A new study has found an intriguing correlation between the functions of the internet and neural activity in the human brain. Hailed by many neurologists as a groundbreaking initiative, the functional similarity between the brain and the internet can potentially open new avenues into the probing of Artificial Intelligence development.
Researchers have discovered that neurons operate, on a certain level of activity, guided by the principles of a mechanism deployed on the internet called a ‘flow-control algorithm’. As is apparent from its name the flow-control algorithm keeps a check on the influx of traffic within the internet and prevents congestion, hence enabling a smooth and uninterrupted information flow. The groundbreaking revelations have the potential to expand our understating of both the complex neural activity and networks within the internet, which are headed for more complexity even as we speak. Another application of the study as suggested by the researchers has to do with the ability to better understand certain mysterious and complex learning disabilities that have crippled millions around the globe.
The algorithm in question is known as ‘AIMD‘ or ‘additive increase multiplicative decrease’ system. The primary function of AIMD is to regulate the flow within the web and keep a constant watch over congestion. It does that by first deploying packets of data through the stream of information and then carefully listens for receiver’s acknowledgement. If the response is prompt the system reads it as smooth and manageable traffic. However, if there is a lag in delay and response is taking a considerable amount of time, it informs the machine to scale down the functions, as low as by half in some cases. This decrease in processing is the multiplicative decrease side of the algorithm AIMD.
Call it a coincidence or one of the mysteries of the peculiar universe, that the internet and the human brain work on a very similar system of information handling. Inside our brains, once a neuron fires very closely after another one, it not only strengthens their synaptic connections, it also makes it possible for the second neuron to trigger if the first one is triggered in the future. This is very similar to what happens on the internet as ‘additive increase’. Multiplicative decrease takes place when the second neuron fires before the first one is triggered. This process significantly weakens the synaptic connection, hence decreasing the possibility of the first one triggering the second neuron in the future; very similar to what the internet does to slow down when information congestion takes place.
Saket Navlakha is an assistant professor of Salk University, who co-authored the new study revealing the perplexing similarity between web networks and neural activity of the brain; the study was published online in Neural Computation. Navlakha said that the revelation of the uncanny resemblance between a human-engineered system and evolved biological mechanism to solve an information flow problem ‘is really interesting’. The formation of the internet took place on the principle that allows it to successfully route the information fairly with great efficiency. This feature of the internet over the years has become the standard to judge its efficiency and reliability; since there was never any need of a centrally distributed system to overlook the entire internet; or you can say the internet never had a central authority. Those who founded the internet spent a colossal amount of time, considering and experimenting with a number of possibilities to make the information flow as efficient as possible.
Let us know what you think of the new study in the comments section.