Welcome to the COVID-19 Misinformation Detector!
There is a lot of misinformation related to the COVID-19 vaccine being posted online from unreliable sources. The COVID-19 Misinformation Detector allows you to check if the information you are reading online (e.g. from Twitter or Facebook) contains misinformation or not!
Enter the text from the online post in the "Hosted inference API" text area to the right to check if it is misinformation. "LABEL_0" means that no misinformation was detected in the post, while "LABEL_1" means that the post is misinformation.
The COVID-19 Misinformation Detector is a modified version of the "bert-base-uncased" transformer model, found here. It is fine-tuned on two datasets containing tweets relating to the COVID-19 pandemic; each tweet is labelled as containing misinformation (1) or not (0), as verified by healthcare experts. The datasets used are:
- ANTi-Vax: a novel Twitter dataset for COVID-19 vaccine misinformation detection
- CoAID (Covid-19 HeAlthcare mIsinformation Dataset)
For a more detailed explanation, check out the technical report here, and check out my literature review on transformers here!
