G oogle Photos recognizes uploaded images automatically objects such as cats, mountains and beaches; Skype transcribed the spoken not only, but translates it in real time in different languages; a chess program called Giraffe brings himself within 72 hours the game itself in and then holds with professionals with.
These are just three examples from a long list of relatively newly released services and applications that all have in common one thing: They are based on artificial neural networks. These networks are currently experiencing its breakthrough and fueling a boom in artificial intelligence. No week passes without new breakthrough products based on “AI” – which may be always meant in detail
Confused terminology
AI, AI, Deep Learning, neural networks, Cognitive computing et cetera … Who deals with artificial intelligence will quickly supplied a fix with keywords. The discipline has now even spawned many techniques and designations and the industry does the rest to zoom with catchy slogans confusion. So here is a little word customer is initially announced.
Artificial Intelligence, AI for short, or the English equivalent Artificial Intelligence (AI) can be succinctly best be described with the words of the computer scientist John McCarthy, who coined the phrase: “the science and the development of intelligent machines”. There are many discussions aloof, exactly what that means – for example, what intelligence really is and whether the German word intelligence corresponds to the English Intelligence. But to get an idea of what is involved, should be quite sufficient McCarthy’s description.
How is he doing it? Microsoft reads faces feelings like anger, contempt, disgust, happiness or surprise.
AI includes the so-called machine learning; is more common for the English term machine learning. Here come algorithms used, which should learn during use on the basis of the data processed or that have undergone a training phase before use. There are several learning algorithms. “Naive Bayes” comes about in spam filters for use, “collaborative filtering” (to German collaborative filtering) at Amazon’s function “Customers who bought this item also bought …” [1]
the current superstars among the learning algorithms but the (artificial) neural networks, more specifically, the deep neural networks (DNN), so deep neural networks. Neural networks are relatively simple replicas of the human brain. Here, nerve cells, neurons, one behind the other in several layers. From Deep Neural Networks is when there are more than two additional layers between the input and output layer. The article on page 130 introduces the functioning of neural networks and the various topologies in detail. …
No comments:
Post a Comment