News Center

Photoelectric superconducting integrated circuit, enabling next generation AI

Release time:

2021-04-10 15:24

Source:

Experts in the field of artificial intelligence have long aimed to build AI systems with human intelligence,which requires high-performance hardware.At present,the AI models for large-scale application and research mostly rely on semiconductor chips.In order to accelerate computing efficiency and enhance computing power,some scientists are trying to combine optical devices with silicon chips,but the manufacturing process is plagued by physical limitations and practical problems.

In order to build the next generation of artificial intelligence systems,Jeffrey Shainline,an engineer and researcher at the National Institute of Standards and Technology(NIST)in Colorado,US,recently called on scientists in the field of AI to abandon the use of semiconductor hardware in favor of superconducting hardware for model development.In his latest opinionated paper,he proposes a new type of general-purpose artificial intelligence(AGI)hardware that combines optics and superconductors,and explores the possibility of using it for large-scale research.

intelligence,hardware

Previous Page

A breakthrough in the mass production of triaurora, the first single piece of 40°FOV volume holographic diffraction waveguide can be mass produced

Set a new industry record! China Star Power Co., Ltd. and South China University of Technology have successfully developed the highest luminous efficiency of quantum dot LED devices in the industry

Next Page

Previous Page

A breakthrough in the mass production of triaurora, the first single piece of 40°FOV volume holographic diffraction waveguide can be mass produced

Next Page

Set a new industry record! China Star Power Co., Ltd. and South China University of Technology have successfully developed the highest luminous efficiency of quantum dot LED devices in the industry