- 17:00Israel vows to strike Iran again if threatened, defense minister warns
- 16:20Morocco sees record 8.9 million tourists in first half of 2025
- 15:50Morocco aims to secure five-month wheat stock amid price drop
- 15:20Macron and Starmer to sign historic nuclear deterrence pact
- 14:50Chemsedine Talbi completes Sunderland move after stellar Club Brugge season
- 14:20Police in Marrakech detain Algerian-French suspect wanted by Interpol
- 13:30EU unveils stockpiling plan to prepare for crises and conflict
- 12:50Spain eases lockdowns as firefighters stabilize Catalonia forest blaze
- 12:20Spaniards favor China over the US in shifting global perceptions
Follow us on Facebook
Brain-Like Supercomputer to Revolutionize AI Development
A groundbreaking supercomputer, designed to replicate the energy-efficient computations of the human brain, is poised to catalyze advancements in AI research and development.
Unveiled by the International Neuromorphic Systems Centre (ICNS) at Western Sydney University, the appropriately named "DeepSouth" supercomputer is set to debut in April 2024, boasting the ability to process an impressive 228 trillion synaptic operations per second. This brain-inspired marvel aims to simulate intricate neural networks while drastically reducing energy consumption compared to conventional supercomputers.
The comparison to the human brain is nothing short of remarkable. With a mere 20 watts of power, the brain achieves the equivalent of an exaflop – a staggering quintillion floating-point operations per second. DeepSouth's primary objective is to match this computational prowess while minimizing energy wastage.
“DeepSouth represents a pivotal moment in leveraging the inherent efficiency of the brain to enhance data processing for AI and other disciplines,” remarked ICNS Director Dr. Emma Johnston. “At last, we have a supercomputer customized for the future.”
Mimicking Neurobiology for Smarter Computing
DeepSouth challenges the norm of supercomputers that simulate neural networks using graphics and multicore processors. Although powerful, such traditional configurations are energy-intensive compared to the modest energy requirements of the brain.
By closely emulating neurobiological architecture, DeepSouth aims to overcome this energy barrier, enabling transformative applications in biomedicine, robotics, aerospace, and, notably, artificial intelligence.
The moniker "DeepSouth" pays homage to IBM’s revolutionary TrueNorth supercomputer, which pioneered the simulation of expansive neural networks in hardware. It also draws parallels with DeepBlue, the iconic supercomputer that triumphed over World Chess Champion Garry Kasparov 25 years ago.
Diverging from its predecessors, DeepSouth’s neuromorphic design is adaptable, allowing its architecture to scale up or down efficiently to meet computing demands. This flexibility, coupled with the energy savings inspired by the brain, has the potential to rapidly propel AI development in the years to come.
As Dr. Johnston succinctly put it: “By unlocking more of the brain’s secrets, we are also unlocking AI’s full potential.”