RSS
 

The Machine Learning

30 Sep

Trend Even from the 21st century artificial intelligence was in a great crisis, could do a lot of logical reasoning, treat and rationalize equations, make inferences, but the volume of “human” data was small, the emergence of the Web and the adhesion of billions. of people enabled a new scenario: Big Data, a large amount of human data.

This large and complex data can be processed with large-scale accurate results, so that it can be processed by machine so that the machine “learns” trends, this (not by) machine learning is called “Machine Learning”.

With this it is possible to identify opportunities and avoid mistakes such as insisting on “logical” discourses but out of trend, it is not a question of going into fashion, but rather identifying it to correct it or propose other alternatives, and that is the opportunities Real.

Big data identifies data patterns, creates and analyzes the connections between them, and makes the execution of a given also smarter, whether or not it can rely on human supervision. In these two types, supervised and unsupervised, machine “learning” will always be essential, in supervised there is human interaction controlling the input and output of data and thus directly interfering with machine training, while in unsupervised algorithms use deep learning to handle and solve complex tasks without human intervention.

An important algorithm for semantic searches was made by a GitHut team: Hamel Husain. Ho-Hsiang Wu and Tiferet Gazit, and two Microsoft researchers: Miltiadis Allamanis and Marc Brockschmidt, who assess the state of evolution of semantic search algorithms and even make the code available through the GitHub site.

https://arxiv.org/pdf/1909.09436v1.pdf

 

Comentários estão fechados.