Go Legend Chapter 530

Chapter 530 A special haircut

In the 18th round of , Li Xiangping, who had just won the "Fujitsu Cup", missed this place. He was defeated by Li Shishi with a classic move.

Of course, chess fans all know that "one piece solves two signs" belongs to the kind of very cool hand tendons. Once the move is hit, there is basically no solution. Therefore, when Li Xiangping saw the move, it took two minutes to figure out that it was indeed unsolvable. He was very refreshed. The thrower conceded defeat.

After the game, the two players did not conduct a long replay. After all, the language was not available, and it was impossible for a full-time translator to be equipped for the replay. Therefore, after a few random gestures, Xiao Li walked away and stayed behind. Li Xiangping was thinking on the spot alone.

And when he was in a daze, more than ten minutes passed.

No one thought there was anything wrong with this, and no one went over to disturb him.

After all, this is Li Xiangping. After all, this is one of his only two defeats this year. It is his second defeat this year after losing to Gu Dali in March.

Since he suffered a defeat, even Li Xiangping, he must have adjusted his mood. Considering that it is rare for him to encounter such a situation now, so he spends a little longer to arrange his hair, which of course everyone understand.

It's just that no one knows that Li Xiangping's “haircut” this time is so special:

"Ah, little friend Xiangping, I only know until now that if we want to defeat the machine, it turns out that this person will be the biggest helper."

Lao Shi's words seemed inexplicable, but Li Xiangping understood it right away. Not only did he understand, but it was also a new understanding he had after playing today's game of chess:

"Hehe Ding'an brother, do you think so too, it seems that heroes see the same thing"

To explain Li Xiangping's new understanding, we must first briefly understand the basic context of this game of chess:

From the very beginning of this game, the two of them practiced a "dog move".

Well, a dog move that has almost become a "dog set", a dog move that is almost well known in today's professional chess world. And not long ago, Xiao Li also used this change, that is, in this year's "Toyota Cup" round of 16, he was eliminated by Kong Erjie.

However, Xiao Li is Xiao Li after all. He takes the changes he has lost without fail. "Not taking the usual path" seems to always be his pursuit. Today, he is no exception. When there were more than 40 moves in the overall game, he changed again. Move, let that "dog set" deviated from the normal track.

It's just that Xiao Li's move this time is different from last time.

The last time he changed his moves, not only did he not take advantage of Kong Erjie, but Li Xiangping even found a good move in the research room, and he found a good move in a short period of time, one that could be defeated immediately. That move was a good move.

Therefore, the change he made last time was objectively not very technical. It was very much like he was forced to "do not take the ordinary way" in order to "do not take the ordinary way".

But his move today is different, completely different!

This move not only has stronger concealment and confusion, but also has a much higher technical content, so high that even Li Xiangping was lost, and so high that he was defeated by others before he passed a hundred moves.

Li Xiangping is now reviewing this game again, he does not think he has lost wrongly.

In fact, when he first saw Xiao Li's changes, Li Xiangping also counted two changes, one way was slightly peaceful, and if he continued to evolve according to that way, he might still benefit a little.

And the other way of change, of course, is the road of no return in actual combat.

The reason why he chose this road of no return in actual combat is because he did not calculate the step of "one son solves two problems". That move was the 93rd move in the overall situation, and Li Xiangping's calculation just happened to stop at Global 92nd hand.

Further in-depth discussion: Li Xiangping's calculation path is normal to stay there.

Because this is the characteristic of the classic trick of "one son solves two problems": as long as you don't count this step, then in the previous move, you must think that you are in your favor, and the person who collapses immediately should be your opponent. �

And this is precisely Li Xiangping's game psychology at the time: Do you dare to change the blatant dog trick? Are you too unreasonable? Aren't you looking for death? So how should I punish you? By the way, it should be the correct solution to evolve like this. If I go like this, Xiao Li seems to collapse within a hundred hands.

It is precisely because of this kind of game psychology that Li Xiangping's calculation stops at 92 moves.

It is precisely because Li Xiangping has this kind of consciousness that he thinks that he has nothing to say if he loses today, and he will agree with his statement of cheating: opponents like Xiao Li are really rare, especially if he wants to challenge the dog. If so, then such an opponent is even more valuable! It is much rarer than an opponent like Da Li Gu Dali.

why would you say so? To make an inappropriate analogy, if you compare the dog tricks you are studying now to "Xuanmen Zhengzong", Da Li is also studying "Xuanmen Zhengzong", Gu Dali is also studying "Xuanmen Zhengzong", and other Many, many chess players are also studying "Xuanmen Zhengzong".

However, Xiao Li is different, he is still studying his "heteropath", and judging from today's game of chess, his "hellman" seems to be upgrading, at least he is much stronger than his previous life.

In his previous life, Xiao Li actually created a "zombie flow" under the framework of the human Go theory, but now, his evil kung fu is already based on "dog tricks" � ��

Li Xiangping doesn't know what name to choose, but he can confirm one thing: Xiao Li's current "hero side" is already one level stronger than the "zombie flow" in his previous life.

Then the question now comes: If Li Xiangping stands in the perspective of sharpening his chess skills, would it be more helpful to be the "Xuanmen authentic" of Da Li Guli? Or is it more helpful for Xiao Li, an "evil outsider" that has been upgraded?

The answer is obvious, everyone knows that Xiao Li's style is more helpful!

"Hey, Xiao Li is really Xiao Li, the unruly Xiao Li, the always maverick Xiao Li, there are opponents like him, Ding'an brother, do you think this is one of life's great joys? It's a pity that opponents like him are still too few."

"Hehe, this little Xiangping friend is relieved. I believe that there will be more and more such opponents in the near future."

"Oh, why did Brother Ding'an say that?"

"Little friend Xiangping, I remember that in a game a few months ago, you said something to your little brother: Those who learn from me live, and those who like me die, since you know this truth, are you still worried? Don't others understand?"

Li Xiangping had a black face at the time: "How does Brother Ding'an speak? Since I know what it means"

But although Li Xiangping has a black face, he thinks it is true. There are so many smart people in the world, especially in the professional Go world, all of them have high IQ.

So if you want to challenge yourself or even beat yourself, it is impossible for everyone to follow behind you. When they understand the dog trick to a certain extent, someone will definitely find a way to find a new way, and gradually become Xiao Li Such "hell outs".

It's almost inevitable!

Because I am not a real dog after all, I can't do a "dimension-reducing blow" to humans like dogs, and let humans kneel down and sing and conquer.

In this case, there must be many people who want to beat themselves, especially those chess players in the professional chess world's first army group. It's quiet and twisting.

"Hey, good, great thing, now it seems that the "magic door" of the Go world may really be able to flourish, but then again, if you are qualified to enter the "magic door" of the Go world, then there is really no one to take it easy, even From a certain point of view, the average aptitude of the "Magic Gate" in the Go world is actually far higher than the average qualification of the "Xuanmen Authentic". Now I just don't know which players will enter the Magic Gate in the next 10 years. ”

While Li Xiangping was thinking like this, Lao Shi suddenly had a question that made his heart beat faster.

"Xiangping little friend, I suddenly thought of a question just now, and I really want to ask your opinion."

"If you have something to say, say it, why do you have to be so serious."

"I was just thinking, if today were replaced by that machine, would they be able to avoid this step of "one child solves two problems"?"

"Huh?!! This"

Hearing Lao Shi's question, it was like a thunderclap, which instantly made Li Xiangping's heart beat faster. After he forced himself to calm down and think for a while, he asked back:

"What did Brother Ding'an think?"

Lao Shi said with a wry smile: "I just can't be sure, so I want to hear the opinion of little friend Xiangping."

At this time, Li Xiangping had completely calmed down. He smiled and said, "I think the dog is likely to be unable to avoid it."

"Why?"

"Without him, because of a method like "one son solves two signs", the previous evolution must be extremely fierce, every step is startling, and every move is inseparable from the back of the head, and this feature is precisely the weakness of the dog."

Lao Shi also thought about it a little: "What does Xiaoyou Xiangping mean, every move before this, the winning rate is shaking violently, there may be blind spots that machines can't find?"

Li Xiangping continued to smile: "Yes, that's what I meant, although I can't be sure now that dogs cannot avoid "one child solves two signs", but the dog's weakness must be hidden in similar means. Amazing! Lee Sedol! I can finally understand why he was the first to defeat a generation of dogs."

"The first generation of dogs has this weakness, what about the second generation of dogs and the third generation of dogs?"

"Haha, Ding'an, don't worry, this weakness is not only for the first generation of dogs, but also for the second generation of dogs and even the third generation of dogs. Anyway, before the two of you and I crossed, humans have not solved this problem for the time being."

When Li Xiangping replied to Lao Shi's question, he recalled several things in real history, and these things may be able to eliminate some chess players' blind worship of Go AI.

After the birth of AlphaGo, especially after Google published an algorithm paper, everyone saw that foreign dogs, large and small, appeared in various countries around the world. In a previous life, Li Xiangping knew the local dog. Among the core members of the "Star Array" R&D team, there are two friends who have a good relationship.

According to their introduction, they developed the "Star Array" based on the open source paper, and then their team decomposed them into about 80 specific topics. After these 80 topics were tackled, the "Star Array" came out.

In fact, it is not only them, but the situation of other teams is almost the same, based on the open source paper, and then decomposed into specific topics.

Just because of their different decomposition methods, we can see that the big and small foreign dogs and local dogs seem to have different "chess styles".

However, what interests Li Xiangping is not the style of chess. What he is interested in is that after the team developed the "Star Array", they took on a new task, which is to develop a smart traffic light system for a city in the Yangtze River Delta.

According to their own words, developing an intelligent traffic light system is much more difficult than the "Star Array". The "Star Array" only needs to tackle 80 problems, but the traffic light system has about 130.

The last traffic light system was also developed and is already operating in that city.

The system worked well most of the time, not only saving manpower, but also making traffic flow more smoothly throughout the city.

It's just that during the trial operation, the system also had two problems, and the command was disordered, causing traffic jams in the entire city.

What's the problem? There was an extreme situation. At that time, a heavy rainstorm occurred in the city before get off work, causing many vehicles to be blocked on the road. When the number of vehicles exceeded the limit of the system, it appeared disordered.

Looking back at Go, there are extreme situations on the chessboard, and dogs are also prone to disorder.

Are these two similar?

And the root of all this lies in the basic algorithm. As long as the algorithm is not corrected, this problem cannot be solved.

Li Xiangping later asked his friends: What is the level of your traffic light system?

My friend smiled. He said that if it was changed to Go AI, it would be no problem to hang and beat the second-generation dog. As for the third-generation dog, considering that Google developed the "dog ancestor" after all, they should be more polite and think that it is slightly better than the third-generation dog. Almost.

However, my friend added that if it is replaced by the strongest Go AI "Exquisite Art" in China at that time, it should be on par with the three-generation dog "Alpha Yuan". If the two really want to compete, it depends on the hardware competition. Whoever configures the stronger hardware wins.

To be honest, when Li Xiangping first heard this, he, a layman, was still very surprised. Maybe it was also because of the "dog ancestor". Like most chess fans, he subconsciously thought that "Alpha Yuan" was stronger.

So Li Xiangping continued to ask, and wanted to know what the professionals who developed AI thought.

The friend told Li Xiangping: Considering that A Fayuan had such a glimpse, he only made a hasty appearance and retired from the rivers and lakes. In this way, if Google is excluded for commercial reasons, they hide more advanced algorithms, then only from public appearances From the point of view of the dogs, they are indeed on the same level as "Exquisite Art".

Li Xiangping asked at the time: "Then what do you mean, in fact, the three generations of dogs have not made up for this flaw?"

The friend's answer was without any hesitation, and his answer was unequivocal:

"No, I just said that, excluding the situation where Google hides more advanced algorithms, this loophole must exist, because such loopholes cannot be repaired simply by patching."

Li Xiangping continued to ask: "Then do you think Google doesn't have a more advanced algorithm?"

The friend smiled and said that he couldn't say it well, but he personally thought that there should be no.

Li Xiangping asked why again.

The friend told Li Xiangping that since the Google paper came out, scientists all over the world have actually discovered the flaw through practice, and they all want to find a solution.

Unfortunately, the algorithm is the most basic basic discipline, but it is the most basic thing, but it is the most difficult to solve, so until now, there is no better algorithm.

The only leader is Google, and one of their teams has come up with an idea, which seems to be the most reliable idea of all, but at the time, it was just an idea. They hadn't built a model, so they couldn't. A mature algorithm.

Next, the friend was excited to introduce Google's new ideas to Li Xiangping.

Obviously, Li Xiangping's friend is also a kind of nerdy science and engineering man. When he told Li Xiangping that, it was just playing the piano to the cow. How could Li Xiangping understand that?

Only know that if there are real numbers, there are imaginary numbers, and if there is positive matter, there is antimatter.

Google seems to want to use similar ideas to make up for that defect, which sounds very advanced.

The only thing that impressed Li Xiangping at that time was listening to the friend's introduction: the core members of the team that proposed this idea at Google were also three Chinese, two from Tsinghua University and one graduate from the Chinese Academy of Sciences.

This made Li Xiangping wonder what it was like at the time.

When thought of this, Li Xiangping was unwilling to continue to recall:

"Haha, bro Ding'an, let's discuss this issue later. Today's game is not over yet, let's go and see how other teammates are doing."

Writing this kind of chapter is really exhausting, so today I am confidently asking for votes, all kinds of votes!

(end of this chapter)

Chapter end

Report