text

FOSSASIA - THE PERFECT TECHNOLOGY BLOG... We are Fossasia Technologies, we are trying to provide you all the latest information about Information Technology, Best Search Engine Optimisation (SEO) Technique. We are doing research on the daily basis to provide you all the best technique for SEO so that you can use them on your projects to get high ranks on the SERP(Search Engine Result Page). These techniques are the best techniques through which you can get organic traffic from all the search engines on your website. We are Fossasia Technologies providing latest news about IT Field.

Saturday, 20 June 2015

Artificial Intelligence is smarter than us now


I'm sure at some point your life you've taken an IQ test, whether it was for school, work, or just for fun. Some people put a lot of stock in their IQ, although it's good to remember it only measures one kind of intelligence.

Still, the IQ test does have some uses, and a new one is seeing how artificial intelligence matches up with human intelligence. Traditionally, AI didn't do so well, but that's just changed.

There are three things the typical IQ test measures: logic, math and verbal reasoning. Computers are good at the first two, but they struggle with verbal reasoning.

That's because when you're dealing with words, there's a lot of nuance. The same word can have multiple meanings depending on the context, and that might change the entire problem. So scientists came at the problem from a new direction.

As the Observer reports, "The researchers found it necessary to go beyond existing technologies to automatically solve verbal comprehension questions, so they created a framework consisting of three components."

The first element is a classifier that recognizes the specific type of a verbal question (Is it an analogy, classification, synonym or antonym problem?). The second prong involves leveraging a word embedding method that considers the multi-sense nature of words and the relational knowledge among words contained in dictionaries. Lastly, for each specific type of question, they developed a simple yet effective solver based on the obtained distributed word representations and relation representations. The overall result is a technique for recognizing the different meanings a word can have.


That explanation is clear as mud, but the result is that it worked. The researchers pitted their machine against 200 people from Amazon's Mechanical Turk program.

The AI's IQ was higher than people with a bachelor's degree and lower than people with a master's degree. In other words, it did better than the average human.

Is this another sign that machines are going to take over the world? Let me know your thoughts on this development in the comments.
Source > komando

No comments:

Post a Comment