The rapid development of artificial intelligence has led to more and more sophisticated models, but these systems still encounter fundamental performance challenges. A team of scientists led by Dr. Sina Yi, assistant professor at Texas A & m College of Engineering, has developed a new approach called Super-Touring AI, which imitates the ability of the human brain to learn and adapt. This innovation can significantly improve artificial intelligence, significantly reducing calculation costs and energy consumption.
Current AI models are based on architectures that separate data storage from processing, requiring huge computing power and energy to migrate information between these two components. On the other hand, the human brain integrates learning and memory by neural connections called synapses, which dynamically strengthen or weaken based on experience – a process known as synaptic plasticity.
The dr Yi team took inspiration from Neuronauka to develop AI systems that act more like biological brains. Traditional AI models largely depend on backward propagation, the optimization algorithm used to adapt neural networks during training. While effective, reverse propagation is intensively computing and biologically unbelievable.
To solve this problem, the team studies alternative mechanisms, such as learning Ebbbian-often summarized as “cells that shoot together, combine”-and plasticity depending on tension (STDP). These biologically inspired learning processes allow AI systems to strengthen connections based on activity patterns, reducing the need for constant retraining and excessive computing resources.
One of the most promising aspects of super-tourctic AI is his ability to efficiently process information in real time. In the last test, the circuit based on these learning principles enabled the drone to move around a complex environment without prior training. In contrast to traditional AI models, which require extensive data sets and pre -reening, this approach allowed the drone to adapt and teach in flight, showing faster reaction times and lower energy consumption.
Integration of neuromorphic processing, which imitates brain-like processing, increases the potential of super-proofing AI. By settling these learning algorithms in specialized equipment, scientists strive to develop AI systems that require minimal power while maintaining a high level of adaptation and intelligence.
The AI industry is growing rapidly, and companies are racing to develop larger and stronger models. However, scalability remains a burning challenge due to hardware restrictions and increasing energy demand. Some AI applications already require entire data centers, increasing both economic and environmental costs.
Dr. Yi emphasizes that the progress in the equipment is as crucial as AI software improvements. “Many people think that artificial intelligence affects almost algorithms, but without efficient computer equipment AI can not really evolve,” he explains. Super-turning AI offers a change of paradigm, combining hardware software and innovations to create balanced, scalable AI solutions.
Re-imagining AI architectures to reflect the performance of the human brain, Super-Turning AI is a significant step towards AI sustainable development. This technology can lead to a new generation of artificial intelligence, which is both more intelligent and responsible for the environment.
“Contemporary artificial intelligence, like chatgpt is powerful, but it is too expensive and energy -consuming. We are working on creating artificial intelligence, which is both smarter and more balanced,” says Dr. Yi. “Super-turning AI can transform the method of building and using AI, ensuring that its progress will benefit both people and the planet.”
You can examine the published research of the team in Scientific progress.
