Saturday, November 18, 2017
Text size

The Artificial Intelligence and Cybernetics

on Wednesday, 04 March 2015. Posted in Artificial Intelligence

Science fiction has, for a long time, looked to a future in which robots are astute and cyborgs. However, science has sped up practices to make the thoughts of the science fiction to come true in the real world. We consider a few distinctive tests in connecting science and innovation together in a robotic manner, basically eventually consolidating people and machines in a moderately changeless merger. Following are some example: -

Robot with a brain:

Neurons cultured under research laboratory conditions on an array of non-intrusive electrodes give an alluring option which to understand another manifestation of robot controller. An experimental control stage, basically a robot body, can move around in a characterized zone absolutely under the control of such a system/brain and the impacts of the brain, controlling the body, can be seen.

The brain was coupled with its physical robot body. Sensory information was fed from the robot to the brain completing the robot-brain loop. Thus the processing of the signal were in two forms:

  1. Brain to robot, in which live neuronal movement is utilized as the choice making instrument for robot control.
  2. Robot to the brain, which includes an information mapping methodology, from robot sensor to stimulate brain/neuron.

Human brain-computer Interface:

Numerous human brain–computer interfaces are utilized for restorative purposes, so as to conquer a therapeutic/neurological issue - an illustration being the profound brain incitement anodes utilized to defeat the impacts of Parkinson's Disease. In any case, even here it is conceivable to consider utilizing such innovation in option approaches to give human capacities not ordinarily possessed by the people.

Sometimes, the individuals who have endured a removal or have endured a spinal damage because of a mishap may have the capacity to recapture control of gadgets through their neural signals. In the interim, stroke patients can be given constrained control of their surroundings, as surely can the individuals who have the motor neuron illness. If such implant are done with the body that does not require any repair as in the above cases, surely these implantations are going to improve the abilities of the normal humans.

In conclusion, we can say that the combination of artificial intelligence and cybernetics can help humans, to improve their abilities from normal to extra-ordinary. Also, the robot can be created, having the ability to think like a human and to do work as humans do.

Primate Inspires Computer To See Image

on Tuesday, 03 March 2015. Posted in Artificial Intelligence

Up to this point, no computer model has held the capacity to match the primate brain at recognizing objects amid a short look. Since year, researchers have tried to build a system that mimics the brain's ability to see images, recognize speech and understand it.  However, the neuroscientist have come up with the latest study on deep neural network that can mimic the primate brain. The model shows the best understanding of the researchers on how the brain recognizes the object.

 Working of the model

Scientists got inspired from the representation of visual information hierarchy in the brain. The image information passes from the retina towards the primary visual cortex and then to the  inferior temporal cortex. The image is processed at each stage until it is recognized by the brain. The neural network designers have created numerous layers for computing in the model, to mimic the hierarchy. The layer performs the linear dot product mathematical operation. The object location, which is not needed, is cast aside. The representation of the object becomes complex at each layer.

For the study, researchers initially measured the object recognition capacity of the brain. The scientists have embedded series of electrodes in the inferior temporal cortex and a part that feeds visual in the inferior temporal cortex. This helps in getting the number of neurons reacting to a particular image by the animal. The  scientist the compared the result of the experiment with the representation created by the deep neural networks. The representations were in the form of a matrix generated by the total computational units in the system. Each image used, created a different series of the numbers and the model accuracy is determined by whether it can gather objects into into comparative groups.

Result

The study has taken a leap in the computer processing capacity. Researchers are using small chips intended for elite in handling the gigantic measure of visual content required in video games. The large dataset is available to feed the system' algorithm. Before the study was conducted, the neural network was unable to identify the images. Now, due to the availability of numerous images, the system can mold calculations accordingly to become more accurate in recognition of the object. 

<<  1 2 3 [4