results than PCR [4]. The researchers also mentioned that Compared to RT-PCR, chest CT imaging might be a more reliable, practical and a rapid method to diagnose and assess COVID-19, especially in the epidemic area.Although many centers recommend the use of CT scans on Xrays, research [5] has revealed that Xrays can effectively predict which patients are more likely to develop their symptoms, especially for people aged 21-50, therefore early detection of COVID-19 using Xrays scans is very critic, and can help quickly isolating the infected people, as presented in "Fig. III-B", where the virus damages occurs differs day by day in the patient's lungs. It is also noted that SARS-CoV-2 Pneumonia should be initially distinguished from other pneumonia diseases since they present quite similar scans in some cases. Therefore, it is important to establish an automated tool that helps doctors in detecting COVID-19 on chest scans. Deep learning, in particular convolutional networks, has demonstrated state-of-art achievements in medical imaging analysis in many application areas, such as neurological disorders, retinal diseases analysis [6], pulmonary infections, digital pathology, breast anatomy imaging, cardiology, abdominal treatment, and musculoskeletal disorders [7]. Typically, convolutional networks perform better on large scale datasets; therefore, in case of small scale dataset, one way to overcome this deficiency is to use transfer learning. In the latter, a model is trained on one task (where a large scale dataset is available) is fine-tuned on the second task (where only a small sale dataset is available). Most traditional learning methods build and train new baselines and classifiers from scratch for each classification task [8]. On the other hand, Transfer learning re-use and transfer knowledge learnt from a source classifier that have been trained on large scale databases to simplify the construction of a classifier for a new target Abstract-Since the novel coronavirus SARS-CoV-2 outbreak, intensive research has been conducted to find s uitable tools for diagnosis and identifying infected people in order to take appropriate action. Chest imaging plays a significant role i n this phase where CT and Xrays scans have proven to be effective in detecting COVID-19 within the lungs. In this research, we propose deep learning models using Transfer learning to detect COVID-19. Both X-ray and CT scans were considered to evaluate the proposed methods.

No abstract

This paper tackles the problem of sentiment analysis in the Arabic language where a new deep learning model has been put forward. The proposed model uses a hybrid bidirectional gated recurrent unit (BiGRU) and bidirectional long short-term memory (BiLSTM) additive-attention model where the Bidirectional GRU/LSTM reads the individual sentence input from left to right and vice versa enabling the capture of the contextual information. On the other hand, the model is trained on two types of embeddings: FastText and local learnable embeddings. The BiLSTM and BiGRU architectures are put into competition to identify the best hyperparameter set for the model. The developed model has been tested on three large-scale commonly employed Arabic sentiment dataset: large-scale Arabic Book Reviews Dataset (ABRD), Hotel Arabic-Reviews Dataset (HARD), and Books Reviews in the Arabic Dataset (BRAD). The testing results demonstrate that our model outperforms both the baseline models and the state-of-the-art models reported in the original references of these datasets, achieving an accuracy score of 98.6 \(\% \) , 96.19 \(\%,95.65\% \) for LARB, HARD and BRAD, respectively. Furthermore, to demonstrate the generalization capabilities of our model, the performances of the model have been evaluated on three other natural language processing tasks: news categorization, offensive speech detection, and Russian sentiment analysis. The results demonstrated the developed model is language and task-independent, which offers new perspectives for the application of the developed models in several other natural language processing challenges.

No abstract

Analyzing social media posts and comments has become a critical task to prevent cyberbullying and hate speech. In this work we present a classification models based on the attention mechanism to analyze Arabic posts and filter out all kinds of inappropriate speech including Religious based hate speech, offensive and abusive content in different Arabic dialects. The attention-based models show promising results for four Arabic datasets. The results are presented and compared in terms of accuracy and training time