The new system wants better search results on COVID on the Internet. (Representative)
Toronto:
A team led by an Indian-origin researcher in Canada has developed a new system that can increase the accuracy and reliability of online health searches by up to 80 percent to help people make better decisions about topics like COVID-19. Increases.
The team from the University of Waterloo in Canada noted that Internet search engines are the most common tool the public uses to look up facts about COVID-19 and its impact on their health.
The spread of misinformation can have real consequences, so the team devised a way to make these searches more reliable.
“With so much new information coming out all the time, it can be challenging for people to know what is true and what is not,” said Raunak Pradeep, a PhD student at Waterloo’s Chariton School of Computer Science and lead author of one. Study.
“But the consequences of misinformation can be very bad, such as people going out and buying medicines or using home remedies that can hurt them,” Mr. Pradeep said.
Researchers said that even the large search engines that host billions of searches every day cannot keep up because there has been so much scientific data and research on COVID-19 in such a short period of time.
Mr Pradeep said, “Most systems are trained on well-curated data, so they do not always know that articles promoting drinking bleach to prevent COVID-19 as opposed to actual health information How to differentiate between
“Our goal is to help people see the right articles and get the right information so they can make better decisions with things like COVID in general,” he said.
Mr Pradeep said that the project aims to refine internet search programs to promote best health information for users.
The team leveraged their two-step neural reranking architecture to search which they augmented with a label prediction system trained to distinguish correct from suspicious and incorrect information.
The system combines with a search protocol that relies on data from the World Health Organization (WHO) and verified information as the basis for ranking, promotion and sometimes excluding online articles.
“Our design could potentially improve consumer health discovery to combat misinformation, a challenge recently exacerbated by the COVID-19 pandemic,” the study authors wrote.
Mr. Pradeep from the University of Waterloo and other authors Xueguang Ma, Rodrigo Nogueira and Jimmy Lin presented a paper on the preliminary findings of the system at SIGIR “21, a conference on research and development in information retrieval, held between 11 July-15 online.”
.