Space-efficient optical computing with an integrated chip diffractive neural network. Nat. Commun. 13, 1044 (2022). This article has been highly influential for neuromorphic hardware, which demonstrates a space-efficient optical computational unit that realizes on-chip diffractive neural network. Article...
These modes are configured by modulating the charge density within the oxygen vacancies via synergistic optical and electrical operations, as confirmed by differential phase-contrast scanning transmission electron microscopy. Using this OEM system, three visual processing tasks are demonstrated: image sensory...
To perform highly parallel optical computing using these, or any other kind of device, will require large arrays of optical processor elements, often involving the integration of a number of devices within each pixel, and probably involving the cascading of a number of such arrays, optically inter...
A local weight update mechanism is required, directly fetching the signals in the network itself. Here, we first summarize the backward propagation algorithm as this helps to understand the merits of the optical signal processor presented thereafter. To train a feedforward DNN, we can use ...
In the field of AI, the high-throughput and large-scale matrix operations required by AI processor chips can be completed by silicon optical neural network computing units [160]. It has been shown that optical neural network chips have two orders of magnitude faster than tradi- tional ...
An integrated circuit processor architecture that implements digital signal processing (DSP) functions with less hardware, improved speed and a more efficient layout. The CPU resources are used in conjunction with an integrated multiply/accumulate unit to perform DSP operations. Use of the CPU's inter...
The advantage of this photonic arithmetic processor is the short (10's ps) computational execution time given by the optical propagation delay through the integrated nanophotonic router. Furthermore, we show how photonic processing in-the-network leverages the natural parallelism of optics such as ...
Using multiple integration technology, large-scale integrated circuit design and manufacturing methods of system integration are used. This enables the mechanical components and the sensors, actuators, microprocessors, and other electronic and optical systems to be integrated in a very small space such as...
The emergence of parallel convolution-operation technology has substantially powered the complexity and functionality of optical neural networks (ONN) by harnessing the dimension of optical wavelength. However, this advanced architecture faces remarkable
(activation function). In a departure from the electronic neural network, our introduced IDNN framework implements a convolution transformation physically in the optical field using PICs. The convolution transformation is special matrix multiplication, and the complex-valued matrix elements are circulant41...