Google and Facebook coming up with a futuristic image learning algorithm.
Google Inc. the search engine giant and Facebook Inc. the social media giant have collaborated to work on a system that will be designed to function as an artificial neural network. It is extremely interesting to learn that both the companies have not only become successful in reaching their objective that will allow them to identify images but will also assist them in automatically generating pictures through what the machine learns on its own.
Currently, the machine has the capability to recognize buildings, faces, cars and other objects of similar nature. Moreover, it also has the ability to create such images on its own. The machine follows a set pattern which makes it able to comprehend to the trends and patterns. This bears striking resemblance with the human cognition that takes use of cognitive reasoning to function.
It is said that the first algorithm is assigned to make the picture on an arbitrary vector whereas the second is responsible to match it with real looking objects. Moreover, it can also nullify bogus images so that a more definite result can be obtained.
Both the companies recently tested the images where they asked people to choose the real image from the bogus images. Ironically, 40 percent of the population was actually fooled by the fake images. According to a researcher working on the artificial intelligence platform of Facebook believes that the technology has the potential to decipher the abstract form of images. This can be done through the white noise pictures found on TV sets. Apart from that, it has the ability to locate images that are governed from someone’s camera that will be under the influence and structure of the imagery pattern.
Apart from this, Google products and services are also working on another project where an image taken by any user has the ability to detect any type of birds present in the sky. Although the results greatly depend on the quality and resolution of the image, howver, a trial and error mechanism needs to be used to come up with more accurate results.
Both the companies now wish to make the mechanism better so that it can develop the imagery along with adding more test trials so that it can assist further in machine learning.
Facebook and Google have taken a great step that is not only different but will also change the fate of augmented reality. It is extremely enchanting to know that whether other firms will actually be interested to work with this technology that has an image learning algorithm to facilitate virtual reality in the times to come.
No comments:
Post a Comment