A new approach developed at the University of Surrey takes direct inspiration from biological neural networks of the human brain./SCREENGRABUniversity of Surrey researchers have developed a new approach to improve artificial intelligence (AI) performance by mimicking the networks of the human brain.
According to a study published in Neurocomputing, mimicking the brain's neural wiring can significantly improve the performance of artificial neural networks used in generative AI and other modern AI models such as ChatGPT.
Topographical Sparse Mapping connects each neuron only to nearby or related neurons, similarly to how the human brain organises information efficiently.
Dr Roman Bauer, senior lecturer, said: "Our work shows that intelligent systems can be built far more efficiently, cutting energy demands without sacrificing performance."
Researchers said the model eliminated the need for vast numbers of unnecessary connections, improving performance in a more sustainable way without sacrificing accuracy.
Dr Bauer added: "Training many of today's popular large AI models can consume over a million kilowatt-hours of electricity.
"That simply isn't sustainable at the rate AI continues to grow."
An enhanced version called Enhanced Topographical Sparse Mapping goes a step further by introducing a biologically inspired "pruning" process during training.
This is similar to how the brain gradually refines its neural connections as it learns.
The research team is also exploring how the approach could be used in other applications, such as more realistic neuromorphic computers - a computing approach inspired by the human brain's structure and function.
