An international team of scientists has developed a cryptography-inspired method that searches publicly available data from NASA’s Fermi Gamma-ray Space Telescope for pulsars about a million times faster than previous approaches. Using 15 years of Fermi observations, the researchers identified PSR J1622-7204, an isolated neutron star rotating 261 times per second. Their findings were published Thursday in Physical Review Letters in a paper explaining the results.
A faster way to find pulsars
Pulsars are rapidly rotating neutron stars that sweep beams of radiation toward Earth like cosmic lighthouses. Finding them in gamma-ray data is exceptionally difficult because Fermi detects only a small number of photons from any individual pulsar. Those photons are mixed with background emissions from other gamma-ray sources and from the wider universe.
Researchers must infer several unknown properties at once, including the star’s rotation rate and its position in the sky. Searching every possible combination would require an impractical number of calculations.
“We haven’t seen all the rotations of the pulsar, and we can’t just count them to figure out where the photons came from,” said lead author Daniel Puckkolb of Germany’s Max Planck Institute for Gravitational Physics, also known as the Albert Einstein Institute. “We have a four-dimensional grid, and we have to search in each direction with millions of trials. There isn’t enough computing power on Earth to do that.”
The new method found PSR J1622-7204 using the equivalent of about 10 hours of calculation on a standard computer processor. By comparison, the best previous methods failed to find pulsars despite using as much computing power as a standard processor would need for 600 years of calculations.
Turning the search into a geometry problem
Traditionally, scientists have divided Fermi’s data into small sections and used computing power donated by volunteers around the world. Even with this distributed approach, identifying a single pulsar typically required the equivalent of hundreds of computer years.
Study co-authors Dotan Gazith and Barak Zackay of the Weizmann Institute of Science in Israel recognized that the search could be represented as a geometry problem. Specifically, the task can be expressed as finding the shortest vector, or an array of regularly spaced points, in a high-dimensional lattice.
The lattice technique uses algorithms developed by cryptographers seeking to break encryption systems designed to withstand future code-breaking efforts. Related lattice methods are also used to improve the efficiency of satellite navigation systems. The researchers’ analysis itself is not related to encryption.
“Although our analysis here has nothing to do with encryption, these algorithms could help solve the problem of searching for pulsars in data,” Gazis said.
Each trial took about five minutes on a single computer core. The method can run on a standard laptop, which typically has two to eight cores. After the lattice method identified a pulsar candidate, the researchers checked it against the full Fermi dataset. The resulting signal was too strong to be explained by coincidence.
Searching for pulsars in binary systems
The team’s next goal is to search for pulsars in binary star systems, where a pulsar orbits another star. These objects have long been considered extremely difficult to detect because the companion star’s gravity distorts the pulsar’s signal, making traditional searches largely ineffective.
“We know that there are probably dozens, maybe even hundreds, of pulsars in this dataset that we didn’t find,” said study co-author Colin Clark, an astrophysicist at the Max Planck Institute. “Our next goal is to expand our search and use a combination of new algorithms and crowdsourced computing power from volunteers in the Einstein@Home program to help us find many new pulsars that we think should be out there.”
What open NASA data makes possible
The discovery depended on NASA’s policy of making scientific data, including Fermi observations, freely available. Launched in 2008, Fermi surveys the entire sky every three hours, observing gamma rays, the highest-energy form of light. After an initial one-year adjustment period, all Fermi data became publicly available, with new data continuing to be released after processing.
“The good thing about NASA data is that it’s all open,” Clark said. “All 18 years of Fermi data are publicly available, and the data continues to grow.”
This openness has allowed researchers outside the Fermi mission, including teams inspired by cryptography, to apply new methods to the dataset. NASA also supports the scientific community with open software and tools, including Fermitools, which teams can use to prepare datasets before beginning their analyses.
The Fermi mission has expanded its catalog from hundreds of gamma-ray sources to thousands, with new discoveries emerging as scientists apply different ideas to the same observations.
“Whatever search one team is doing, another team just has to go and check that search,” said Elizabeth Hayes, Fermi project scientist at NASA’s Goddard Space Flight Center in Greenbelt, Maryland. “It has improved all the science.”
Source: science.nasa.gov


