I am the only one who practices magic: I practice magic in the city

Chapter 221 Are We Really Going to Lose? (Second update with 4000 words!)

"I didn't expect that the income of Goryeo's top chess player is so low. It's not even as high as the income of senior executives of large companies, right? Lee Se-sil has been the world's number one for so many years." Shen Shuyao whispered in the crowded game room.

Fang Yu looked at the big screen at the scene and whispered, "It was still okay before, but the prize money for Go competitions has not increased in 30 years, which is mainly related to the low commercial value of Go."

"It's also thanks to you that you settled the matter with the Goryeo Chess Institute and Lee Se-sil within two days. I thought it would take at least until the first day after the competition to be finalized." Fang Yu smiled.

After receiving the news that Lee Se-sil was going to compete with AlphaGo, Fang Yu immediately asked Shen Shuyao to fly to Goryeo to discuss cooperation with the Goryeo Chess Institute and Lee Se-sil.

For this purpose, Fang Yu specially asked his father to find him a Korean translator to go with Shen Shuyao.

Through my grandfather, I found the Korean Embassy and urgently issued visas to all the employees of Youzi Technology.

It was easy to deal with the Goryeo Chess Institute. With a sponsorship of 5 million Malaysian dollars, they were not asked to reject Gugou. They just wanted them to play "Xiangtan Chess" in the last three games after Lee Se-sil lost the first two games. The Goryeo Chess Institute easily agreed.

The difficult part is Li Shishi's side.

"Li Shishi refused to meet with me. Finally, I went to his daughter's elementary school, and when his daughter was getting out of school, I convinced his wife to do the work for him." Shen Shuyao raised her nose proudly.

"You're so awesome. You're indeed my good wife." Fang Yu scratched Shen Shuyao's nose with his hand.

"Look, his wife is over there." Shen Shuyao's neck turned red when Fang Yu called her a good wife. She quietly pointed to a young woman in her thirties who was also sitting in the first row.

Fang Yu just glanced at it.

general.

"Ah, Li Shishi has made his move." Shen Shuyao saw on the big screen that Li Shishi placed the chess piece at the position 33 according to the placement point given by Juzi.

Neither she nor Fang Yu knows Go. Fang Yu just looked at the rules of Go in the past two days. He can solve simple life and death problems with his IQ, but he has no idea about the fixed patterns.

Seeing that AlphaGo also started to make moves, it didn't take long for the two sides to make more than ten moves. Fang Yu couldn't help but yawn.

This chess game is so boring.
-
"This chess game is like a fight between gods!" Liu Weiping's deep voice was extremely excited.

"The black chess piece started with a three-three move, followed by a double flying swallow. The white chess piece was not to be outdone. Although I still don't understand what this flying move means, I believe it will be useful later."

"Generally speaking, we would not have played the cutoff in Lee Se-sil's 23rd move. I don't know why he chose to cut off on the right side."

"Generally speaking, in this kind of chess pattern, it would be more advantageous to cut off on the left side." Liu Weiping moved some of the chess pieces on the board and arranged the chess pieces at the corners.

"If we play like this, the original advantage of black pieces in the lower left corner will be weakened, giving white pieces a chance to attack." In the studio of Tengxun Video, Liu Weiping pointed at the chessboard in front of him with a fan.
-
"Li Shishi's cut is indeed a bit strange. If it is cut on the right side, the risk of black pieces in the lower left corner will indeed be greater. Gu Jie, what do you think is Li Shishi's intention in cutting on the right side?" In the studio of Huanshi TV, Chen Ying threw the topic to Gu Jie.

"Uh... it should be for future external support." Gu Jie randomly placed a few black pieces, "If this chess piece is here, there may be a lot of changes in the future."

Chen Ying covered her mouth and laughed, "You mean, Li Shishi and the foreign aid want to compete with AlphaGo in computing? Professor Liu, you also said before that the thing that humans can't compete with AlphaGo the most is computing, right?"

"That's right. For move 23, the reason why professional chess players cut on the left side is because of the problem of calculation amount. The space on the right side is too large, and there are too many changes in the future. And based on the first two games, AlphaGo is obviously better than any human in this kind of spatial calculation." Liu Ziqing nodded firmly.

Upon hearing this, Gu Jie remained silent.

"What if Li Shishi's foreign aid is also an AI?" Gu Jie was silent for a few seconds and suddenly spoke.

"Haha." Liu Ziqing laughed a few times. "Gu Jie seems to have gone insane. How can it be so easy to have an AI comparable to AlphaGo?"

"Google was able to develop AlphaGo based on its strong foundation in neural network deep learning."

"Currently, the alphaGo algorithm is not open source. By the way, speaking of this, I would like to apologize first."

"During yesterday's live broadcast, I hadn't figured out the principle of AlphaGo, and was criticized online. After I went back, I also read the paper published by Google. The core of it is actually an algorithm problem."

"The tree search used in Google's algorithm is actually studied by many universities in China, but Google is still very powerful in terms of the depth and complexity of neural networks and the understanding and application of tree algorithms. There is probably no company in China that can compare with Google at present."

"To put it bluntly, the research on neural networks in our country is far behind that of everyone else in the world, except for a few large companies in Miami."

"Most of them know about this and just take existing stuff and modify the code to do experiments. The gap in level is as big as the gap between Shimen Electronic Information Vocational Technical College and Xinda University, for example."

"There are issues with resources, talent, and the academic environment."

"But as long as we have a direction and with the investment of major domestic companies, I believe we will definitely catch up."

When Liu Ziqing said this, he realized that his butt seemed to be crooked just now, and he quickly moved back to make up for it.

"But AlphaGo has just appeared. It is impossible for an AI to be as good as AlphaGo, let alone the one that defeated Fan Hui a few months ago."

"I'm not saying it's impossible for it to appear in Da Zhou or Maiguo. I saw a report yesterday. The director of the AI ​​laboratory of Facebook is also Chinese. He was interviewed and said that Facebook is also doing the same research. The reporter asked him if he could defeat AlphaGo now, but he avoided the question. Why?"

“It’s because Google and Deepmind have such a huge advantage in this area.”

"Otherwise, without this confidence, how could they jump out and challenge Li Shishi at this time? Don't you think so?"

Liu Ziqing's words made the other two hosts nod their heads frequently.

"Hey, let's stop chatting. We've already reached move 33. The game is going very fast today. Let's take a look at the situation. Gu Jie, what do you think of these moves?"
-
"Julian, the winning rate has dropped! What happened? Was there any problem with the last move?" Hassabis nervously held David Silva's arm beside him.

As soon as Black made move 33, the winning rate curve on the screen dropped from the original evenly matched 50% to 49% with a slight jump.

Although it was only a 1% change in win rate, in the first third of the game, AlphaGo had never experienced a negative fluctuation in win rate.

Julian, a handsome young blond boy, licked his lips and said, "The problem is not the 33rd move, but AlphaGo just discovered that there was a problem with its layout in the 18th move, while the opponent discovered this flaw in the 19th move and made a targeted layout."

"At the 33rd move, the layout just made in the 19th move echoed each other. At this time, AlphaGo judged that the subsequent defensive pressure in the middle of the abdomen would increase, so it reduced the winning rate."

"However, judging from AlphaGo's simulation of the future situation, there shouldn't be any problem in the end." Julian clicked on a few simulation pictures. These were different predictions and analyses made by AlphaGo based on the current situation.

"Fan, what do you think?" Hassabis was worried and began to consult the only Go professional in the room - Fan Hui.

Fan Hui's eyebrows, nose and eyes were all wrinkled together. He thought to himself, you think too highly of me. I'm just a second-dan player. I didn't even understand six or seven of the more than thirty moves. What do you want me to think?
Fan Hui's reaction made Hassabis even more unsure. "David, go to the lounge of the Korean Chess Institute and see what they say."

David Silva was also very nervous, even more nervous than Hassabis. After hearing this, he quickly ran to the lounge of the Korea Chess Institute opposite and came back in less than ten minutes.

"I didn't understand the Korean language they spoke." David's face flushed.

"Fuck! Why don't you bring a translator with you?" Hassabis rarely cursed in Chinese, using a curse word starting with F.

Before David could make any move, as AlphaGo made its 40th move, Hassabis saw on the screen that the winning rate had returned to 50%.

Phew, Hassabis finally breathed a sigh of relief.

He now finally understood Li Shishi's mood yesterday and the day before yesterday.

Shit! Where on earth did this chess player come from?
After watching for another ten minutes, based on Hassabis's own chess skills, it was basically impossible to tell which move was good and which was bad.

"Demi, did you notice that Lee Sedol and his team didn't take a long time to think today? They played very quickly. I recorded the average time it took Lee Sedol to make a move, and it was only about 58 seconds."

Fan Hui's facial features were still wrinkled, as if this was the only way to prevent his glasses from sliding off his nose.

Hassabis was stunned.

"Such a fast move time is not much different from AlphaGo." Fan Hui seemed to be muttering to himself, and seemed to be talking to Hassabis.

"Moreover, this move..."

"Snap!" As Lee Se-sil's 53rd black move fell, AlphaGo seemed to be confused all of a sudden.

"Broken! It's actually broken!" Fan Huiteng stood up suddenly and pushed his glasses on the bridge of his nose in disbelief.

"What do you mean?" Hassabis asked nervously, and then asked Julian, "Have you calculated the winning rate?"

Julian quickly clicked the mouse and keyboard. "Not yet. This move seems to have caused great confusion for AlphaGo. It is searching the neural network through MCTS."

"Fortunately, we have fixed the bug in AlphaGo, so this move will not cause it to crash. Now it depends on how it responds next."

"This situation usually only occurs when AlphaGo is about to lose the game." Julian turned around and said to Hassabis with a serious expression.

Losing the game? Is AlphaGo going to lose?

"This move has completely cut off the connection between AlphaGo's two pieces." Fan Hui rubbed his eyes, "AlphaGo lost at least 7 points in this move."
-
"Black 53 breaks! Beautiful!" Liu Weiping snapped his fan shut. "By now, everyone should have understood, right? When Black moved at 41, everyone said it was a slow move, but I said this move was interesting. 53 breaks! Once this move was made, it was the best of the game!"

"I can't believe this move was made by Lee Se-sil, given his imagination and overall vision. Is this foreign aid so effective? Could it be Lee Chang-ho? Hahaha, just kidding."

"AlphaGo is now in a passive position. Let's see how it will respond. Huh? It's been three minutes, right? Is this the first time AlphaGo has taken so long to think?"

"By the way, it seems that Li Shishi played very quickly today, almost without thinking."
-
Li Shishi really did not think twice, because there was no need to think.

He was now completely shocked by this game of chess.

It turns out that AlphaGo’s chess skills are even stronger than he imagined!

In the past fifty or so moves, AlphaGo made almost no slow or bad moves. Every move it made was almost perfect, and its use of human patterns reached the level of being completely focused.

Only then did he realize that when AlphaGo played chess with him, it had not really shown its true strength.

No, it should be said that his strength is totally insufficient to allow AlphaGo to exert its full strength!

AlphaGo is a machine. Once it calculates that it can win, it will use the simplest method to ensure victory. Therefore, the gap between him and AlphaGo in each game seems to be not much, but he just can't catch up.

Even though he played his best game yesterday, he still lost to AlphaGo by half a point.

but!

And that mysterious chess player, no, it should be the artificial intelligence that is stronger!
The man named Hua sitting next to him had a headset on his ear. After each move made by AlphaGo, the headset would transmit the next move's location. Then the man named Hua would tell him, and he would make the next move.

And every move he made would make him so excited that he would get goosebumps all over his body!

Unlimited imagination!
The artificial intelligence that gives the landing point is hardly affected by any human patterns. It will take off first if it is told to, and will not complete the pattern if it is told not to.

If AlphaGo is the epitome of human stereotypes, then this mysterious artificial intelligence completely believes that human Go may be fundamentally wrong!

By the 53rd move, Li Shishi's hands had already started to shake.

It’s too powerful, too powerful, there’s no way a human could produce this!
who are you?
Is it the orange on the Internet?

"Snap!" After nearly five minutes of thinking, AlphaGo finally placed a piece, right above Black's 53.

A very normal response.

But after this move, the winning rate of White chess dropped sharply to 42% on the screen of the DeepMind control room on the sixth floor!

Everyone on the deepmind team had a look of shock and panic on their faces.

Is it possible that AlphaGo is really going to lose?
(End of this chapter)