Rewrite the Technological Landscape

Chapter 819 Wind Valley CP?

The PA Waterloo mentioned by Meng Qian refers to the personal artificial intelligence that was not recognized by the market after it went online at the end of last year, but was scolded bloody.

The actual application of personal artificial intelligence is a software that can be connected with designated applications. Of course, the computer terminal and mobile terminal are interoperable, and the scope of designated applications is very wide. This is the advantage that Dafeng Group's huge ecosystem can provide.

After using the software to connect the cloud and the Internet, you can purchase PA services at different prices. The annual fee ranges from 80 yuan to 800 yuan. Each level can achieve different functions. For example, some are basic life data management, and some can assist work , according to the description, the highest gear can become a personal assistant with certain initiative.

However, after purchasing the service, many users have not felt the convenience brought by personal artificial intelligence for a long time, and then began to mock Dafeng Group.

Maybe it’s because I finally found an opportunity to ridicule Dafeng Group. Although the company has made a lot of related explanations, people don’t seem to listen at all, especially in foreign countries. Eighteen floors.

In addition, the current Internet has turned into a carnival to watch the winners fail. The Dafeng Group has been successful for too long, and it is hard to catch an opportunity to watch the Dafeng Group fail. The Internet is full of people who want to follow suit .

So much so that the company had to consider asking Meng Qian to come forward, hoping to use Meng Qian's influence to restore the situation. Of course, the premise is that the technology itself is not really a big problem, it is a problem of public perception.

Therefore, Meng Qian approached the media to arrange a live interview on March 17 to respond positively to the personal artificial intelligence issue.

Coincidentally, on March 17, the Internet was still hyping about AlphaGo, and it was still hyping about Demis Hassabis provoking Meng Qian.

So Meng Qian simply sent a rumor on this day: AlphaGo is just AlphaGo, nothing more.

Tens of thousands of comments appeared in just one sentence, and most of them were asking what Meng Qian meant by that. On the night of the 17th, the live interview started.

"Mr. Meng, the news you posted on the Internet today has attracted more than 50,000 comments, but everyone is obviously curious about what you mean. Can you give a positive answer?" The host first asked about today's affairs .

Meng Qian responded very calmly, "AlphaGo is just AlphaGo. In fact, that is what it means. You can search for the definition of AlphaGo first. AlphaGo is a set of deep learning engines designed for Go optimization. It uses neural networks and MCTS. , combined with Google's huge cloud computing resources, and equipped with a large number of GPUs to provide general computing capabilities.

To put it bluntly, this is a deep learning algorithm about Go, because with sufficient resources and hardware, the algorithm can achieve a good self-learning ability.

I remember that Demis Hassabis also said yesterday that AlphaGo is a combination of advanced search tree and deep neural network. The neural network transmits the description of the chessboard through 12 processing layers. These processing layers contain millions of similar connection point to the nerve. One of the neural networks chooses the next move, and the other predicts the winner of the game.

So, this is actually a successful application of deep learning. I don’t deny that this is a very successful application, but if you say that because of AlphaGo, it is a new milestone in artificial intelligence, or even mentioning that artificial intelligence replaces humans, it would be ridiculous. . "

"According to Demis Hassabis, AlphaGo has made artificial intelligence a qualitative change from a quantitative change."

"Hey, the purpose of deep algorithms from the very beginning is to make artificial intelligence change from quantitative to qualitative. If you think that AlphaGo has finally achieved this, it can only show that their understanding of deep learning is still relatively shallow.

What is quantitative change to achieve qualitative change?

A person memorizes 10,000 poems and then exports them into poems. Is this the quantitative change we understand and the qualitative change? "

"certainly."

"Then if a person recites a hundred poems and then exports them into poems, is this a quantitative change into a qualitative change?"

"That should be the case."

Meng Qian said affirmatively, "Of course it is, why not? It's just that the latter is smarter than the former, but if the latter has never read poetry, he can't write poetry, so the process of quantitative change to qualitative change of human beings requires two Things, one is smart, one is learning skills, and the other is learning content.

When it comes to artificial intelligence, it means algorithm computing power and data. I believe everyone can understand it.

So does AlphaGo make sense? Of course, AlphaGo has confirmed what kind of level a computer can achieve in a specific field under the algorithmic computing power big data provided by Google. We can understand that AlphaGo is a relatively smart machine, but it is just a machine after all. One of the many cases of deep learning from quantitative change to qualitative change, because it does not break away from the limitations of specific applications, that is, it is useless outside the field of Go.

If you want to say what is a milestone breakthrough, I think it is a matter at this stage, so that computers can achieve understanding in the face of more data in different fields, and finally break through the limitations of specific applications. "

The host changed the topic very sensiblely, "Mr. Meng is talking about personal artificial intelligence, right?"

Meng Qian also took it very naturally, "That's right.

I can clearly say that personal artificial intelligence is our key development direction in the future, because I was more modest and did not say this before, but in fact, it is a milestone in the true sense of artificial intelligence.

You can think about it, if AlphaGo's achievements in the field of Go are worthy of such publicity recently, then our personal artificial intelligence is processing multi-line data of users at the same time, and the amount of data in the early stage is very small. Artificial intelligence can't be boasted on the hot search every day?

The word-of-mouth of personal artificial intelligence has not been very good in the past four months. I have also seen many friends say that this may be the first overturned work of Dafeng Group, but what I want to tell you today is the technical support behind personal artificial intelligence. How huge it is you cannot imagine.

I believe some people will say that no matter how great the technical support behind you is, it has nothing to do with us. If your product is not good, it is not good. But the problem is that one of the characteristics of deep learning is the need to feed data.

The data needed for personal artificial intelligence is our user data, so this will inevitably take time.

We have said before that one of the great pleasures of personal artificial intelligence is that you can cultivate it as a companion, and go through a process from strange to familiar, it will gradually understand you, slowly understand you, and finally slowly To become a robot companion that can bring you help.

This is the value and performance of personal artificial intelligence. Personal artificial intelligence is here to serve you. Are you not afraid that you will know your situation well from the very beginning? "

After Meng Qian finished his explanation, he began to show some relatively successful cases. In the case, someone’s personal artificial intelligence has been able to help him complete the relatively repetitive daily recording work, and personal artificial intelligence has already assisted the user in making a schedule for the next week. .

This is also the reason why this interview has to be delayed until March. There must be some relatively successful user cases to be convincing, and if you want to have such user cases, you will definitely need time for users to feed the data, otherwise just talk about how good it is , even if Meng Qian personally launched the game, users would not buy it too much.

It’s different now. With real success stories, clear theoretical explanations, and the market enthusiasm promoted by Google, user reactions have begun to change.

Although I have been talking about it before the theoretical explanation, but before everyone just didn’t listen, and scolded Dafeng Group for wanting to make money, they just said it. Now when I think that AlphaGo can be used to brag like this, it seems that personal artificial intelligence is even more powerful.

On Google's side, Demis Hassabis frowned watching Meng Qian's video, and quickly went to Page, "This Meng Qian is a bit of a sneaky change of concept."

At this moment, Paige's expression was indescribably weird, "Have you noticed the comments on the Internet?"

"I didn't read the comments on the Internet, what else is there to read in the current online comments."

"But the direction of public opinion this time is a bit inexplicable."

Faced with Page's strange expression and strange expression, Demis Hassabis was still curious to understand the public opinion on the Internet, and then he also looked a little confused.

"AlphaGo is a marketing campaign that Dafeng Group paid Google for, right?"

"It's been a joint marketing campaign for a long time. AlphaGo is so popular just to show how awesome PA is."

"I don't know if the PA is awesome or not, but it can actually let Google help with marketing. The Dafeng Group is really awesome in doing this."

"It's over, it's over, what should I do if I suddenly hit the CP of Fenggu Valley?"

Demis Hassabis came back to his senses for a long time, "Is this because the netizens are too brainy or Meng Qian is just doing a slapstick operation again!"

Chapter 828/1045
79.23%
Rewrite the Technological LandscapeCh.828/1045 [79.23%]