Other Applications

Model Uncertainty in CNNs

Industry is becoming increasingly more reliant on computer vision, with large numbers of diverse applications like robotic surgery and autonomous vehicles. Convolutional neural networks (CNNs) are the typical architecture that underpins such systems. This article explores the use of Bayesian neural networks for safety-critical machine vision tasks, and how optical processing could have some unique advantages over conventional electronics…

Optical computing for computational fluid dynamics: An introduction to CFD

In this article, we aim to provide an illustration of where the field is now so that we can later talk in detail about where it might end up in the future. If you’re in a hurry, there’s a summary you can skip to at the end of this article. If not, we start by describing a method of making approximations for complex problems…

Deep learning and fluid dynamics: Solving the Navier-Stokes equations with an optical neural network

Modern CFD techniques give us impressive capabilities when it comes to designing things like safer aircraft, more efficient wind turbines and better streamlined cars, but they also come at an incredibly high cost in money, time and energy. We asked if there was a better way of doing this.

There is…

Simulating wave propagation with the optical Fourier engine

It’s no exaggeration to say that the concept of a wave is one of the most useful ideas in engineering and physics. Waves are so omnipresent in the world around us that just about everyone has an idea of what they are. Some might think of the waves on the ocean surface, while others might think of their use in the vast network of wireless transmitters and receivers that surrounds most of us in daily life. The ability to work out how waves travel, convey information from one point to another, or dissipate away, is essential to their practical use. However, in most cases, working this out isn’t a trivial task…

Self-driving vehicles and Fourier-optical computing: A way forward?

While our previous work usually focuses (often in great detail) on whether we can perform specific operations, it’s also worth looking at the broader picture and considering what a dramatic shift in computational efficiency might mean for the world around us.

We start by considering one of the most challenging applications for high-performance computing:

Self-driving vehicles…

Neurosymbolic AI as a self-driving solution: Convolutional Logic Tensor Networks

Single end-to-end neural networks have some clear limitations, one of which is an inability to represent first order logic. One of the recent attempts to fuse ideas from both formal logic and deep learning are so-called Logic Tensor Networks (LTNs)

Optical computing for predicting vortex formation

In this article, we see whether deep neural networks can be used in combination with our existing hardware to detect the formation of small features, such as vortices, when the simulated domain is much larger. We will concentrate on the 2D case as an illustration and see how it can be used to build neural networks for detecting and predicting the formation of vortices, a task which is highly sensitive to the accuracy of a machine learning approach

AI super resolution using an optical computer

In this article we will discuss super resolution, an AI technique that uses deep neural networks to increase the resolution of images and videos. Gone are the days when the command “Enhance!” from the TV show C.S.I. would cause all but the most open minded to scoff uncontrollably.

Today, we really can enhance and upscale low resolution data using the power of deep learning.

Semantic segmentation using an optical computer

Convolutional neural networks have many uses; we’ve covered some of these in detail, from the conventional through to somewhat more unusual applications. Along the way, we’ve been examining how we can disrupt this field through the use of a technology that allows us to apply a useful piece of mathematics, the convolution theorem, to wildly shift the computational effort and power consumption involved in executing these networks.

In this article, we will take a look at very popular and powerful application of convolutional networks, semantic segmentation, which is one of the most challenging yet useful computer vision tasks…

Attention: Fourier Transforms. A giant leap in transformer efficiency.

In this article we will discuss a promising neural network architecture, FNet, proposed recently by Google research. FNet is a interesting approach to neural network design, in that it replaces the traditional attention mechanism in transformer networks with a novel operation: the Fourier transform (FT)…


Sign up with your email address to receive news and updates from Optalysys.

We respect your privacy.