The rapid advancement of communication technology has led to an alarming increase in the spread of misinformation and fake news. This phenomenon has become increasingly common in the current COVID-19 pandemic. A study was conducted last year by the National Centre of Infectious Diseases (NCID) on the Singapore population’s knowledge, risk perception, and behavior during the COVID-19 outbreak. The majority of the public chose social media as their preferred source of information on COVID-19. Messaging platforms such as WhatsApp and Facebook Messenger are the most common channels used to share information related to the pandemic.
The term "infodemic" has been used to describe this tsunami of information that is likely to be inaccurate and is difficult to be verified. Those who recirculate such information frequently are also the ones most likely to adopt irrational behaviors like panic buying. Hence, there is an urgent need to develop an effective tool for Singaporeans to verify the trustworthiness of information related to COVID-19, to alleviate the spread of panic, chaos, and potential health hazards.
What it does
We propose a product that aims to address the issue of detecting misleading information related to COVID-19 for pioneer generation. The COVID pandemic and its consequences have been actively discussed on social media. Several issues has been aroused concerning the pioneer generation. Information from these non-traditional sources is often inaccurate and difficult to verify, and their rapid dissemination could exacerbate the negative impact of the pandemic. The pioneer generation needs more explanation rather than just a single "valid" or "invalid" result. Therefore, we propose to develop a bilingual fact-checking tool by exploring various Machine Learning and Deep learning methods.
How we built it
We develop a COVID-19 fact-checking tool by exploring Machine Learning and Deep learning methods using fine-tuned DistilBert model. We also made our models results easy to be interpreted by the elderly. Our website can be accessed easily.
Accomplishments that we're proud of
We have used the recent advancement in deep learning(fine-tuned DistilBert model) together with great explainable AI in front end and back end.