DeepMind AI Breaks 50-Year-Old Matrix Multiplication Record
DeepMind's AlphaZero AI discovers a faster method for matrix multiplication, beating a 50-year-old computer science record. This breakthrough promises to speed up thousands of everyday computing tasks and reduce energy consumption.
DeepMind uses its game-playing AI, AlphaZero, to break a 50-year-old record in computer science by discovering a faster way to perform matrix multiplication. This fundamental math problem sits at the heart of countless everyday applications, from displaying images on screens to simulating complex physics and running machine learning models.
Despite being a basic concept taught in high school, finding the absolute fastest algorithm for matrix multiplication remains one of the biggest open problems in computer science. The challenge lies in the sheer number of possible ways to multiply two matrices, which exceeds the number of atoms in the universe for certain cases.
To solve this, researchers turn the mathematical problem into a three-dimensional board game called TensorGame, where each move represents a step in the multiplication process. By training a new version of AlphaZero to play this game, the AI successfully discovers novel algorithms that outperform the previous record-holding methods, promising significant cost and energy savings for future computing tasks.