menu
techminis

A naukri.com initiative

google-web-stories
Home

>

ML News

>

Photonic p...
source image

Mit

3w

read

122

img
dot

Image Credit: Mit

Photonic processor could enable ultrafast AI computations with extreme energy efficiency

  • Scientists from MIT have demonstrated a fully integrated photonic processor that can perform all the key computations of a deep neural network optically on the chip, making it an ideal candidate for computationally demanding applications like high-speed telecommunications, scientific research, and lidar.
  • Deep learning models require a lot of energy as they involve many layers of computation that can push the limits of traditional electronic computing hardware.
  • Photonic processors, however, can perform these computations with light, offering faster and more energy-efficient alternatives.
  • The newly developed optical device can complete the key computations for a machine-learning task in less than half a nanosecond, while achieving more than 92% accuracy.
  • The chip is composed of interconnected modules that form an optical neural network and is fabricated using commercial foundry processes, which could enable scalable integration into electronics.
  • The photonic processor may lead to faster and more energy-efficient deep learning for computationally demanding applications, such as scientific research and high-speed telecommunications.
  • Designing devices that combine nonlinear functions through electronics and optics called nonlinear optical function units (NOFUs) has enabled the chip to perform nonlinear operations on the chip.
  • The entire circuit was fabricated using the same infrastructure and foundry processes that produce CMOS computer chips, which can make the chip manufacturable at scale using tried-and-true techniques with minimal errors.
  • The researchers want to explore algorithms that can leverage the advantages of optics to train systems faster and with better energy efficiency, as well as scaling up their device and integrating it with real-world electronics like cameras or telecommunications systems.
  • This research was funded, in part, by the U.S. National Science Foundation, the U.S. Air Force Office of Scientific Research, and NTT Research.

Read Full Article

like

7 Likes

For uninterrupted reading, download the app