The deadly coronavirus virus (COVID‐19) was confirmed as a pandemic by the World Health Organization (WHO) in December 2019. It is important to identify suspected patients as early as possible in order to control the spread of the virus, improve the efficacy of medical treatment, and, as a result, lower the mortality rate. The adopted method of detecting COVID‐19 is the reverse‐transcription polymerase chain reaction (RT‐PCR), the process is affected by a scarcity of RT‐PCR kits as well as its complexities. Medical imaging using machine learning and deep learning has proved to be one of the most efficient methods of detecting respiratory diseases, but to train machine learning features needs to be extracted manually, and in deep learning, efficiency is affected by deep learning architecture and low data. In this study, handcrafted local binary pattern (LBP) and automatic seven deep learning models extracted features were used to train support vector machines (SVM) and K‐nearest neighbour (KNN) classifiers, to improve the performance of the classifier, a concatenated LBP and deep learning feature was proposed to train the KNN and SVM, based on the performance criteria, the models VGG‐19 + LBP achieved the highest accuracy of 99.4%. The SVM and KNN classifiers trained on the hybrid feature outperform the state of the art model. This shows that the proposed feature can improve the performance of the classifiers in detecting COVID‐19.

Tuberculosis (TB) a disease caused by Mycobacterium tuberculosis (Mtb) poses a significant threat to human life, and current BCG vaccinations only provide sporadic protection, therefore there is a need for developing efficient vaccines. Numerous immunoinformatic methods have been utilized previously, here for the first time a deep learning framework based on Deconvolutional Neural Networks (DCNN) and Bidirectional Long Short-Term Memory (DCNN-BiLSTM) was used to predict Mtb Multiepitope vaccine (MtbMEV) subunits against six Mtb H37Rv proteins. The trained model was used to design MEV within a few minutes against TB better than other machine learning models with 99.5% accuracy. The MEV has good antigenicity, and physiochemical properties, and is thermostable, soluble, and hydrophilic. The vaccine's BLAST search ruled out the possibility of autoimmune reactions. The secondary structure analysis revealed 87% coil, 10% beta, and 2% alpha helix, while the tertiary structure was highly upgraded after refinement. Molecular docking with TLR3 and TLR4 receptors showed good binding, indicating high immune reactions. Immune response simulation confirmed the generation of innate and adaptive responses. In-silico cloning revealed the vaccine is highly expressed in E. coli. The results can be further experimentally verified using various analyses to establish a candidate vaccine for future clinical trials.

Smart tourism is a developing industry, and numerous nations are planning to establish smart cities in which technology is employed to make life easier and link nearly everything. Many researchers have created object detectors; however, there is a demand for lightweight versions that can fit into smartphones and other edge devices. The goal of this research is to demonstrate the notion of employing a mobile application that can detect statues efficiently on mobile applications, and also improve the performance of the models by employing the Gaussian Smoothing Filter (GSF). In this study, three object detection models, EfficientDet—D0, EfficientDet—D2 and EfficientDet—D4, were trained on original and smoothened images; moreover, their performance was compared to find a model efficient detection score that is easy to run on a mobile phone. EfficientDet—D4, trained on smoothened images, achieves a Mean Average Precision (mAP) of 0.811, an mAP-50 of 1 and an mAP-75 of 0.90.

It is essential to use highly antigenic epitope areas, since the development of peptide vaccines heavily relies on the precise design of epitope regions that can elicit a strong immune response. Choosing epitope regions experimentally for the production of the SARS-CoV-2 vaccine can be time-consuming, costly, and labor-intensive. Scientists have created in silico prediction techniques based on machine learning to find these regions, to cut down the number of candidate epitopes that might be tested in experiments, and, as a result, to lessen the time-consuming process of their mapping. However, the tools and approaches involved continue to have low accuracy. In this work, we propose a hybrid deep learning model based on a convolutional neural network (CNN) and long short-term memory (LSTM) for the classification of peptides into epitopes or non-epitopes. Numerous transfer learning strategies were utilized, and the fine-tuned method gave the best result, with an AUC of 0.979, an f1 score of 0.902, and 95.1% accuracy, which was far better than the performance of the model trained from scratch. The experimental results obtained show that this model has superior performance when compared to other methods trained on IEDB datasets. Using bioinformatics tools such as ToxinPred, VaxiJen, and AllerTop2.0, the toxicities, antigenicities, and allergenicities, respectively, of the predicted epitopes were determined. In silico cloning and codon optimization were used to successfully express the vaccine in E. coli. This work will help scientists choose the best epitope for the development of the COVID-19 vaccine, reducing cost and labor and thereby accelerating vaccine production.

On device machine learning has in recent times attracted growing attention and applications, ranging from advanced OCR and handwriting recognition, smart sleep tracking in Apple iOS, mobile virtual assistants and Image recognition, text translation in android; and more. As a result, there is increasing demand for user privacy and security in machine learning applications. Federated Learning (FL), a new method, allows users to cooperatively train a global model while keeping the data local, thereby resolving the issue of data privacy protection through the encryption mechanism. The server aggregates models until convergence while the clients train their local models. In this procedure, the server employs a reward system to persuade users to provide huge amounts of high-quality data in order to enhance the overall model. This paper explores a few FL techniques that can be applied alongside on-device learning, in order to achieve better privacy and security.