Engineers bring efficient optical neural networks into focus

EPFL researchers have published a programmable framework that overcomes a key computational bottleneck of optics-based artificial intelligence systems. In a series of image classification experiments, they used scattered light from a low-power laser to perform accurate, scalable computations using a fraction of the energy of electronics.

As digital artificial intelligence systems grow in size and impact, so does the energy required to train and deploy them -- not to mention the associated carbon emissions. Recent research suggests that if current AI server production continues at its current pace, their annual energy consumption could outstrip that of a small country by 2027. Deep neural networks, inspired by the architecture of the human brain, are especially power-hungry due to the millions or even billions of connections between multiple layers of neuron-like processors.

To counteract this mushrooming energy demand, researchers have doubled down on efforts to implement optical computing systems, which have existed experimentally since the 1980s. These systems rely on photons to process data, and although light can theoretically be used to perform computations much faster and more efficiently than electrons, a key challenge has hindered optical systems' ability to surpass the electronic state-of-the art.

"In order to classify data in a neural network, each node, or 'neuron', must make a 'decision' to fire or not based on weighted input data. This decision leads to what is known as a nonlinear transformation of the data, meaning the output is not directly proportional to the input," says Christophe Moser, head of the Laboratory of Applied Photonics Devices in EPFL's School of Engineering.

Moser explains that while digital neural networks can easily perform nonlinear transformations with transistors, in optical systems, this step requires very powerful lasers. Moser worked with students Mustafa Yildirim, Niyazi Ulas Dinc, and Ilker Oguz, as well as Optics Laboratory head Demetri Psaltis, to develop an energy-efficient method for performing these nonlinear computations optically. Their new approach involves encoding data, like the pixels of an image, in the spatial modulation of a low-power laser beam. The beam reflects back on itself several times, leading to a nonlinear multiplication of the pixels.

"Our image classification experiments on three different datasets showed that our method is scalable, and up to 1,000 times more power-efficient than state-of-the-art deep digital networks, making it a promising platform for realizing optical neural networks," says Psaltis.

The research, supported by a Sinergia grant from the Swiss National Science Foundation, has recently been published in Nature Photonics.

A simple structural solution

In nature, photons do not directly interact with each other the way charged electrons do. To achieve nonlinear transformations in optical systems, scientists have therefore had to 'force' photons to interact indirectly, for example by using a light intense enough to modify the optical properties of the glass or other material it passes through.

The scientists worked around this need for a high-power laser with an elegantly simple solution: they encoded the pixels of an image spatially on the surface of a low-power laser beam. By performing this encoding twice, via adjustment of the trajectory of the beam in the encoder, the pixels are multiplied by themselves, i.e., squared. Since squaring is a non-linear transformation, this structural modification achieves the non-linearity essential to neural network calculations, at a fraction of the energy cost. This encoding can be carried out two, three or even ten times, increasing the non-linearity of the transformation and the precision of the calculation.

"We estimate that using our system, the energy required to optically compute a multiplication is eight orders of magnitude less than that required for an electronic system," Psaltis says.

Read more at Science Daily

World's largest childhood trauma study uncovers brain rewiring

The world's largest brain study of childhood trauma has revealed how it affects development and rewires vital pathways.

The University of Essex study -- led by the Department of Psychology's Dr Megan Klabunde -- uncovered a disruption in neural networks involved in self-focus and problem-solving.

This means under-18s who experienced abuse will likely struggle with emotions, empathy and understanding their bodies.

Difficulties in school caused by memory, hard mental tasks and decision making may also emerge.

Dr Klabunde's cutting-edge research used AI to re-examine hundreds of brain scans and identify patterns.

It is hoped the research will help hone new treatments for children who have endured mistreatment.

This could mean therapists focus on techniques that rewire these centres and rebuild their sense of self.

Dr Klabunde said: "Currently, science-based treatments for childhood trauma primarily focus on addressing the fearful thoughts and avoidance of trauma triggers.

"This is a very important part of trauma treatment. However, our study has revealed that we are only treating one part of the problem.

"Even when a child who has experienced trauma is not thinking about their traumatic experiences, their brains are struggling to process their sensations within their bodies.

"This influences how one thinks and feels about one's 'internal world' and this also influences one's ability to empathise and form relationships."

Dr Klabunde reviewed 14 studies involving more than 580 children for the research published in Biological Psychiatry Cognitive Neuroscience and Neuroimaging.

The paper re-examined functional magnetic resonance imaging (fMRI) scans.

This procedure highlights blood flow in different centres, showing neurological activity.

The study discovered a marked difference in traumatised children's default mode (DMN) and central executive networks (CEN) -- two large scale brain systems.

The DMN and the posterior insula are involved in how people sense their body, the sense of self and their internal reflections.

New studies are finding the DMN plays an important role in most mental health problems -- and may be influenced by experiencing childhood trauma.

The CEN is also more active than in healthy children, which means that children with trauma histories tend to ruminate and relive terrible experiences when triggered.

Dr Klabunde hopes this study will be a springboard to find out more about how trauma affects developing minds.

She said: "Our brain findings indicate that childhood trauma treatments appear to be missing an important piece of the puzzle.

"In addition to preventing avoidance of scary situations and addressing one's thoughts, trauma therapies in children should also address how trauma's impacts on one's body, sense of self, emotional/empathetic processing, and relationships.

"This is important to do so since untreated symptoms will likely contribute to other health and mental health problems throughout the lifespan."

Read more at Science Daily

Artificial neural networks learn better when they spend time not learning at all

Depending on age, humans need 7 to 13 hours of sleep per 24 hours. During this time, a lot happens: Heart rate, breathing and metabolism ebb and flow; hormone levels adjust; the body relaxes. Not so much in the brain.

"The brain is very busy when we sleep, repeating what we have learned during the day," said Maxim Bazhenov, PhD, professor of medicine and a sleep researcher at University of California San Diego School of Medicine. "Sleep helps reorganize memories and presents them in the most efficient way."

In previous published work, Bazhenov and colleagues have reported how sleep builds rational memory, the ability to remember arbitrary or indirect associations between objects, people or events, and protects against forgetting old memories.

Artificial neural networks leverage the architecture of the human brain to improve numerous technologies and systems, from basic science and medicine to finance and social media. In some ways, they have achieved superhuman performance, such as computational speed, but they fail in one key aspect: When artificial neural networks learn sequentially, new information overwrites previous information, a phenomenon called catastrophic forgetting.

"In contrast, the human brain learns continuously and incorporates new data into existing knowledge," said Bazhenov, "and it typically learns best when new training is interleaved with periods of sleep for memory consolidation."

Writing in the November 18, 2022 issue of PLOS Computational Biology, senior author Bazhenov and colleagues discuss how biological models may help mitigate the threat of catastrophic forgetting in artificial neural networks, boosting their utility across a spectrum of research interests.

The scientists used spiking neural networks that artificially mimic natural neural systems: Instead of information being communicated continuously, it is transmitted as discrete events (spikes) at certain time points.

They found that when the spiking networks were trained on a new task, but with occasional off-line periods that mimicked sleep, catastrophic forgetting was mitigated. Like the human brain, said the study authors, "sleep" for the networks allowed them to replay old memories without explicitly using old training data.

Memories are represented in the human brain by patterns of synaptic weight -- the strength or amplitude of a connection between two neurons.

"When we learn new information," said Bazhenov, "neurons fire in specific order and this increases synapses between them. During sleep, the spiking patterns learned during our awake state are repeated spontaneously. It's called reactivation or replay.

"Synaptic plasticity, the capacity to be altered or molded, is still in place during sleep and it can further enhance synaptic weight patterns that represent the memory, helping to prevent forgetting or to enable transfer of knowledge from old to new tasks."

When Bazhenov and colleagues applied this approach to artificial neural networks, they found that it helped the networks avoid catastrophic forgetting.

"It meant that these networks could learn continuously, like humans or animals. Understanding how human brain processes information during sleep can help to augment memory in human subjects. Augmenting sleep rhythms can lead to better memory.

"In other projects, we use computer models to develop optimal strategies to apply stimulation during sleep, such as auditory tones, that enhance sleep rhythms and improve learning. This may be particularly important when memory is non-optimal, such as when memory declines in aging or in some conditions like Alzheimer's disease."

Read more at Science Daily

Even smartest AI models don't match human visual processing

Deep convolutional neural networks (DCNNs) don't see objects the way humans do -- using configural shape perception -- and that could be dangerous in real-world AI applications, says Professor James Elder, co-author of a York University study published today.

Published in the Cell Press journal iScience, Deep learning models fail to capture the configural nature of human shape perception is a collaborative study by Elder, who holds the York Research Chair in Human and Computer Vision and is Co-Director of York's Centre for AI & Society, and Assistant Psychology Professor Nicholas Baker at Loyola College in Chicago, a former VISTA postdoctoral fellow at York.

The study employed novel visual stimuli called "Frankensteins" to explore how the human brain and DCNNs process holistic, configural object properties.

"Frankensteins are simply objects that have been taken apart and put back together the wrong way around," says Elder. "As a result, they have all the right local features, but in the wrong places."

The investigators found that while the human visual system is confused by Frankensteins, DCNNs are not -- revealing an insensitivity to configural object properties.

"Our results explain why deep AI models fail under certain conditions and point to the need to consider tasks beyond object recognition in order to understand visual processing in the brain," Elder says. "These deep models tend to take 'shortcuts' when solving complex recognition tasks. While these shortcuts may work in many cases, they can be dangerous in some of the real-world AI applications we are currently working on with our industry and government partners," Elder points out.

One such application is traffic video safety systems: "The objects in a busy traffic scene -- the vehicles, bicycles and pedestrians -- obstruct each other and arrive at the eye of a driver as a jumble of disconnected fragments," explains Elder. "The brain needs to correctly group those fragments to identify the correct categories and locations of the objects. An AI system for traffic safety monitoring that is only able to perceive the fragments individually will fail at this task, potentially misunderstanding risks to vulnerable road users."

Read more at Science Daily

Deep neural network to find hidden turbulent motion on the sun

Scientists developed a neural network deep learning technique to extract hidden turbulent motion information from observations of the Sun. Tests on three different sets of simulation data showed that it is possible to infer the horizontal motion from data for the temperature and vertical motion. This technique will benefit solar astronomy and other fields such as plasma physics, fusion science, and fluid dynamics.

The Sun is important to the Sustainable Development Goal of Affordable and Clean Energy, both as the source of solar power and as a natural example of fusion energy. Our understanding of the Sun is limited by the data we can collect. It is relatively easy to observe the temperature and vertical motion of solar plasma, gas so hot that the component atoms break down into electrons and ions. But it is difficult to determine the horizontal motion.

To tackle this problem, a team of scientists led by the National Astronomical Observatory of Japan and the National Institute for Fusion Science created a neural network model, and fed it data from three different simulations of plasma turbulence. After training, the neural network was able to correctly infer the horizontal motion given only the vertical motion and the temperature.

The team also developed a novel coherence spectrum to evaluate the performance of the output at different size scales. This new analysis showed that the method succeeded at predicting the large-scale patterns in the horizontal turbulent motion, but had trouble with small features. The team is now working to improve the performance at small scales. It is hoped that this method can be applied to future high resolution solar observations, such as those expected from the SUNRISE-3 balloon telescope, as well as to laboratory plasmas, such as those created in fusion science research for new energy.

From Science Daily