Researchers at the Salk Institute say they have found parallels between an Internet algorithm and human brain activity that improve the understanding of engineered and neural networks, with potential insight into learning disabilities.
Internet users' ability to find their "sweet spot" in which to channel data while avoiding congestion is enabled by the additive increase, multiplicative decrease (AIMD) algorithm. As computers in a network employ this approach, the entire system can constantly maximize efficiency by adapting to changing conditions.
Salk professor Saket Navlakha and Duke University's Jonathan Suen mathematically modeled neural activity using the AIMD template, as the algorithm efficiently maintained smooth information flow, adjusting traffic rates whenever pathways got too clogged. AIMD also could best explain what was occurring in neurons experimentally, and Suen says both the brain and the Internet follow simple rules facilitating global stability.
From Salk News
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA
No entries found