Tag Archives: Lightbased

Low-Power Light-Based AI Image Generators

Posted on by
Reply

Vibrant Vincent van Gogh-inspired artwork generated by a traditional diffusion model (left in each set of three) and an optical image generator (right).

Shiqi Chen et al. 2025

AI image generators that utilize light for image creation instead of conventional computing hardware can consume significantly more energy—potentially hundreds of times greater.

When an AI model generates an image from text, it typically employs a method known as diffusion. The AI first presents a collection of vast datasets, demonstrates how to use statistical noise to disrupt them, and then encodes these patterns using a specific set of rules. For a new noisy image, these rules can be applied to reverse the process, with several steps resulting in a coherent image that aligns with a specific text prompt.

For producing realistic, high-resolution images, diffusion requires numerous sequential steps that depend on considerable computational power. In April, OpenAI reported that its new image generator created more than 700 million images within its first week. Achieving this scale requires an enormous amount of energy and water to both power and cool the machinery supporting the model.

Now, Aydogan Ozcan at the University of California, Los Angeles, along with his colleagues, has designed a diffusion-based image generator that operates using light beams. The encoding phase is digital and requires minimal energy, while the decoding phase is wholly optical and demands no computational resources.

“Unlike traditional digital diffusion models that need hundreds or thousands of iterative steps, this method accomplishes image generation with snapshots and requires no additional calculations beyond the initial encoding,” states Ozkan.

The system initially utilizes a digital encoder trained on a publicly available image dataset. This mechanism can produce patterns that can be transformed into images. This encoder then engages a liquid crystal display known as a spatial light modulator (SLM) to physically imprint the static laser beam. When the laser beam travels through a second decoding SLM, the desired image is instantly produced on the screen recorded by the camera.

Ozkan and his team employed this system to generate black-and-white images of simple objects, such as the digits 1-9, to test the diffusion model, as well as vivid images in the style of Vincent van Gogh. The outcomes appeared to be comparably similar to those generated by conventional image synthesis methods.

“This might represent the first instance of optical neural networks serving as a computational tool that can produce not only lab-based demos but also practical results,” remarks Alexander Lvovsky from Oxford University.

For the Van Gogh-inspired images, the system consumed merely a few millijoules of energy per image, primarily for the liquid crystal screens, contrasting sharply with the hundreds or thousands of joules necessary for conventional diffusion models. “This indicates that the latter consumes energy equivalent to that of an electric kettle, while the optical system only uses millionths of a joule,” notes Lvovsky.

The system will need modifications to function in a data center environment compared to widely adopted image generation tools, but Ozcan believes it could also be suitable for portable electronics like AI glasses, thanks to its low energy demands.

Topic:

Source: www.newscientist.com

Light-based computers are nearing their commercial debut

Posted on by
Reply

Lightweight based computer chip made by Pace, LightElligence

Light Ergens

Computers that use light rather than data to represent and manipulate data can reduce data center power requirements and at the same time speed up calculations. Two studies published today describe breakthroughs in performing real problems on light-based computers, creating techniques that are on the verge of commercial applications, the researchers say.

Electronic computers have historically followed Moore’s law, as we all use today. The power of the machine doubled every two years. However, in recent years, progress has slowed down as transistor miniaturization reaches its fundamental physical limits.

Researchers are working on many potential solutions, including quantum and photonic computing. However, Quantum Computing still struggles to achieve true utility, but Photonic Computing has reached the point where chip designs like those set in two new research are performing authentic calculations. In addition, these photonic chips can be manufactured using the same factory that manufactures silicon chips for electronic computers.

Photonic computers offer greater potential benefits than electronic computers. One is that photons travel faster than electrons do in the circuit, allowing for faster calculations and less pauses between each step of the calculation. Second, photons move without resistance and are rarely absorbed by the material on which the chip is made, allowing the same job to be performed using less energy than an electric computer that requires energy-intensive cooling.

In its research, Lightelligence, a Singapore-based company, shows that a device called a Photonic Arithmetic Computing Engine (PACE), which combines photonic and microelectronic chips, can successfully execute ISING problems that apply directly to the logistics industry and many other areas.

Meanwhile, US startup LightMatter claims that its own chip can run AI model BERT to create text in Shakespeare’s style. New Scientist Could not reach Lightmatter due to comments.

Bo Peng At LightElligence, the sector is increasingly busy with start-ups and technology is rapidly maturing. “We’re more or less pre-production,” says Peng. “It’s more like a real product than just a lab demonstration.”

Just as the world of quantum computers is trying to demonstrate the benefits of quantum, quantum machines are the point where classical computers can provide useful things. He won’t draw when this will happen, but says that this technology is closer to being ready for commercial applications – perhaps it works as a photonic chip that works with the electric chip, rather than completely replacing them to handle the specific tasks that it can provide boost.

Needless to say, hardware based on the research and Lightelligence PCI Express format. This is a standard motherboard add-on format for desktop computers that allow you to add graphics cards and other devices. Company devices can already be added to any commercial desktop, but require the appropriate software to communicate.

Robert Hadfield At the University of Glasgow in the UK, two studies show that “it’s a kind of boiling area.” “This is close to the point where the industry may consider photonic processors a viable alternative,” he says. “It’s really interesting to see how mature this architecture has become. These are photonic chips manufactured in one of the world’s leading foundries, so they can be expanded for mass production.”

Stephen SweeneyThe University of Glasgow also says that they have already seen optical data transmissions roll out around the world, with optical optical computing approaching too. “With Photonics, you can do things at a lower loss than electronics can,” says Sweeney. “And if you need to be able to do a huge amount of calculations, you need to start looking at it.”

topic:

Source: www.newscientist.com