The brain is arguably the world’s greatest computer. Sure, it doesn’t act like a computer, and it certainly doesn’t look like your average laptop – it's far more squishy for one thing. But in computing power alone, it is unmatched.
Capable of executing as many as 100 billion operations per second, the brain model makes your average modern computer look dated. So why not make computers more like the grey matter between our ears?
Well, people have been trying. Since the 1980s, the idea of neuromorphic computing (modelling machines on the human brain and nervous system) has gained traction.
There have been plenty of attempts over the years, but while many have advanced the technology, none have come close to a complete model. Until now.
DeepSouth, a project being run out of the International Centre for Neuromorphic Systems at Western Sydney University, is set to go online in April 2024. If the team behind it are to be believed, it could perform a staggering 228 trillion actions per second (that's over 2000 more than you can do).
undefined
How does a brain-computer work?
Well, we don't know an awful lot at the moment – DeepSouth has kept the current state of the project fairly under wraps. But there is enough detail to understand where it's going.
“DeepSouth is inspired, not just by the brain but the body, essentially mimicking and using the same kind of approach that the brain takes,” Domenico Vicinanza, an associate professor of intelligent systems at Anglia Ruskin University, told BBC Science Focus.
“What makes our brains unique is a very different approach to computation compared to a lot of other devices, such as smartphones which use a microprocessor with all of the power concentrated in one place and the memory found elsewhere.”
By keeping memory and processing power separate, there is a lot of back and forth in computers taking up a huge amount of energy to accomplish. To try to resolve this, researchers looked at a way of combining the two systems, resembling what goes on under your dome.
“The brain is so powerful because every single neuron and synapse is a little bit of computing and memory together. The power is in connecting all of these trillions of points, while also being able to reconfigure them at any time,” said Vicinanza.
“The brain takes little power to run and it changes when we learn new skills and open new synapses. We could utilise this model to make computers run more efficiently and be able to achieve more.”
Essentially, by distributing power to billions of small units (our neurons) that interact through trillions of different connections (synapses), the brain can battle the most powerful computers while using next to no energy.
This is the important part. Supercomputers take up both huge amounts of space and energy. The brain, on the other hand, keeps things light.
What does this mean for the future of computers?
This all sounds like quite a complicated way to design a computer. So what is the point (other than scientific bragging rights)? Well, there are a couple of potential benefits to the tech that we use.
By condensing the technology, and drastically improving its energy efficiency, we can improve the quality of batteries. Smartphones that last for a week instead of a day, for example.
It also means being able to drastically reduce the size of powerful computers, condensing some of the best technology into smartphones, tablets and small laptops.
With this technological advancement, personal devices could be better equipped to deal with tasks. Advanced programs could be run in the background, like antivirus checks, or simply the ultimate dream – hundreds of tabs without the intense lag.
This doesn’t just mean a faster smartphone, but the ability to bring high-powered computing to places that weren’t possible before. Better life support systems in hospitals, more advanced AI models in all fields and even areas like the aviation industry where huge streams of data need to be managed.
How DeepSouth could help fight ageing
The end goal of DeepSouth is to make computers better, but there is a side effect. By studying the brain so intensely, and trying to understand its ability to act as a computer, scientists are learning more about how the brain works.
Neuromorphic computing offers a glimpse into the mechanisms of the brain. It could give us a better understanding of illnesses that affect the brain, such as Alzheimer’s, dementia or Parkinsons'.
“By replicating the very intricate network of units that make up the brain, we can understand the uncertainty of the brain, and the dynamics of certain diseases,” said Vicinanza.
“We could even understand how our brain reacts to drugs, or how our brain develops as we age.”
This might seem like a strange idea, that a computer could give us an insight into the brain, but as we’ve mentioned, that's all the brain is – a very powerful and complicated computer.
It just takes some reframing of the issue. Think of diseases as a virus affecting a computer. To deal with a computer virus, we find the cause, and follow the route, seeing where it originated and where it has gone.
By mirroring the brain in a computer system, we can see how neurons translate to computer signals. How they communicate, when and why they spike and how problems can arise.
Take learning a new skill, for example. This fundamentally changes the neural networks in our brains. However, if you were to look at this action in an MRI or brain scan, we can only see what part of the brain has been activated – the end result essentially.
Neuromorphic computing can help us understand the route that got us there. “Having a model which in principle can access all the little individual processors and memory units can help to understand how the brain forms,” said Vicinanza.
“This, with enough analysis, could help us to better understand the ageing process, drug treatment, disease or even just our reactions to certain prompts.”
It all sounds like a miracle, a way to get ahead of all the issues that are inflicted on the brain, but right now it is entirely theoretical. First, we need to see if DeepSouth can work. Then we can turn our attention to the somewhat incredible art of understanding disease through a digitally created brain.
Read more:
