By Sarah Wang
As malicious actors become more adept in their attempts to circumvent international nuclear non-proliferation safeguards, the United States government has invested in research to better detect the malicious activities of malicious actors. To address non-proliferation threats, agencies like the International Atomic Energy Agency (IAEA) use careful monitoring techniques to ensure that nuclear materials subject to agreements are not used to produce nuclear weapons. They also use sophisticated forensic methods to determine the origin of nuclear material recovered by law enforcement. However, these techniques are often time-consuming and laborious.
New research from the Pacific Northwest National Laboratory (PNNL) uses machine learning, data analytics and artificial reasoning to facilitate and accelerate threat detection and forensic analysis in the nuclear field. By combining these computational techniques with expertise in non-proliferation and safeguards, PNNL strives to uncover innovative methods by conducting research and transforming that research into concrete, actionable solutions.
Benjamin Wilson, a nonproliferation analyst at PNNL, is in a unique position to merge these data analysis and machine learning techniques with nuclear analysis. As a former IAEA safeguards inspector, Wilson knows exactly what kind of information the IAEA is looking for in order to expose possible malicious activity by actors.
“Preventing nuclear proliferation requires vigilance,” Wilson said. “It involves work, from audits of nuclear materials to investigations into who is handling nuclear materials. Techniques based on data analysis can be exploited to facilitate this. »
With support from the National Nuclear Security Administration (NNSA), the Mathematics for Artificial Reasoning in Science (MARS) initiative, and the Department of Defense, PNNL researchers are working on several projects aimed at making non-proliferation more effective and nuclear safeguards. Here are some highlights of their recent research.
Detection of nuclear material diversion
Nuclear reprocessing facilities recover spent nuclear fuel and separate it into waste. The products are then used to produce compounds that can be recycled as new fuel for nuclear reactors. These compounds contain uranium and plutonium, which could be used to produce nuclear weapons. The IAEA monitors nuclear facilities to ensure that none of the nuclear material is diverted to nuclear weapons. This usually involves regular inspections as well as the collection of samples for subsequent destructive analysis.
“We could save a lot of time and labor if we could create a system that automatically detects anomalies based on plant processing data,” Wilson said.
In a study published in The International Journal of Nuclear Safeguards and Non-Proliferation, Wilson worked with researchers at Sandia National Laboratories to build a virtual replica of a reprocessing facility. They then trained a machine learning model to detect patterns in process data representing the diversion of nuclear materials. In this simulated environment, the model showed encouraging results. “While this approach is unlikely to be used in the near future, our system offers a promising start to complement existing safeguards,” Wilson said.
Analyze texts for signs of nuclear proliferation
Who does research on nuclear materials? Is this research consistent with international agreements and declarations of this State? Where did they acquire the specialized equipment they use? Where are nuclear materials currently used for research?
These are the kinds of questions that IAEA analysts strive to answer every day. To find these answers, they usually have to spend many hours reading research papers and manually sifting through lots of data.
PNNL data scientists Megha Subramanian and Alejandro Zuniga along with Benjamin Wilson, Kayla Duskin and Rustam Goychayev are working to make this task much easier with research presented in The International Journal of Nuclear Safeguards and Non-Proliferation.
“We wanted to create a way for researchers to ask nuclear-specific questions and receive correct answers,” Subramanian said.
The team developed a machine learning tool based on Google’s BERT: a language model trained on Wikipedia data for general knowledge queries. Language models allow computers to “understand” human languages: they can read texts and extract important information, including context and nuance. People can ask BERT questions, like “what is the capital of France?” and receive the correct answer.
While the model trained by Wikipedia excels at answering general knowledge questions, it lacks nuclear knowledge. Therefore, the team created AJAX – Artificial Judgment Aid from Text – to fill this knowledge gap.
“Although AJAX is still in its infancy, it has the potential to save many hours of analyst time by providing both a direct response to queries and the evidence for that response,” Subramanian said. The evidence particularly intrigues researchers because most machine learning models are often described as “black boxes” that leave no trace of evidence for their answers, even if they are correct. AJAX seeks to ensure verifiability by retrieving documents that contain evidence.
“When the field is as important as detecting nuclear proliferation, it’s critical for us to know where our information comes from,” Subramanian said.
Using Image Analysis to Determine the Origin of Nuclear Material
Occasionally, a law enforcement officer, in the United States or elsewhere, will encounter nuclear material that is beyond regulatory control and of unknown origin. In these cases, determining where the material came from and/or its origins of creation is crucial, as the recovered material may be only part of the material that is at risk of being trafficked. In the event of such incidents, the IAEA maintains a database and encourages countries to cooperate and collaboratively combat illicit nuclear trafficking in order to enhance nuclear security. Forensic analysis of nuclear materials is one of the analytical tools used in this vital effort.
PNNL researchers, in collaboration with the University of Utah, Lawrence Livermore National Laboratory, and Los Alamos National Laboratory, have developed a way to use machine learning to aid in forensic analysis of these samples. . Their method uses electron microscopy images to compare the microstructures of different nuclear samples. Different samples contain subtle differences that can be identified using machine learning.
“Imagine synthesizing nuclear material was like baking cookies,” said Elizabeth Jurrus, MARS initiative lead. “Two people can use the same recipe and end up with different-looking cookies. It’s the same with nuclear materials.
The synthesis of these materials could be affected by many factors such as local humidity and the purity of the starting materials. The result is that nuclear material produced at a specific facility ends up with a specific structure – a “characteristic appearance” that can be seen with an electron microscope.
“Our collaborators at the University of Utah have built a library of images of different nuclear samples,” said Alexander Hagen, co-lead author of the study published in the Journal of Nuclear Materials. “We used machine learning to compare images from their library to unknown samples to determine where the unknowns came from.”
This could help nuclear analysts determine the source of a material and guide further investigations.
Machine learning to keep citizens safe
While it may take some time for agencies like the IAEA to adopt machine learning techniques in their nuclear threat detection process, it is clear that these technologies can have an impact and streamline the process.
“While we don’t expect machine learning to replace anyone’s job, we see it as a way to make their jobs easier,” Jurrus said. “We can use machine learning to identify important information so analysts can focus on what’s most important.
Pacific Northwest National Laboratory draws on its distinctive strengths in chemistry, earth science, biology, and data science to advance scientific knowledge and address the challenges of sustainable energy and national security. Founded in 1965, PNNL is managed by Battelle for the Department of Energy’s Office of Science, which is the largest supporter of basic physical science research in the United States. The DOE’s Office of Science is working to address some of the most pressing challenges of our time. For more information, visit https://energy.gov/science. For more information on PNNL, visit the PNNL News Center. follow us on TwitterFacebook, LinkedIn and Instagram.
Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of press releases posted on EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.