Brain-Inspired AI Could Cut Energy Use and Boost Performance

Strategic Insights -

Researchers at the University of Surrey have developed a novel approach to artificial intelligence (AI) that takes direct inspiration from the human brain's neural networks. By mimicking the brain's sparse and structured neural wiring, the team has shown that AI systems can be made more energy-efficient and faster without sacrificing accuracy. The method, called Topographical Sparse Mapping (TSM), rethinks how AI systems are wired at their most fundamental level, eliminating the need for vast numbers of unnecessary connections and computations.

Key Takeaways:

  • The University of Surrey's Nature-Inspired Computation and Engineering (NICE) group has developed a new approach to AI called Topographical Sparse Mapping (TSM), which takes inspiration from the human brain's neural networks.
  • TSM connects each neuron only to nearby or related ones, unlike conventional deep-learning models that connect every neuron in one layer to all neurons in the next.
  • The enhanced version of TSM, called Enhanced Topographical Sparse Mapping (ETSM), introduces a biologically inspired "pruning" process during training, similar to how the brain refines its neural connections as it learns.
  • Surrey's enhanced model achieved up to 99% sparsity, meaning it could remove almost all of the usual neural connections, while still matching or exceeding the accuracy of standard networks on benchmark datasets.
  • The model trains faster, uses less memory, and consumes less than one percent of the energy of a conventional AI system.
  • The brain-inspired mapping can be extended to deeper layers of the network, making it even leaner and more efficient.
  • The research team is exploring how the approach could be used in other applications, such as more realistic neuromorphic computers, where the efficiency gains could have an even greater impact.

Statistics:

  • The AI models used in the study consumed over a million kilowatt-hours of electricity, equivalent to the annual use of over a hundred US homes and costing tens of millions of dollars.
  • The University of Surrey's enhanced model achieved up to 99% sparsity.
  • The model consumes less than one percent of the energy of a conventional AI system.

Sources:

  • "Brain-inspired AI could cut energy use and boost performance" (University of Surrey news release)
  • Neurocomputing (journal where the study was published)
  • https://doi-org.sdpl.idm.oclc.org/10.1016/j.neucom.2025.131740 (full paper)
  • https://www.surrey.ac.uk/news/brain-inspired-ai-could-cut-energy-use-and-boost-performance (original text)
  • "Topographical Sparse Mapping: A Novel Approach to Efficient Neural Network Design" (research paper by Mohsen Kamelian Rad et al.)