DeepMind Breaks 50-Year old Math Record Using AI

Improving the efficiency of algorithms for fundamental computations can have a widespread impact, as it can affect the overall speed of a large number of computations. Matrix multiplication is one such primitive task, that occurs in many systems. It involves multiplying two rectangular arrays of numbers and is usually found at the heart of speech recognition, image recognition, smartphone image processing, compression, and generating computer graphics. So, finding the fastest way to solve matrix multiplication is a task. However now, DeepMind found a faster way to perform matrix multiplication using AI and it also broke a 50-year-old math record.

DeepMind Breaks 50-Year-old Math Record Using AI

Check Also: Meta Announces First AI-Speech Translator For Unwritten Language

Graphics processing units (GPUs) are particularly good at performing matrix multiplication due to their massively parallel nature. They can dice a big matrix math problem into many pieces and attack parts of it simultaneously with a special algorithm. In 1969, a German mathematician named Volker Strassen discovered the previous-best algorithm for multiplying 4×4 matrices. His algorithm reduces the number of steps necessary to perform a matrix calculation. For example, multiplying two 4×4 matrices together using a traditional method would take 64 multiplications. However, Strassen’s algorithm can perform the same calculation in 49 multiplications.

Using a neural network called AlphaTensor, DeepMind discovered a way to reduce that count to 47 multiplications. Its researchers published a paper about the achievement in Nature last week. You can check out more about it by clicking here.

See Also: Adobe’s Upcoming AI Project Will take Image Compositing to a Next Level: Report