Inference in artificial intelligence with deep optics and photonics

By Gordon Wetzstein, Aydogan Ozcan, Sylvain Gigan, Shanhui Fan, Dirk Englund, Marin Soljačić, Cornelia Denz, David A. B. Miller, Demetri Psaltis. Nature, 2020

901 citations recorded by the source corpus.

Artificial intelligence tasks across numerous applications require accelerators for fast and low-power execution. Optical computing systems may be able to meet these domain-specific needs but, despite half a century of research, general-purpose optical computing systems have yet to mature into a practical technology. Artificial intelligence inference, however, especially for visual computing applications, may offer opportunities for inference based on optical and photonic systems. In this Perspective, we review recent work on optical computing for artificial intelligence applications and discuss its promise and challenges. Recent work on optical computing for artificial intelligence applications is reviewed and the potential and challenges of all-optical and hybrid optical networks are discussed.

Subfields: Artificial Intelligence, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/5f6b2a12cb4fa2e8509736ca45091f17491ada75

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Peter Leonard McMahon
2. Dirk Englund
3. Yichen Shen

In Electrical and Electronic Engineering

1. Peter Leonard McMahon
2. Dirk Englund
3. M. Eichenfield

Citations in the corpus

• The physics of optical computing (2023, methodology): “There is also a substantial thrust in performing computations for neural networks that are not explicitly engineered to be matrix–vector multiplications or convolutions 1,14–18 .”
• Photonic matrix multiplication lights up photonic accelerator and beyond (2022, background): “Electrical input vector Xconv[n] Wavelength, i P ow er Wconv ((3))[25] Wconv ((2))[25] Wconv ((1))[25] Wconv ((3))[1] Wconv ((2))[1] Wconv ((1))[1] Wconv [R − i + 1]”
• High-speed programmable photonic circuits in a cryogenically compatible, visible–near-infrared 200 mm CMOS architecture (2021, methodology): “1a for a 4-mode programmable SU(4) in a rectangular mesh arrangement.”
• Deep physical neural networks trained with backpropagation (2021, neutral): “9, 10).”
• Integrated electro-optics on thin-film lithium niobate (2024, neutral): “Conversely, photonics enriches electronics with the massively parallel processing capabilities, evident in advancements in photonic computing 30–34 and microwave photonics 35–39 , for example.”