Scientists at Los Alamos National Laboratory (LANL) of the U.S. Department of Energy developed an artificial-intelligence–based system to detect illegal use of supercomputers for cryptocurrency mining.
“Based on recent hacks of computers in Europe and other regions, this type of program controller will soon become critically important for preventing crypto miners from breaching high-performance computing resources and stealing resources,” noted LANL researcher Gopinath Chennupati.
The lab-developed solution uses graphs to detect hidden miners. The graphs represent all programs as nodes and the connections between them.
The AI compares software graphs, much like fingerprints, and identifies programs that should not be running on that computer.
LANL says that the graph-based approach cannot be fooled by the usual methods attackers use to mask mining programs.
According to the researchers, tests using “malicious Bitcoin mining code” showed that the AI-based system detected illegal cryptocurrency mining operations faster and more reliably than ordinary analytical tools.
They acknowledged that the solution does not offer 100% guarantees, but “significantly expands the set of effective approaches” in the fight against cybercriminals.
Earlier, in October last year, Andrey Rybkin of the All-Russian Scientific Research Institute of Experimental Physics received three years in a penal colony for mining on a supercomputer. Another defendant in the case, Andrey Shatohin, was sentenced to four years’ probation and fined 250,000 rubles. The third, Denis Baykov, was fined 450,000 rubles and banned from government service for 2.5 years.
Subscribe to the Forklog channel on YouTube!