[ad_1]
Receiving input from a number of smaller panels, the survey committee takes into account an enormous amount of information to formulate research strategies. While the academies won’t release the committee’s final recommendation to NASA for a few more weeks, scientists are dying to know which of their questions will succeed and which will be left out.
“The Decadal Survey really helps NASA decide how to manage the future of human exploration in space, so it’s really important that they are well-informed,” he says. Brant RobertsonProfessor of astronomy and astrophysics at UC Santa Cruz.
A research team wants to use artificial intelligence to facilitate this process. Their recommendations are not for a specific task or line of inquiry; instead, they say their AI could help scientists make tough decisions about which other recommendations to prioritize.
The idea is that by training an AI to identify rapidly growing or dwindling areas of research, the tool could make it easier for survey committees and panels to decide what should make the list.
“What we wanted was to have a system that would do a lot of the work that Decadal Survey did and let the scientists working on the Decadal Survey do their best,” he says. Harley Throne, a retired senior scientist and lead author at NASA’s Goddard Space Flight Center your offer.
Although the members of each committee are selected based on their expertise in their fields, it is impossible for every member to grasp the nuances of every scientific theme. According to the authors, the number of astrophysics publications is increasing by 5% each year. That’s too much for anyone to handle.
This is where Thronson’s artificial intelligence comes into play.
It took a little over a year to build, but in the end, Thronson’s team was able to train it on more than 400,000 studies published in the decade leading up to the Astro2010 survey. They were also able to teach the AI to review thousands of summaries from two- and three-word topic phrases like “planet system” or “extrasolar planet” to identify both low- and high-impact areas.
According to the researchers’ white paper, artificial intelligence has successfully “retrospected” six popular research themes of the past 10 years, including a meteoric rise in exoplanet research and the observation of galaxies.
“One of the challenging aspects of AI is that it sometimes predicts, uncovers, or analyzes things that are completely surprising to humans,” Thronson says. “And we’ve seen that a lot.”
Thronson and his collaborators think the board should use their AI to help the panel review and summarize the vast amount of text the panel needs to review, leaving the final call to human experts.
His research is not the first to try to use artificial intelligence to analyze and shape the scientific literature. Other AIs already used to assist peer review of scientists work of colleagues.
But can he be trusted for a task as important and effective as the Decadal Survey?
[ad_2]
Source link