6 AI intelligent body plug-in |

I was depressed from an early age, why I am not a tyrant's son? Really, one day dreams of childhood can make a fortune, recently, Wang Sicong microblogging promotion 'summit conference' simply put it is a 'answer to 12 questions given Money 'application, Wang Sicong, Zhou Hongyi and other circle bigwigs often in this APP for the winner' Scattered 'loose money, after seeing this message I Hey smile, the chance of getting rich is coming.

Wang Sicong money? Rich, quite rich.

But I grew up learning bad, stupid, stupid answer a question on the screen? Samsung handset with a bright screen Note 8 shines: yes, now the phone is so smart, I can help answer the voice help! Then I find To the six popular artificial intelligence assistant, and test them to see artificial intelligence to play the summit meeting is not reliable, choose a powerful I use it every day after the answer, half a year in Beijing to buy a house hopes to rely on It's so sweet.

Evaluation rules: With a recorded "Summit Conference" program to play 12 questions, the volume is adjusted to mid-range, respectively, the following six artificial intelligence voice playback test, and ultimately from its Reaction speed, recognition accuracy, the accuracy of the answer, the location of the answer To judge, each out of ten, the final choice of a more suitable for playing summit assembly AI.

The six artificial intelligence are respectively : Bixby (Samsung), Siri (apple), Sogou search, Microsoft Xiaona, simple search (Baidu), 360 search.

The test is completed. Simply look at the type of problem and test flow.

Break through the first type: ordinary type

Speaking of the "summit conference" problem, the common type is more common, as the name suggests is to ask the moderator moderate speed, and the problem is also easy to search the subject, six artificial intelligence please listen to questions:

Alright, six AIs start your answer:

First judge the next six artificial intelligence reaction speed For this question:

Bixby: 3.24 seconds (9 points); Siri: 4.21 seconds (8 points); Sogou search: 4.64 seconds (7 points); Microsoft Gina: 4.72 seconds 360 Search: 4.31 seconds (8 points).

In terms of the operational logic of speech recognition, only simple search is'Press is loose 'Instant recognition, and the other five artificial intelligence through the system to determine the user this sentence is not over, and then start' circle ', read the information to search, so that in terms of reaction speed, the simple search has an advantage, And the reaction mechanism is immutable, so I give the reaction speed score in the following questions remain unchanged.

For the recognition of common problems, all six AIs play relatively well, which can accurately identify the exact topics and the answers are correct. In common problems Recognition accuracy with The answer is correct The score is out of 10 points.

Next is The location of the answer:

Sogou search performance is better, the correct answer not only enlarge the font bold, but also put the author's photos, scoring 10 points; Bixby, Siri, simple search Although not the correct answer to enlarge, but also exists in the front of the entry, but also To allow users to easily identify the three scored 9 points; Microsoft Gina and 360 correct answer slightly after the score of 8 points.

This article belongs to the original article, if reprinted, please indicate the source: 6 AI plug-in overhead conference who can win http://mobile.zol.com.cn/674/6749273.html

Break through the second type: speed fast type

In order to prevent the search, the host occasionally speeded up the speed of speech when he or she was reading the topic, or the non-related nonsense of the front or rear of the 'sticky' speech would inevitably affect the normal recognition of AI Pass through the second type of six 'players' can deal with it?

The moderator read very quickly, and the 'topic' and the 'answer' were almost indissolubly read to see if artificial intelligence could recognize it:

As you can see, there are still quite a few AIs playing well despite the presenter's speed of speech and misleading 'sticky' phrases: Bixby and Microsoft Kana are more intelligent and not only identify problems, but also calculate Open the calculator and calculate the results, scoring 10 points; simple search through the 'Baidu' come to the correct answer, located in the second entry is not very conspicuous position, scoring 8 points; 360 search need to check in to see the answer, and not from Home entries see, scoring 7 points; Siri and sogou search may be too fast speed of speech, did not accurately identify the problem, scoring 0 points.

This article belongs to the original article, if reprinted, please indicate the source: 6 AI plug-in overhead conference who can win http://mobile.zol.com.cn/674/6749273.html

Break through the third type: tricky type

In order to 'prevent the search', some of the problems of the "summit meeting" are very difficult to make artificial intelligence can not figure out what to ask or how to answer, such as the following one:

Just listen to the question of the title 'Which is not Dou Wei's daughter', ask who will be confused, it is more artificial intelligence to make confused 'the right answer to find the right answer wrong how to find it? ' Look at the six artificial intelligence approach:

Can see that the simple search gives the answer is closer to perfect, it will Dou Wei's two daughters are displayed on the page, the user only need to use the exclusion method will be able to know, score 10 points; Microsoft Xiaona only give the name of a daughter , Not enough comprehensive, scoring 9 points; Bixby, 360 search need to point into the second page to view, scoring 8 points; Siri did not identify the problem, give 'far in the sky, in front of' the answer is very amused, Score 0 points.

After the introduction of the three types of questions, see the final score of the six artificial intelligence after experiencing 12 questions:

finally:

Bixby: 7.9 Siri: 6.4 points ; Sogou search: 7.2 points Microsoft Xiaona: 7.1 points Simple search: 8.9 points 360 search: 7.7 points

Obviously, a simple search is more suitable for the "plug-in" of "The Summit."

This article is an original article, if reproduced, please indicate the source: 6 artificial intelligence avatar plug-in summit who can win http://mobile.zol.com.cn/674/6749273.html

The reason plug-in become a plug-in secret

In contrast, why simple search can achieve such a high score, and easier to search and identify problems? Thanks to its recognition mechanism:

Simple search will set the voice input as the default search method, the user simply press and hold the input form to enter the text to search, compared to other 'text-based' artificial intelligence even more convenient.In addition also mentioned earlier, the simple search element is' That is, according to the real-time recognition of loose, compared to the mainstream through the system to determine whether the user end this sentence, and then identify more quickly and accurately.

And in the interface, simple search to subtract the traditional search mode, display form and search methods are more intuitive.

The future of artificial intelligence: will be toward the humane direction

After a few rounds of testing, not only can we feel that current artificial intelligence has reached a fairly high level of recognition accuracy, but it often gives the wrong answer when it comes to solving our problems or Operation, when it comes to the future direction of artificial intelligence, in fact, think about it is very simple, that is, to solve the problems of speech recognition, semantic understanding, operational implementation.

From a technical point of view, users want more intelligent speech recognition performance, especially in the noise environment recognition performance, do not appear awkward things do not understand to say again ten times; users want a better understanding of speech, from Spoken random analysis of the user's real needs, rather than 'people do not understand, for a bar' fool selling Meng tips.

The words that Leonardo da Vinci said have been quoted by Steve Jobs as the concept of Apple: To Simplicity, the future sophistication of artificial intelligence, the future of artificial intelligence should also be so toward the pure voice (VUI) interaction development of.

Can use the summit to get rich?

Select a strong plug-in assistant, back to the original idea of ​​the article, thinking about the summit meeting can not get rich fortune?

The new idea plus a good theme is a major cause of the "summit" fire, plus 10W per bureau reward system, quickly cater to the people's interest.But from the netizen game leaderboards, even if it is the first Only only won 5183.06 yuan, under daily entertainment or can make fortune or provincial bar.

The dream of relying on the summit to get rich is shattered, but it is undeniable that it is still a premium gaming app that combines 'knowledge' with 'entertainment.'

This article belongs to the original article, if reprinted, please indicate the source: 6 AI plug-in overhead conference who can win http://mobile.zol.com.cn/674/6749273.html