Google Releases NLU-Powered Tool for Coronavirus Researchers

The traditional search engine can provide a relevant answer to most COVID-19 questions.

For example, queries such as “What are the symptoms of coronavirus?” or “Where can I get tested in my country?” will produce relevant results. Although the answers may satisfy regular users, coronavirus researchers have more specific needs.

Since the COVID-19 questions from researchers are more pointed, a search engines’ keyword-based approach can’t deliver accurate results. That’s where Google’s new AI-powered search tool comes in.

The new Natural Language Understanding-powered tool provides a way to explore COVID-19 scientific literature.

NLU is a subset of Natural Language Processing (NLP), and it focuses on a smaller context. At the same time, NLU works on deriving the meaning of questions while drawing distinct insights.

In a blog post announcement, a Natural Language Understanding scientist at Google Research, Keith Hall, said:

“…we are launching the COVID-19 Research Explorer, a semantic search interface on top of the COVID-19 Open Research Dataset (CORD-19), which includes more than 50,000 journal articles and preprints.”

Here’s what you should know about the NLU-powered tool.

Creating an NLU-Powered Tool Specifically for Coronavirus Researchers

As said earlier, Google’s AI team designed the NLU-powered tool for researchers and scientists.

With the new research explorer tool, scientists can efficiently read through literature for answers to coronavirus-related questions. Aside from returning answers to questions, the tool also highlights parts of the paper that contain said solutions.

What’s more, users can ask a follow-up question to narrow down the result further.

Google’s popular BERT language model helps power the COVID-19 research explorer’s semantic search. Also, the artificial intelligence team at Google trained the AI on BioASQ — a biomedical semantic model — to enhance search results.

There’s also a hybrid term-neural retrieval model for better results.

The term-based model focuses on improving the result’s accuracy. On the other hand, the neural model helps the search engine understand the meaning and context of the query.

“We hope these features will foster knowledge exploration and efficient gathering of evidence for scientific hypotheses,” Hall concluded.