acm-header
Sign In

Communications of the ACM

ACM TechNews

Humans Beat DeepMind AI in Creating Algorithm to Multiply Numbers


Multiplying numbers is a common computational problem.

Credit: narvikk/Getty Images

Jakob Moosbauer and Manuel Kauers at Austria's Johannes Kepler University Linz bested an algorithm developed by artificial intelligence company DeepMind with a program that can perform matrix multiplication more efficiently.

Earlier this month, DeepMind unveiled a method for multiplying two five-by-five matrices in just 96 multiplications, out-performing a more-than-50-year-old record.

Moosbauer and Kauers reduced the process to 95 multiplications by testing multiple steps in multiplication algorithms to see if they could be combined.

Said Moosbauer, "We take an existing algorithm and apply a sequence of transformations that at some point can lead to an improvement. Our technique works for any known algorithm, and if we are lucky, then [the results] need one multiplication less than before."

From "Humans Beat DeepMind AI in Creating Algorithm to Multiply Numbers"

New Scientist (10/13/22) Matthew Sparkes
View Full Article


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account