科学家开发一个光子芯片 以高精度加速机器学习任务 Scientists develop a photonic chip that speeds up machine learning tasks with high accuracy.

麻省理工学院的科学家创造了一个新的芯片,利用光来进行深层神经网络计算,使机器学习速度更快,能源效率更高。 Scientists from MIT have created a new chip that uses light to perform deep neural network computations, making machine learning faster and more energy-efficient. 这个光子芯片在半毫秒内完成关键任务 精度超过92% This photonic chip completes key tasks in under half a nanosecond with over 92% accuracy. 它可以通过克服以往光子硬件的局限性,使利达尔、天文学和电信等领域发生革命性变革。 It could revolutionize fields like lidar, astronomy, and telecommunications by overcoming previous limitations in photonic hardware.