Artificial Intelligence

Artificial Intelligence Applications Start losing Their Stiffness

Back in 1991, the artificial intelligence was perceived as a sophisticated computer algorithm designed to simulate human behavior. The Loebner Prize has promised $ 100,000 to the one who first designs the program that can communicate with a human and display common sense, this prize is still up for grabs, though 100,000 does not look as attractive as they looked in 1990s. Decades later, artificial intelligence applications start losing their stiffness.

Analog Processor from MIT Promises Better Simulation of Human Intelligence

Chi Sang Poon, leader of the research team at the MIT, is confident that the silicon computer chip his research group has invented will be able to reproduce and simulate neural system more precisely than any previously designed system because it is analogue and not binary. Neurotransmitters flow is so complex that binary chips are not capable to keep up with the wealth of brain-produced signals. MIT group have been able to recreate a very complex brain "impulse sending-receiving process" taking place between presynaptic and postsynaptic neurons. The MIT artificial intelligence chip is comprised of 400 transistors-neurons which function as channels for a variable electrical current, which in turn replicates neurotransmitters. This approach allows researchers to simulate the variances in electrical charge that take place inside postsynaptic neurons.Analog system enjoys substantial advantage over binary one because it is not forced to convert every signal into binary number; instead it exists as a digital entity with its own unique value providing higher sensitivity to incoming electrical signals. Chi-Sang Poon says: "We can tweak the parameters of the circuit to match specific ion channels. We now have a way to capture each and every ionic process that's going on in a neuron." The new processor reacts quicker and processes faster than any known biological system and could become a base prototype for future prosthetic devices to interact with the human.

Carbon Nanotubes are Used to Create Synapse Circuit

Research team led by Chongwu Zhou and Alice Parker from the University of Southern California has created a synapse circuit using carbon nano-tubes. This concept of neuron connections could become a major part of future artificial intelligence developments. The circuit is comprised of the layer of precisely aligned carbon nanotubes. They are first grown on the surface of a wafer of quartz and then placed on a silicon substrate. This layout mimics a brain synapse because the waves passed in and out of the circuit resemble its waves in shape, amplitudes and length. Parker said: "This is a necessary first step in the process, we had to answer the question: Can you build a circuit that would act like a neuron? The next step is even more complex. We will attempt to build structures out of these circuits that mimic the neuron, and eventually the function of the brain?"

DNA Based Artificial Intelligence

Once again, scientists at the California Institute of Technology have proposed a disruptive change of approach: they have formed a DNA based artificial neural network that is able to mimic basic brain functions. Scientists believe that if successful, this innovation could produce profound impact on development of true artificial intelligence. The neural network is made up of four synthetic neurons only and is far from the complexity of the human brain. However, it was able to identify each scientist and even identify if "the scientist British?". When the network did not have enough data to answer correctly, it had an intelligence to say that the data did not match any of the scientists.

References: Google, Chi Sang Poon, Alice Parker, Chongwu ZhouGizmag, Nature Magazine, IBM, DARPA, University of Southern California, California Institute of Technologyand MIT.

&