Can shifting computation into hardware cut AI power use? This chip design can focus on efficiency in time-based data processing.
Researchers at Loughborough University have developed a brain-inspired computing chip that can make certain artificial intelligence (AI) tasks up to 2,000 times more energy efficient. The device processes time-dependent data directly in hardware, reducing reliance on traditional software-based computation and offering a potential path toward more sustainable AI systems.
The approach addresses a growing challenge in AI development. As models become more powerful, their energy consumption continues to rise, particularly for tasks involving pattern recognition and prediction over time. By shifting part of this computation into physical hardware, the new chip can significantly lower energy requirements while maintaining functional accuracy. This could benefit applications such as sensor data analysis, weather prediction, and biological modeling, where continuous data processing is essential.
At the core of the innovation is a memristor-based design built from nanoporous oxide materials. These materials form complex, random electrical pathways that mimic the behavior of neural networks, enabling the chip to perform a form of reservoir computing directly within the hardware. Instead of processing data through multiple software layers, the material itself transforms incoming signals into patterns that can be more easily analyzed.
In experimental tests, the system successfully handled tasks including predicting chaotic system behavior, reconstructing missing data, recognizing simple images, and performing basic logic operations. The results demonstrate that a single device can support multiple functions while maintaining high efficiency.
Dr Pavel Borisov, Senior Lecturer in Physics, who led the research team funded by the Engineering and Physical Sciences Research Council (EPSRC), says, “Inspired by the way the human brain forms very numerous and seemingly random neuronal connections between all its neurons, we created complex, random, physical connections in an artificial neural network by designing pores in nanometre-thin films of niobium oxide as part of a novel electronic device”