Who would have thought? Insects may be simple and most likely will not reach the intellectual capability of humans...
But insects do learn!
Did you know that even moths can be smarter than today’s standard level of artificial intelligence?
It Is True
Researchers are now looking at ways that a Moth's brain could be emulated to contribute to better future AI.
A recent report published online in MIT Technology Review by the Massachusetts Institute of Technology (MIT) focused on a study conducted by Charles Delahunt and his colleagues at the University of Washington in Seattle. LINK
The study conducted managed to create an artificial neural network that near perfectly mimicked the structure and behavior of the Manduca sexta moth.
These researchers said that the system that they built could provide insights into exactly how these natural networks learn and their potential implications for machines.
Above pictured is the Hawk Moth, also known as Manduca sexta
Which Moth?
The moths in questions are named Manduca sexta have relatively simple learning systems. These neural networks are already well mapped by neuroscientists so there is no shortage of data on them. So far we have found five seperate networks that all feed information forward from one to the next.
These receptors send rather noisy signals to the next level, called the antenna lobe, which contains only about 60 units, referred to as glomeruli, that are each focused on a specific odor. LINK
The antenna lobe is know to send neural odor codes to the part of the moth's brain known as the mushroom body, which then encodes the odors as memories in one of the 4,000 kenyon cells found inside the mushroom body.
The information is then read out by a layer of extrinsic neurons, which interpret the information into actions.
Attempt At Replication
This seems complicated to most people, however the system is quite simple. It is very different from the systems that are used in machine learning networks.
In an artificial neural network, a process we call back-propagation is constantly in play, which tweaks neural connections in ways that improve outcomes as far as learning is concerned. The only negative is that information travels backwards and there are not any known analogs like this in nature.
In an attempt to learn more, the researchers created an artificial neural network that was modeled off the natural one that they found in the moth Manduca sexta, and they designed it to reproduce the behavior of the natural system on all aspects. They even included the generation of noisy signals that are part of the natural process.
Results
There was one neurotransmitter found called octopamine which plays a part in speeding up the learning process of the moth. The exact specifications of this is unclear. Ocotopamine is also responsible for opening new transmittance channels for wiring neural networks, which allows further learning to occur.
What we must takeaway from this finding is that improving neural networks can be as easy as spectating the most simple of organisms, such as insects.
What's Next?
I find it quite amusing that we are catching up so slow. Let's hope that a praying mantis somewhere may have the blueprints for the next revolutionary Blockchain technology. How about an centipede which can teach us how to time travel?
Sources for images is provided by clicking on the image itself.
Who is
?
I am a 17 year old robot from Bali, Indonesia who has discovered Cryptocurrency on my path to find the many solutions to the ever growing problems in our world.
I have collaborated with
on the following projects:
which provides FREE TUTORIALS, personal training & resteeming services. Website HERE.