The Whale That Outswam Evolution: Swarm Intelligence Maximises Memory in Connectome Reservoirs
Researchers have explored the use of bio-inspired optimization algorithms to enhance the memory capabilities of neural networks based on biological connectomes. By applying techniques like Particle Swarm Optimization and Whale Optimization Algorithm to the synaptic weights of echo-state networks derived from six different species, they found significant improvements in performance. The Whale Optimization Algorithm, in particular, demonstrated substantial gains, including a 17x increase in memory capacity in some cases. AI
IMPACT Demonstrates a novel method for enhancing neural network memory using bio-inspired optimization, potentially leading to more efficient AI systems.