AlphaDev: AI Reinforcement Learns Faster Sorting Methods

AlphaDev discovers faster sorting algorithms

The digital society drives an increase in energy consumption and computation. In the last 50 years, we have relied on hardware improvements to keep up. As microchips reach their physical limit, it is critical to improve code running on them in order to make computing more efficient and sustainable. It is particularly important for the algorithms which make up the code that runs trillions of time a day.

Our paper, published in Nature today, introduces AlphaDev – an artificial intelligence system (AI) that uses reinforcement-learning to discover computer science algorithms. These algorithms are superior to those developed by scientists and engineers for decades.

Nature Publication.

Source:
https://www.deepmind.com/blog/alphadev-discovers-faster-sorting-algorithms

发表回复

您的电子邮箱地址不会被公开。 必填项已用 * 标注