Brain Inspired AI Breakthrough Cuts Energy Use by 99% Without...
Dr Singularity@Dr_Singularity
38 views
Nov 02, 2025
~1 min read
Advertisement
1
Brain Inspired AI Breakthrough Cuts Energy Use by 99% Without Losing Accuracy
Researchers from the University of Surrey’s NICE group developed a new AI wiring method called Topographical Sparse Mapping (TSM), inspired by how the brain connects neurons efficiently.
Instead of linking every neuron to all others, TSM connects only nearby or related ones, cutting unnecessary computations.
An improved version, Enhanced TSM (ETSM), mimics the brain’s pruning process during learning.
This approach achieves up to 99% sparsity, matching or surpassing standard AI accuracy while using less than 1% of the energy and training faster.
This method could lead to far more efficient AI and neuromorphic computing systems.
Researchers from the University of Surrey’s NICE group developed a new AI wiring method called Topographical Sparse Mapping (TSM), inspired by how the brain connects neurons efficiently.
Instead of linking every neuron to all others, TSM connects only nearby or related ones, cutting unnecessary computations.
An improved version, Enhanced TSM (ETSM), mimics the brain’s pruning process during learning.
This approach achieves up to 99% sparsity, matching or surpassing standard AI accuracy while using less than 1% of the energy and training faster.
This method could lead to far more efficient AI and neuromorphic computing systems.
2
source surrey.ac.uk/news/brain-ins…
