Element AI recently launched a source of more than 40,000 articles from scholarly journals containing COVID-19 and other related infectious diseases for research materials that are potentially useful to researchers
(Photo : Unschool India on Flickr)

(Photo : Pixabay)

Element AI recently introduced a search tool that explores the COVID-19 Open Research Dataset. This is a source of more than 40,000 articles from scholarly journals containing COVID-19 and other related infectious diseases for research materials that are potentially useful to researchers.

In addition, one can also look for natural-language terminologies, keywords and phrases to surface works of literature that have the semantically-the same content, or copy wordings from paragraphs, or queries that are related, into the search bar to bring back such articles with only the most essential statements emphasized or highlighted.

Moreover, a surge of studies on COVID-19, forecasted to hit millions of people, has paved the way to the Internet in the months from the time this pandemic started.

According to Reuters, approximately 153 pre-printed studies on this infectious disease have been publicly made as of March 24. The said studies promise understandings on the spread of this virus although many have not gone through peer-review, making it a struggle for the stakeholders to distinguish the so-called "wheat from the chaff."

Up to this time, the platform influences the technology of Element AI, from its product, the Knowledge Scout, which employs artificial intelligence to catch the links between the different bits of information in order to learn and enhance over time, while developing a source of unspoken knowledge.

ALSO READ: AI's New Antibiotics Discovery: Will it Really Save a Person's Life?

The Platform to be Progressively Updated

According to Element AI, the said platform will be increasingly updated in the coming days and weeks, with added COVID-19 data sets along with features such as the capabilities for addressing queries which are free-domain, the detection of the topic, and the summarization that drives questions.

Additionally, this presentation of the new search tool of the Element AI follows that the "CORD-19 Search" of Vespa, is somewhat similarly scanning through the COVID-19 Open Research Dataset for significant materials that have been carefully reviewed. 

Meanwhile, the DMIS Lab of the Korea University, for its month, released "Covidsearch" this month, providing question-answering in actual time on more than 30,000 COVID-19 related pieces of literature with results highlighting applicable biomedical articles. Another similar institute is offering a basic platform that helps explore this COVID-19 dataset's complete text.

Relatively, the AI behind these, as well as the other search tools for COVID-19 learns from indications like the data resulting from several inputs, for one. Every indicator informs the predictions of the system such that it is learning how different resources are pertinent (or not) to a particular search query.

Natural Languages in AI

The processing of natural languages enables the said resources to understand a single study in a data set's context. Natural language search, on the other hand, is AI's specialized application, creating a "word mesh" coming from the text that's free-flowing.

This is quite similar to a knowledge graph, connecting similar ideas released to larger notions to return the same answer no matter how a question or query is phrased. Having gathered these, "the jury is out on how big" of an effect semantic search tools may possibly have on the ongoing research on COVID-19. However, as referred to earlier, they tend to force down on the more doubtful research study that has come to light. One of the most recent papers suggests a connection between COVID-19 and HIV, while other research claims, the illness came from outer space.