The United States military has a program to create an cat level artificial intelligence. This project is called systems of neuromorphic adaptive plastic scalable electronics (SyNAPSE). The term neuromorphic means the creation of digial neurons on a microchip that is able to mimic the functioning of real neurons. So its basically a chip designed to replicate the functioning of a real brain. The brain of a cat has approximately 10^8 neurons and 10^12 synapses compared to a human brain which has 10^11 neurons and 10^15 synapses. So this chip will likely have to mimic a fairly large proportion of those neurons to function properly. Basically they want to emulate a cat’s cortex or the outer layer of the brain which is important for cognitive processes.
The SyNAPSE project has a total of 5 different phases that need to be completed. In the first phase (0), they will have to demonstrate the underlying architecture of the synapses and neurons on the chip. They will also need to demonstrate that this technology can actually be scaled up to higher densities and also that it has a relatively low power consumption.
In Phase 2, they are going to start to run benchmark tests on the technology. They will test the chip out in a sensory enviroment like a virtual reality training ground. The ultimate goal of the project is to create a chip with a low power consumption and a compact design. They want the chip to have a neuroscience inspired architecture and an analog circuit design. When the chip is completed it will be able to address a variety of different cognitive abilities such as planning, perception, decision making and motor control. Scientists want this chip to be able to seek out resources and survive in a manner similar to a real cat.
In Phase 3 of this project researchers will have created a full-scale chip. The researchers hope that this artificial brain will be able to have several different attributes such as touch, vision, symbol manipulation and also survivability in a virtual environment. The other phases in the SyNAPSE program will consist of researchers further testing the chip.
It’s still too early to tell whether this project will be successful or not. Can you drastically simplify a brain on a chip and get it to function in a similar manner to a real brain? That is still not clear. However, this project will certainly give researchers a keener insight into how a brain works.