Inventors:
Yang, Jiang-ming (Beijing, CN) Cai, Rui (Beijing, CN) Wang, Xin-jing (Beijing, CN) Zhang, Lei (Beijing, CN) under the auspices of Microsoft Corporation were awarded patent rights for an artificial intelligence process.

The Basis of the Patent
Here is a patent you can read about on our Artificial Intelligence RSS feed. Date May 22 titled “Modeling semantic and structure of threaded discussions”. How will this software affect our social lifestyle?
This is intelligence computer software for the management and automation of DATA logic streams. It is designesd to increase logic streams to develop the art of discussion closer to the level of the fact of the word and begin interpreting automate dicisions on the possible meaning word, given the context in which itthe word is found in the discussion threads.
The patent was awarded on May 22, 2012 and was applied for, almost 3 years ago. It seems they are a bit backed up over there in the US Office of Patents and Trademarks. This patent will help bloggers, forum and discussion group moderators and is based on models consisting of semantic terms and structural terms. It will help in organizing posts on forums, finding the experts on a subject for good information sorting and identify junk mail and spin content that tries to appear on blogs and forums these days.
The patent comes under the heading of artificial intelligence because this modeling system makes decisions and constructs organized ideas in the form dialogue output based on dialogue input.
If you are looking for the origins of Robby, Terminator the Jetsons or Futurama or even HAL from 2001 type robots, then look no further. It is verbal logic systems like the one patented this week that will be the cyborgs of tomorrow.
Here is a discussion of experts model example used in the Patent application.
This is the kind of discussion that a robot would understand as easily as a programmer discussing topics online.
Very tidy piece of organization. And presumably the junk posts have been dispensed with as well. This next figure contained in the patent give a hint of the formula used for that function.
Assigning numeric probability values to semantic and structural terms to come up with the likelihood of a bogus comment.
By this method it would in principle be possible to construct laws and ordinances. By creating a discussion framework to solve a certain problem. Then sorting through the comments, discovering who were the experts in the discussion and who were just letting off steam. You could have a political module, a logistics module an engineering module, behavior statistics module and run the comments through them and set up the elements in the law that would give it highest probability to pass legislation.
This patent has been pending for almost 3 years now and so has probably been in use since then. That means it may already be outdated. Who knows what they are actually implementing now in the artificial intelligence world.
Keep Posted.