AI is Behind NASA’s Out-of-This-World Science Search Engine

The Science Discovery Engine uses neural nets and natural language processing to search through NASA’s open science datasets.

This week, enterprise search company Sinequa announced that its AI-powered platform is being used to power NASA’s new Science Discovery Engine (SDE).

Sinequa’s search solution takes advantage of both neural networks and natural language processing to streamline the data retrieval process. Sinequa says it worked closely with developers and researchers at NASA to create a tool for accessing the vast array of data publicly available from the Science Mission Directorate.

“The Science Discovery Engine is a fantastic infrastructure which helps both NASA scientists and communities around the world to find the information they need in just a matter of seconds, rather than minutes and hours… we look forward to assisting in future as NASA looks to pull in more datasets and sources of information to power its insights engine,” said Ulf Zetterberg, Co-CEO of Sinequa, in a company press release.

Making data more accessible

With the new tool, NASA aims to make it easier for the public and researchers to access all their disparate data. The tool now integrates data from all five of the focus areas of NASA’s Science Mission Directorate, including astrophysics, earth science, biological and physical sciences, heliophysics, and planetary science.

The tool allows users to search over 84,000 datasets, including over 715,000 documents, in one easy-to-use website. All these enormous datasets are currently stored across the internet, on websites, data repositories and archives, making the new search engine a powerful tool for researchers to quickly locate data.

“To me, the most exciting thing about the SDE is how it makes the rich wealth of NASA’s open science data and information more accessible to an ever-growing community of users,” said Kaylin Bugbee, leader of SDE team operations and a NASA research scientist, in a NASA-written article on the SDE from January 2023. “This increased accessibility will open new pathways to scientific discovery and encourage more people to make use of the open science data and information NASA provides.”

One important characteristic of the new search engine is its ability to use scientific terminology as part of its natural language processing capabilities. Nearly 9,000 different scientific terms can be understood by the search engine, and Sinequa and NASA expect this number to grow as AI learns from the search engine history. As many scientists and researchers know, this is an incredibly useful feature, as scientific terminology lets users integrate increasingly specific language in their search engine queries to accelerate the data location process.

The open science future

NASA declared 2023 “A Year of Open Science” to bring awareness to and celebrate the successes of open science and engineering. In launching the Science Discovery Engine, NASA is continuing to support ongoing efforts to move towards an open research environment. All these efforts are part of a broader five-year “Transform to Open Science” mission at NASA, which aims to create a long-term, inclusive culture of open science in research communities.

“Sharing science and open information is the only way for science, technology and innovation to prosper, and it’s fantastic to be playing a part in making this happen,” said Zetterberg.