Titan on the Red

LANL Weapons Program scientists and engineers will soon have automated research support with the implementation of artificial intelligence/machine learning.

By Riz Ali | September 9, 2021

Titan@2x

Think Alexa or Siri, but meet Titan on the Red.

LANL Weapons scientists and engineers who long for an automated research assistant will soon have that support when it comes to poring through millions of records and immense data. The National Security Research Center (NSRC) plans to implement artificial intelligence/machine learning (AI/ML) on its red network and as a part of its processes to make digitized documents easier for researchers to find.

This AI/ML system is called Titan on the Red and will ensure faster, more efficient results for the Lab’s Weapons Program researchers, as well as the classified library’s collections management, said Mott Linn, the NSRC’s Chief Librarian.

“We see artificial intelligence as a tool to help us go through the monumental tasks we have in digitizing, cataloging, and searching our collections,” Linn said. “After completing a successful six-month test run with unclassified materials, we’re confident AI/ML will save the Lab a lot of time and even more money when it comes to our research efforts.”

What is AI/ML?

In addition to digital assistants, other examples of AI include chatbots, smart thermostats, and preferences in the things you read and watch. AI involves developing computers to mimic human cognition. Meanwhile, ML is a type of AI that uses sophisticated algorithms to perform specific functions.

“Although AI/ML is a relatively new advancement, it’s proven,” Linn said. “It will really be a gamechanger for the way the NSRC operates, and more importantly, the way we help the Lab meet its national security mission.”

How does it work?

The goal of Titan on the Red is to extract metadata from the Lab’s various digital data repositories, such as the Online Vault, PDMLink, shared drives, and SharePoint sites. The software then presents researchers with a natural-language enhanced, AI/ML-based interface to search through the NSRC’s entire digitized collection in an intuitive manner. In short, it helps users conduct exploratory-style research for documents, regardless of their original location.

“One of the greatest assets at LANL is the information that we have generated in over 75 years of nuclear weapons work,” said Charlie Nakhleh, the Associate Laboratory Director for Weapons Physics. “This is what distinguishes us from any other weapons laboratory in existence. The Titan on the Red system will make this valuable information discoverable.”

These collections contain information on nuclear weapons modeling and simulation, weapons designs, and pit production, all of which are critical to the Lab’s stockpile stewardship mission and pit production benchmarks, Nakhleh said. Researchers access NSRC materials on a daily basis, though it can be a cumbersome process that often requires a librarian to manually search records.

Implementing Technology to Improve Research

Titan on the Red also addresses a significant backlog in digitizing materials that are not yet searchable, said Nanette Mayfield, who leads the NSRC’s Digital Collections Team.

With tens of millions of materials in its collections, the NSRC is one of the largest research libraries in the United States. However, less than 10 percent of these holdings have been digitized, and less than 10 percent of those have been cataloged. This affects the speed at which the NSRC is able to provide researchers with materials.

These materials would become accessible in a matter of months once the system comes online, rather than in double-digit decades if the NSRC were to continue work without AI/ML software. The NSRC has 2.4 million digitized documents that are not yet searchable. It would take NSRC staff more than 400 years at the current rate to catalog these files.

“This advancement is really the only solution to making the Lab’s vast collections searchable to our researchers,” Mayfield said. “Investing in AI/ML saves countless manpower hours and money, while directly contributing to the Lab’s mission success.”