in countries like Pakistan so that they can gain new drive for social and economic growth from the global trend of low carbon transition.

Embedded Sensing System for Recognizing Citrus Flowers Using Cascaded Fusion YOLOv4-CF + FPGA.

Florescence information monitoring is essential for strengthening orchard management activities, such as flower thinning, fruit protection, and pest control. A lightweight object recognition model using cascade fusion YOLOv4-CF is proposed, which recognizes multi-type objects in their natural environments, such as citrus buds, citrus flowers, and gray mold. The proposed model has an excellent representation capability with an improved cascade fusion network and a multi-scale feature fusion block. Moreover, separable deep convolution blocks were employed to enhance object feature information and reduce model computation. Further, channel shuffling was used to address missing recognition in the dense distribution of object groups. Finally, an embedded sensing system for recognizing citrus flowers was designed by quantitatively applying the proposed YOLOv4-CF model to an FPGA platform. The mAP@.5 of citrus buds, citrus flowers, and gray mold obtained on the server using the proposed YOLOv4-CF model was 95.03%, and the model size of YOLOv4-CF + FPGA was 5.96 MB, which was 74.57% less than the YOLOv4-CF model. The FPGA side had a frame rate of 30 FPS; thus, the embedded sensing system could meet the demands of florescence information in real-time monitoring.

Green Citrus Detection and Counting in Orchards Based on YOLOv5-CS and AI Edge System.

Green citrus detection in citrus orchards provides reliable support for production management chains, such as fruit thinning, sunburn prevention and yield estimation. In this paper, we proposed a lightweight object detection YOLOv5-CS (Citrus Sort) model to realize object detection and the accurate counting of green citrus in the natural environment. First, we employ image rotation codes to improve the generalization ability of the model. Second, in the backbone, a convolutional layer is replaced by a convolutional block attention module, and a detection layer is embedded to improve the detection accuracy of the little citrus. Third, both the loss function CIoU (Complete Intersection over Union) and cosine annealing algorithm are used to get the better training effect of the model. Finally, our model is migrated and deployed to the AI (Artificial Intelligence) edge system. Furthermore, we apply the scene segmentation method using the "virtual region" to achieve accurate counting of the green citrus, thereby forming an embedded system of green citrus counting by edge computing. The results show that the mAP@.5 of the YOLOv5-CS model for green citrus was 98.23%, and the recall is 97.66%. The inference speed of YOLOv5-CS detecting a picture on the server is 0.017 s, and the inference speed on Nvidia Jetson Xavier NX is 0.037 s. The detection and counting frame rate of the AI edge system-side counting system is 28 FPS, which meets the counting requirements of green citrus.

Relation Between Gut Microbiota Composition and Traditional Spontaneous Fermented Dairy Foods Among Kazakh Nomads in Xinjiang, China.

The human gastrointestinal tract represents one of the most densely populated microbial ecosystems studied to date. Although this microbial consortium has been recognized to have a crucial impact on human health, its precise composition is still subject to intense investigation, as people from different regions have different gut microbiota structures. The Kazakh nomads in Xinjiang, China still retain their nomadic lifestyle and traditional diet. Their specific diet style and ancient genetic background shaped their gut microbiota to contain unique characteristics. In present study, the compositions of the gut microbiota and fermented dairy foods were assessed by high-throughput sequencing of the 16S rRNA gene. Twenty-nine Kazakh nomads were recruited, and 33 traditional fermented dairy foods were collected from five pasturing areas (Buerjin, Zhaosu, Nilka, Tekes, and Fuhai) in northern Xinjiang, China. The correlation of the physical index with the gut microbiota was also analyzed. The unique diet style of Kazakh may be a critical factor in keeping their gut microbiota in a balanced state and help them to remain in good health. PRACTICAL APPLICATION: This research shows that the consumption of spontaneous fermented dairy food plays an important role in increasing gut microbial diversity. Some probiotics in fermented dairy food, such as Bifidobacterium and Lactobacillus, have positive correlation with human body health index such as body mass index and blood glucose. These may provide some theoretical supports to adjuvant therapy of obesity and diabetes through scientific dietary intervention.

Characterization of bacteria and yeasts isolated from traditional fermentation starter (Fen-Daqu) through a 1H NMR-based metabolomics approach.

Daqu is a traditional fermentation starter for the production of baijiu and vinegar. It is an important saccharifying and fermenting agent associated with alcoholic fermentation and also a determining factor for the flavour development of these products. Bacterial and yeast isolates from a traditional fermentation starter (Fen-Daqu) were examined for their amylolytic activity, ethanol tolerance and metabolite production during sorghum-based laboratory-scale alcoholic fermentation. The selected strains (Bacillus licheniformis, Pediococcus pentosaceus, Lactobacillus plantarum, Pichia kudriavzevii, Wickerhamomyces anomalus, Saccharomyces cerevisiae, and Saccharomycopsis fibuligera) were blended in different combinations, omitting one particular strain in each mixture. 1H nuclear magnetic resonance (NMR) spectroscopy coupled with multivariate statistical analysis was used to investigate the influence of the selected strains on the metabolic changes observed under the different laboratory-controlled fermentation conditions. Principal component analysis showed differences in the metabolites produced by different mixtures of pure cultures. S. cerevisiae was found to be superior to other species with respect to ethanol production. S. fibuligera and B. licheniformis converted starch or polysaccharides to soluble sugars. Lactic acid bacteria had high amylolytic and proteolytic activities, thereby contributing to increased saccharification and protein degradation. W. anomalus was found to have a positive effect on the flavour of the Daqu-derived product. This study highlights the specific functions of S. cerevisiae, S. fibuligera, B. licheniformis, W. anomalus and lactic acid bacteria in the production of light-flavour baijiu (fen-jiu). Our results show that all investigated species deliver an important contribution to the functionality of the fermentation starter Daqu.