Conducting research that Google for some time in the field of artificial intelligence is the latest since the acquisition of Deep Mind no longer a secret: in 2015 the company released the results of a research project in which the artificial intelligence should learn old Atari game classic autonomously. According to Google DeepMind succeeded Google has released AI Software 2016 both to learn the rules of the game and to develop successful tactics autonomously. Deep Mind sets to simulate the ability of an artificial memory in his studies on a combination of neural networks and short-term memory.
A further step in the development of artificial intelligence has done with Deep Dream Google: As visualization software Deep Dream was originally developed by the procedures and principles to better understand neural networks. To this end, Deep Dream was lined with millions of images and observed how the levels of the neural network to identify objects or people and animals in general, interpret information. By the time Deep Dream got the recognition of shapes and the visualization of the learning process has produced fascinating images. The flood of images of the artificial neural network called Google itself as Inceptionism.
A step closer to artificial intelligence
Together with Geoffrey Hinton expert on artificial neural networks the search giant for around two years has been working on a new algorithm to recognize logical relationships in language and so cause natural conversations and even flirting can should.
People will talk not only to exchange information with computers but for fun – and because the devices will then be in a position they can reflect empathy through language are the friend or partner:
In future it will be Possible to Predict thought vectors from the sequence of previous thought vectors and this will capture natural human reasoning. With Sufficient effort it may even be Possible to train a search system to ignore nearly all of the contents of its thoughts and to make predictions based on Purely Those features of the thoughts did capture the logical form of the sentences used to express them.
Thought Vectors – on the way to autonomous AI
Thought Vectors are already being used in a simple version in Google Translate but lead it still often meaningless translations. In the future, that will change: Thought Vectors give each word a number set (or vector) that describes its position in a theoretical „Meaning Space“ (or cloud). A set can be regarded as the path between each word in turn is divided into its own set of numbers. The „Thought“ serves as a bridge between two languages because it is transmitted, for example, in the French version of the Meaning Space and then translated into a new path between the words.
First, however, has to be found out which numbers to which word in the language include – and here is where Deep Learning. Initially, the positions of the words in the cloud are still arranged without system and the first translations probably produce nonsense. Through a feedback loop, an error signal is output that the placement of a word in a sentence as long as changes to the correct position comes out.
Google is ignoring everything the Terminator franchise taught us and is releasing „TensorFlow“, open-source software that helps computers learn quicker than ever before. The software is a branch of artificial intelligence called „machine learning,“ tech that has already found a home in Google Search, Google Photos and Gmail. Tech-giants, like Google, Facebook and Amazon are all working with machine learning to better the services that they offer like smart search, ad targeting and product recommendations. Machine learning is now shifting into a complex realm where researchers are creating computer models that can see and even understand what it’s looking at.
TensorFlow is modelled after how the brain recognizes things and can be used for both research and application purposes. Google noted that TensorFlow is a great start to advancing the world’s technology. By making the program open-source, more developers adopt it to convert their research into a workable application without having to rewrite code. The future may not be now, but it seems just around the computer-identified bend.