Have you ever attempted math problems involving matrix multiplication? Did you know that it is utilised in every aspect of life, from processing images on smartphones to recognising vocal commands to creating graphics for video games? If you learned it in school, you already know it is not that tough.
And when more and more computing processes are utilised, it becomes more complicated at each stage. A new Artificial Intelligence system called Deep Mind from Google has just been released, and it has the potential to provide novel, effective, and demonstrably right solutions to basic problems like matrix multiplication.
The technology, known as AlphaTensor, could answer the long-standing dilemma of how to multiply two matrices quickly.
In a paper published in the journal Nature, researchers said that 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.
Researchers created a single-player game to solve the problem of developing effective matrix multiplication algorithms, and discovered that there are far more potential algorithms than there are atoms in the universe.
They began by teaching AlphaTensor agents to play the game using reinforcement learning, with no prior knowledge of the available matrix multiplication techniques.
Post Your Comments