COVID-19 Fake News Detection using Pre-Tuned BERT-Based Transfer Learning Models

The World Health Organization declared the Coronavirus disease in 2019;that was a very hard time for people, and every single day, a new crisis emerged. In that case, everyone shares their stories on the social media platform on a daily basis, but no one is sure that information is true or misleading, and that becomes a challenge to detect differences between them. And to tackle this problem, this paper has explored the veracity of social media stories using some machine learning models. The goal of this paper is to test three different BERT base pre-trained transform learning models (BERT, DistilBERT, and RoBERTa) on an English COVID-19 fake news dataset to detect the fake and true news separately. We explore their capability with precision, recall, and F1-score to achieve better results compared with the previous research. © 2022 IEEE.

Analyzing the Performance of Sentiment Analysis using BERT, DistilBERT, and RoBERTa

Sentiment analysis or opinion mining is a natural language processing (NLP) technique to identify, extract, and quantify the emotional tone behind a body of text. It helps to capture public opinion and user interests on various topics based on comments on social events, product reviews, film reviews, etc. Linear Regression, Support Vector Machines, Convolution Neural Networks (CNN), Recurrent Neural Networks (RNN), LSTM (Long Short Term Memory), and other machine learning and deep learning algorithms can be used to analyze the sentiment behind a text. This work analyses the sentiments behind movie reviews and tweets using the Coronavirus tweets NLP dataset and Sentiment140 dataset. Three advanced transformer-based deep learning models like BERT, DistilBERT, and RoBERTa are experimented with to perform the sentiment analysis. Finally, the performance obtained using these models on these two different datasets is compared using the accuracy as the performance evaluation matrix. On analyzing the performance, it can be seen that the BERT model outperforms the other two models. © 2022 IEEE.

Natural language processing analysis applied to COVID-19 open-text opinions using a distilBERT model for sentiment categorization.

COVID-19 is a disease that affects the quality of life in all aspects. However, the government policy applied in 2020 impacted the lifestyle of the whole world. In this sense, the study of sentiments of people in different countries is a very important task to face future challenges related to lockdown caused by a virus. To contribute to this objective, we have proposed a natural language processing model with the aim to detect positive and negative feelings in open-text answers obtained from a survey in pandemic times. We have proposed a distilBERT transformer model to carry out this task. We have used three approaches to perform a comparison, obtaining for our best model the following average metrics: Accuracy: 0.823, Precision: 0.826, Recall: 0.793 and F1 Score: 0.803.

Exploring Public Attitude Towards Children by Leveraging Emoji to Track Out Sentiment Using Distil-BERT a Fine-Tuned Model

Sentiment analysis is a computational method that extracts emotional keywords from different texts through initial emotion analysis (e.g., Happy, Sad, Positive, Negative & Neutral). A recent study by a human rights organization found that 30% of children in Bangladesh are being abused on online in the COVID-19 epidemic by various obscene comments. The main goal of our research is to collect textual data from social media and classify the way children are harassed by various abusive comments online through the use of emoji in a text-mining method and to expose to society the risks that children face online. Another goal of this study is to set a precedent through a detailed study of child abuse and neglect in the big data age. To make the work effective, 3373 child abusive comments are collected manually from online (e.g. Facebook, Newspapers and various Blogs). At present, there is still a very limited number of Bengali child sentiment analysis studies. Fine-tuned general purpose language representation models, such as the BERT family model (BERT, Distil-BERT), and glove word embedding based CNN and Fast-Text models have been used to successfully complete the study. We show that Distil-BERT defeated BERT, Fast-Text, and CNN by 96.09% (relative) accuracy, while Bert, Fast-Text and CNN have 93.66%, 95.73%, and 95.05%, respectively. But observations show that the accuracy of the Distil-BERT does not differ much from the rest of the models. From our analysis, it can be said that the pre-trained models performed outstanding and in addition, child sentiment analysis can serve as a potential motivator for the government to formulate child protection policies and build child welfare systems. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Query-focused Extractive Summarisation for Biomedical and COVID-19 Complex Question Answering

This paper presents Macquarie University’s participation to the two most recent BioASQ Synergy Tasks (as per June 2022), and to the BioASQ10 Task B (BioASQ10b), Phase B. In these tasks, participating systems are expected to generate complex answers to biomedical questions, where the answers may contain more than one sentence. We apply query-focused extractive summarisation techniques. In particular, we follow a sentence classification-based approach that scores each candidate sentence associated to a question, and the n highest-scoring sentences are returned as the answer. The Synergy Task corresponds to an end-to-end system that requires document selection, snippet selection, and finding the final answer, but it has very limited training data. For the Synergy task, we selected the candidate sentences following two phases: document retrieval and snippet retrieval, and the final answer was found by using a DistilBERT/ALBERT classifier that had been trained on the training data of BioASQ9b. Document retrieval was achieved as a standard search over the CORD-19 data using the search API provided by the BioASQ organisers, and snippet retrieval was achieved by re-ranking the sentences of the top retrieved documents, using the cosine similarity of the question and candidate sentence. We observed that vectors represented via sBERT have an edge over tf.idf. BioASQ10b Phase B focuses on finding the specific answers to biomedical questions. For this task, we followed a data-centric approach. We hypothesised that the training data of the first BioASQ years might be biased and we experimented with different subsets of the training data. We observed an improvement of results when the system was trained on the second half of the BioASQ10b training data. © 2022 Copyright for this paper by its authors.

COVID-19 Fake News Prediction On Social Media Data

It is, to tell the truth, that the COVID-19 pandemic has put the whole world in a tough time, and sensitive information concerning COVID-19 has grown tremendously online. Most importantly, the gradual spread of fake news and misleading information during these hard times can have dire consequences, causing widespread panic and exacerbating the apparent threat of a pandemic that we cannot ignore. Because of the time-consuming nature of evidence gathering and careful truth-checking, people get confused between fallacious and trustworthy statement. So, we need a way to keep track of misinformation on social media. Most people think that all social media information is real information though, at the same time, it is a shame that some people misuse this social media platform for their own benefit by spreading misinformation. Many individuals take advantage by playing with the weaknesses of others. As a result, people around the world not only are facing COVID-19, they are also facing infodemics. To get rid of this kind of fake news, we have proposed a research model that can predict fake news related to the COVID-19 issue on social media data using classical classification methods such as multinomial naive bayes classifier, logistic regression classifier, and support vector machine classifier. Moreover, we have applied a deep learning based algorithm named distil BERT to accurately predict fake COVID-19 news. These approaches have been used in this paper to compare which technique is much more convenient for accurately predicting fake news about COVID-19 on social media posts. In addition, we have used a data-set that included 6424 social media posts.

Pre-trained ensemble model for identification of emotion during COVID-19 based on emergency response support system dataset.

The COVID-19 precautions, lockdown, and quarantine implemented throughout the epidemic resulted in a worldwide economic disaster. People are facing unprecedented levels of intense threat, necessitating professional, systematic psychiatric intervention and assistance. New psychological services must be established as quickly as possible to support the mental healthcare needs of people in this pandemic condition. This study examines the contents of calls landed in the emergency response support system (ERSS) during the pandemic. Furthermore, a combined analysis of Twitter patterns connected to emergency services could be valuable in assisting people in this pandemic crisis and understanding and supporting people's emotions. The proposed Average Voting Ensemble Deep Learning model (AVEDL Model) is based on the Average Voting technique. The AVEDL Model is utilized to classify emotion based on COVID-19 associated emergency response support system calls (transcribed) along with tweets. Pre-trained transformer-based models BERT, DistilBERT, and RoBERTa are combined to build the AVEDL Model, which achieves the best results. The AVEDL Model is trained and tested for emotion detection using the COVID-19 labeled tweets and call content of the emergency response support system. This is the first deep learning ensemble model using COVID-19 emotion analysis to the best of our knowledge. The AVEDL Model outperforms standard deep learning and machine learning models by attaining an accuracy of 86.46 percent and Macro-average F1-score of 85.20 percent.

COBERT: COVID-19 Question Answering System Using BERT.

In the current situation of worldwide pandemic COVID-19, which has infected 62.5 Million people and caused nearly 1.46 Million deaths worldwide as of Nov 2020. The profoundly powerful and quickly advancing circumstance with COVID-19 has made it hard to get precise, on-request latest data with respect to the virus. Especially, the frontline workers of the battle medical services experts, policymakers, clinical scientists, and so on will require expert specific methods to stay aware of this literature for getting scientific knowledge of the latest research findings. The risks are most certainly not trivial, as decisions made on fallacious, answers may endanger trust or general well being and security of the public. But, with thousands of research papers being dispensed on the topic, making it more difficult to keep track of the latest research. Taking these challenges into account we have proposed COBERT: a retriever-reader dual algorithmic system that answers the complex queries by searching a document of 59K corona virus-related literature made accessible through the Coronavirus Open Research Dataset Challenge (CORD-19). The retriever is composed of a TF-IDF vectorizer capturing the top 500 documents with optimal scores. The reader which is pre-trained Bidirectional Encoder Representations from Transformers (BERT) on SQuAD 1.1 dev dataset built on top of the HuggingFace BERT transformers, refines the sentences from the filtered documents, which are then passed into ranker which compares the logits scores to produce a short answer, title of the paper and source article of extraction. The proposed DistilBERT version has outperformed previous pre-trained models obtaining an Exact Match(EM)/F1 score of 80.6/87.3 respectively.

Combat COVID-19 infodemic using explainable natural language processing models.

Misinformation of COVID-19 is prevalent on social media as the pandemic unfolds, and the associated risks are extremely high. Thus, it is critical to detect and combat such misinformation. Recently, deep learning models using natural language processing techniques, such as BERT (Bidirectional Encoder Representations from Transformers), have achieved great successes in detecting misinformation. In this paper, we proposed an explainable natural language processing model based on DistilBERT and SHAP (Shapley Additive exPlanations) to combat misinformation about COVID-19 due to their efficiency and effectiveness. First, we collected a dataset of 984 claims about COVID-19 with fact-checking. By augmenting the data using back-translation, we doubled the sample size of the dataset and the DistilBERT model was able to obtain good performance (accuracy: 0.972; areas under the curve: 0.993) in detecting misinformation about COVID-19. Our model was also tested on a larger dataset for AAAI2021 - COVID-19 Fake News Detection Shared Task and obtained good performance (accuracy: 0.938; areas under the curve: 0.985). The performance on both datasets was better than traditional machine learning models. Second, in order to boost public trust in model prediction, we employed SHAP to improve model explainability, which was further evaluated using a between-subjects experiment with three conditions, i.e., text (T), text+SHAP explanation (TSE), and text+SHAP explanation+source and evidence (TSESE). The participants were significantly more likely to trust and share information related to COVID-19 in the TSE and TSESE conditions than in the T condition. Our results provided good implications for detecting misinformation about COVID-19 and improving public trust.