Model Generalization on COVID-19 Fake News Detection (preprint)

Amid the pandemic COVID-19, the world is facing unprecedented infodemic with the proliferation of both fake and real information. Considering the problematic consequences that the COVID-19 fake-news have brought, the scientific community has put effort to tackle it. To contribute to this fight against the infodemic, we aim to achieve a robust model for the COVID-19 fake-news detection task proposed at CONSTRAINT 2021 (FakeNews-19) by taking two separate approaches: 1) fine-tuning transformers based language models with robust loss functions and 2) removing harmful training instances through influence calculation. We further evaluate the robustness of our models by evaluating on different COVID-19 misinformation test set (Tweets-19) to understand model generalization ability. With the first approach, we achieve 98.13% for weighted F1 score (W-F1) for the shared task, whereas 38.18% W-F1 on the Tweets-19 highest. On the contrary, by performing influence data cleansing, our model with 99% cleansing percentage can achieve 54.33% W-F1 score on Tweets-19 with a trade-off. By evaluating our models on two COVID-19 fake-news test sets, we suggest the importance of model generalization ability in this task to step forward to tackle the COVID-19 fake-news problem in online social media platforms.

Will Proton Pump Inhibitors Lead to a Higher Risk of COVID-19 Infection and Progression to Severe Disease? A Meta-analysis (preprint)

BackgroundPrevious researches on the association between proton pump inhibitors (PPIs) use and the treatment and prevention of COVID-19 have generated inconsistent findings. Therefore, this Meta-analysis was conducted to clarify the outcome in patients who take PPIs. MethodsWe carried out a systematic search to identify potential studies until November 2020. Heterogeneity was assessed using the I-squared statistic. Odds ratios (ORs) with its 95% confidence intervals (CIs) were calculated by fixed-effects or random-effects models according to the heterogeneity. Sensitivity analyses and tests for publication bias were also performed. ResultsEight articles with more than 268,683 subjects were included. PPI use was not associated with increased or decreased risk of COVID-19 infection (OR:3.16, 95%CI = 0.74-13.43, P=0.12) or mortality risk of COVID-19 patients (OR=1.91, 95% CI=0.86-4.24, P=0.11). While it can add risk of severe disease (OR=1.54, 95% CI=1.20-1.99, P<0.001;) and secondary infection (OR=4.33, 95% CI=2.57-7.29). No publication bias was detected. ConclusionsPPI use is not associated with increased risk infection and may not change the mortality risk of COVID-19, but appeared to be associated with increased risk of progression to severe disease and secondary infection. However, more original studies to further clarify the relationship between PPI and COVID-19 are still urgently needed.