England Launches CIP AI 2,000 Times More Energy Saving
new innovations in the development of artificial intelligence systems
Researchers from Loughborough University in the UK have created a new computer chip or semiconductor inspired by the workings of the human brain. This technology has great potential to make the artificial intelligence (AI) system much more energy efficient, it can even achieve up to 2,000 times better efficiency on certain tasks.
Unlike the AI system which generally relies on software on the computer, this chip is able to process data that changes over time directly on the hardware. This approach not only accelerates the process, but also reduces energy consumption significantly.
The research was led by Pavel Borisov and supported by the Engineering and Physical Sciences Research Council. According to Borisov, this approach opens up new ways of designing AI systems. He explained that by using physical processes rather than being completely dependent on software, we can dramatically reduce the energy required for this type of task.
Memristor structure that reflects the human brain
This chip uses a special device called memristor made of niobium oxide which has a very small structure with random pores. This structure mimics complex networks in the human brain, thus being able to handle various tasks such as simple logic operations, image recognition, and time-based predictions.
One of the trials was carried out using the Lorenz-63 System mathematical model, which is known for describing the “butterfly effect”, which is a small change that can produce big impacts. In these tests, the CHIP is able to predict short-term patterns and reconstruct the lost data quite accurately.
In addition, this chip also manages to recognize simple numbers from pixelated images and perform basic logic operations. This shows that one device can be used for various types of AI tasks.
significant energy efficiency advantages
The main advantage of this technology is its energy efficiency. By imitating the way the brain forms complex and random neural connections, the researchers succeeded in creating a system that was not only intelligent, but also very power-efficient. Borisov explained that this technology was inspired by the way the human brain forms very many neuronal connections and seems random among its neurons.
future potential in neuromorphic electronics
In the future, this technology has the potential to become the basis for the development of neuromorphic electronics, namely devices that are designed to resemble the way the brain works and are able to process time signals efficiently for various modern AI applications.
With this innovation, the world of technology is getting closer to the ability of the human brain to process information and make decisions. Researchers believe that this technology will be an important foundation in the development of AI in the future, especially in areas that require high speed and energy efficiency.
