Global monopoly of technology

As the game enters the confrontation stage, Luo Sheng is looking forward to watching Nafu's first perspective to see how it performs.

This confrontation is [Human Race VS Zerg], the human opponent is a Zerg player, and Neve chose Human Race.

As soon as the game started, Luo Sheng, who watched Naf's first-view perspective, felt extremely embarrassed, because "she" controlled the four farmers to sway around the map, and then went back to mining for about half a minute.

Yes, the start was bad.

This is also peat, which is really mentally retarded.

This scene directly saw Luo Sheng drunk.

Those who didn't understand thought that "she" was full of curiosity about the world, and she was indeed mentally retarded.

Subsequently, Nafu built two houses in a row. Luo Sheng thought that "she" would put down the barracks building, but "she" did not build the barracks.

Instead, he saved a sum of money to open a second mine...

Luo Sheng, who watched the first perspective, suddenly felt that Nave’s current behaviors were just like the first time a pure newcomer came into contact with the game StarCraft, there was no difference.

As time went by, the Zerg players who were fighting against it came to scout and saw the opponent's start.

Zerg players also looked silly.

What the hell?

What kind of play is this?

So a wave of puppies came over. At this time, Neve had just built the first machine gunner and hadn't come out yet.

With a wave of tentative offensive puppies from the opponent's reconnaissance unit entering the mining area, Neve resisted for a while and was helpless, and finally played GG to end the game.

The first confrontation was killed by a small group of scout troops.

In fact, the human player on the opposite side is not very powerful, just an ordinary player. If you want to meet a real top professional player, Nave needs to fight constantly in the ladder and defeat a personal opponent. Every victory will match even more. Strong opponents until you reach the top of the pyramid to fight against real top professional players.

Of course, with Luo Sheng's technology, the algorithm of the matching system can be forcibly modified to allow Nave to directly match the top players, but he did not do that.

Currently, StarCraft is a popular e-sports game, and there are a lot of people playing it, especially in Korea.

There is even an exaggerated statement that at the moment, in South Korea, a man aged 10-40 will be ridiculed insanely if he says he can't play StarCraft.

...

Soon, the second game of confrontation began, this time it was the TVT confrontation game, namely [Human Race VS Human Race].

In the second round of confrontation, Nave made significant progress. Although this round of confrontation was also lost, at least he pushed the soldiers to the opponent's home.

After a wave of main forces were sent away, they were pushed back by the opponent's mechanized tactics.

Play GG again.

Reopen quickly next.

...

When it came to the fifth game, it was TVZ again, which was [Human VS Zerg].

Nave, who lost the game one after another, finally showed the extraordinary side of "she". Her strong learning ability allowed "she" to adapt to this level of "TVZ" confrontation, and easily defeated the opponent with operation.

Although this match against Neve defeated the opponent with a level comparable to top professional players, one of the details is that "she" never makes a second mistake in the same place.

To put it another way, Neve's learning progress is too terrifying, and every confrontation is learning for "her".

Through these five matches against human players, Luo Sheng also saw a defect exposed by Neve, that is, apart from specific division tactics, "she" cannot form a flexible force distribution.

In the fourth game that was lost, a high-level human player continuously harassed Neve’s base on multiple lines, and once "she" returned to defense, she ran away and waited for "her" soldiers to go out. When attacking, he turned back and continued to harass.

In this confrontation, in the face of the human players' flexible strategy of dispatching troops, Nave can only passively respond to the battle, unable to form an effective attack on the human players, which also led to the final snowball effect.

Finally lost.

After the fifth round of the confrontation, Luo Sheng asked Nafu to stop matching his opponent, and then re-executed a major upgrade and optimization.

He feels that such learning efficiency is still too slow.

Nave's agent uses deep learning algorithms. After Luo Sheng has carefully upgraded it, it includes neural networks, self-play through reinforcement learning, multi-agent learning, imitation learning and other algorithms or technologies.

The so-called self-play is simply self-learning, through constant trial and error to find the game that maximizes your winning rate.

Through self-play and reinforcement learning, it constitutes the basic paradigm of open-end learning algorithms, learns human strategies, and ensures that Nave’s agents continue to explore these strategies throughout the game, so that "she" can be upgraded. And progress.

In addition, because StarCraft has a huge space for action, many previous reinforcement learning methods become ineffective.

Therefore, Luo Sheng also developed a new algorithm for non-strategic reinforcement learning for Neve, so that "she" can effectively update the decision from the old historical decision game.

And Luo Sheng's ultimate goal is not to make Neve an "IMBA" StarCraft AI player, but to use this as a starting point to make "she" competent in other complex and changeable application scenarios.

If "she" just became a super master in the StarCraft game, and changed the scene, such as in Go, language or other fields, it would not work, it would still be a pure artificial mental retardation, and there is a huge gap between true artificial intelligence.

...

As time passed, Luo Shengmao spent more than three hours in his office unknowingly, completing the second optimization and upgrading for Neve before leaving get off work. It was too early.

Before Luo Sheng left the office, he put Neve on the Blizzard Star Battle.net ladder, letting "she" constantly confront human players.

After getting all this done, I’m not in charge. I have to go to the Cote d’Azur company in the afternoon to do some work, and there is a meeting.

The stocked Nave is doing one thing, constantly fighting against human players.

Behind Nave, "she" is supported by almost the entire Sioux City data center, which consumes computing power and electricity.

In fact, if there is no money, it is impossible to raise an artificial intelligence, let alone the development stage.

If it is an artificial intelligence research and development institution like IBM, the developer may restrict its development, but Luo Sheng, who has witnessed artificial intelligence, is the easiest for development.

But there are still problems restricting him.

That is the hardware. Nave's growth is limited by the current hardware level. If there is a better hardware foundation, it will usher in a qualitative leap.

Deep networks require high-end GPUs to train in a reasonable time with a large amount of data. These hardwares are very expensive. Of course, there is no such thing as expensive in Luo Sheng.

But the problem is that there is no hardware that meets his requirements. If you have money, you don’t have to buy it. From now on, you will have to start hardware yourself

It is not feasible to train deep networks to achieve high performance without advanced hardware foundation.

Not only to use high-end GPUs, but also fast CPUs, storage, and fast and large-capacity RAM.

...

With the passage of time, through constant battles, Nave's winning rate started from the first game of "her" tournament career, and gradually reversed, and the winning rate is improving rapidly.

At the operational level, Neve is undoubtedly the number one in the world. No human player can compete with "her" at this level, even if it is restricted by APM and EPM.

But this is not the point. The point is the "consciousness", "behavior" and "decision-making" that Nave showed in the game. This is the embodiment of intelligence.

...

Unknowingly, Nave has already played 27 games on the ladder, and the winning rate has climbed to about 70%, which is a very high winning rate.

In order to improve the efficiency of Neve’s "evolution", Luo Sheng allowed "she" to enter the massive database of Bluestar Technology Group and some StarCraft websites to access the "battle report" videos of some top players while fighting against humans. Players play back the video of the game and train in parallel, and then fight against other opponents for training.

A script running in the background will make "she" iteratively upgrade every once in a while, update it through reinforcement learning, and then conduct new confrontations.

In each iteration, a new opponent is matched and the opponents that have been defeated are frozen at the same time. The probability of matching the opponent determines the learning objective function used by the agent, thus maintaining diversity while increasing the difficulty.

With the progress of the self-game, Nave gradually developed more and more mature tactics, or relatively optimal solutions. This process is similar to the process of human players discovering tactics. The new tactical system continues to defeat the old tactical system. Then replace the old tactical system.

...

In the 32nd round of the confrontation, just like climbing the ladder, Nave constantly defeated his opponents, constantly self-iteration, and became stronger and stronger, and the human players he encountered became stronger and stronger, but 90% of them have become "her" "'S men are defeated.

At this moment, another round of confrontation began, and the players that Nave matched had reached the level of professional confrontation.

All the human players who were defeated by "she" didn't know that their opponent only started contacting the game today, let alone that their opponent was not a human, but an AI.

This round of confrontation is [Human Race VS Zerg], and the IDs of the two sides are [Nov] and [=YJ=F91].

At the beginning of the battle, the two sides started operations and reconnaissance normally. The human player whose ID is [=YJ=F91] is located in a room in Hangzhou. He has already left the "Yijing" team.

But now this ID is still retained, and he is now a real interstellar professional player, in the domestic interstellar world is a first-class top master, Jianghu people gave the nickname "Jiangnan Dog King", as a Zerg player, his puppy operation Five minutes before the game starts, it is known as world-class.

In five minutes...

Well, if F91 goes to South Korea, his strength is a dishwasher over there. Anyone who squats at the drinking fountain can sling him, and any Korean interstellar professional player can sling him.

"Nov...? Whose vest is this ID? It's a good operation, a bit of strength." At this time, the confrontation has entered the mid-stage, and F91 did not win the game with the same light as the previous one. On the contrary, it was quite difficult to play, and he even felt the pressure of his opponents.

The key is that he is very confused now. He has never seen this ID. The players in the domestic interstellar circle who can have this strength are basically the cards, and they basically know them.

But this is the first time I have seen the ID of Nov, and the style of play is also the first time I have seen it, which directly allows F91 to exclude those domestic interstellar masters he is familiar with.

"It doesn't matter, wherever the gods come, if you meet your brother, you will die." F91 meditated in his heart, and then his main force began to go out, ready to attack.

[The chicken field is under attack!]

But shortly after his main force went out, the sound of the headset frantically reminded that the base was under attack. F91 subconsciously cut the screen to his main mine, and then saw a picture that shocked him extremely.

"Roasting, roasting, what is it? What is this? Spicy butt operation... roasting!!"

It turned out that at the same time that F91 was preparing to launch an attack, Nave had already sent five ships to launch an attack, and also deliberately went around the six o'clock position on the map to avoid the vision of the opponent in the middle of the map, and repeatedly Use radar to do anti-reconnaissance in advance to ensure that your airdrop troops are not detected by the other party.

If Luo Sheng saw it, he would be very satisfied. This might seem an unusual decision for human players, but it would be different if it was executed by an AI.

This progress is not trivial.

It is worth mentioning that although Neve is an AI, "she" does not "open the picture", there is also the fog of war, and the operation also needs to cut the screen like a human player, and ensure that these parameters are the same as a human player, otherwise it is Bullying.

The F91 troops had just left the house, and Nave's five ships successfully arrived at the opponent's main mine highland.

Then, F91 saw the operation screen that shocked him.

The opponent's five transport planes actually dropped troops while flying, making the combat troops spread on the map like a carpet, so as to achieve the greatest firepower output in the shortest time.

This wave of airdrops directly destroyed the main mine of F91, but for a while, another wave of carpet airdrops by three ships completely destroyed the sub-mining base of F91 and the few remaining troops.

Seeing the two blossoms at home, the main force has advanced for more than half of the time, and it is meaningless to return to aid at this time, and it will definitely lose.

So F91 decided to take a shot.

Change home with your opponent!

The main force does not return aid, only in this way there may be a chance of victory.

As a result, the Zerg troops rushed to the mining area where the human race was located regardless of the blooming situation at home.

"The architecture is a bit good..."

F91’s troops pushed to the opponent’s mine, and was a little uncomfortable with some construction cards in the middle, but after a while, he exclaimed again: "What kind of hot butt? This... how could he still have so much in his house? Troops? Riot soldiers are like a diabetes collapse?"

The shocked F91 was struggling and failed to replace the opponent's sub-mining, and the main force had lost two-thirds, and the family had already blossomed twice.

Having been unable to return to the sky, he sent all the remaining troops A up and gave himself the reason to play GG.

After exiting, F91 did not qualify. He immediately clicked on the video to watch the game. From the beginning to the end, after watching the entire first-view playback of the opponent, the opponent's operations, rioters, climbing technology, and offense including defense and many more.

This ID, which has never been seen before, was extremely shocked by the operations and operations it showed. This is definitely a level that only top Korean players have.

Especially the carpet airdrop troops when the opponent was attacking. The moment he saw that scene, he felt that this ID was the operation of the interstellar "human emperor" Boxer, and the carpet airdrop was the ultimate operation he created.

In fact, F91 knows how to operate a carpet airdrop, but he can't do it at all. This is a flaw.

For the carpet airdrop of the opponent’s 5 ships, the operation process requires that each transport plane be numbered one by one, and then when landing the troops, the left hand needs to switch the formations in turn, and the right hand operates the mouse to click on the soldiers in the transport plane. Complete all airdrops.

It's simple to say, but it's actually extremely difficult. It is impossible for ordinary players to do it. To complete such an operation, it means that it needs to achieve 40 precise and coherent left- and right-handed coordinated operations of the mouse and keyboard in about three seconds. The sequence is wrong.

If you just do this, then StarCraft won't be so difficult to play, you have to run the home while doing this.

At this moment, F91's mind is full of question marks: When did interstellar players with such fraudulent operations appear in China?

This level of operation has definitely reached the world's top level, including smooth operation and terrifying violent abilities. If it were not for playing GG in person, fighting in person, and watching the video in person, F91 would never believe that this is a domestic player, it would definitely believe it. This is the vest of a top Korean player.

After the shock, he continued to qualify.

I hope I can touch it again, even if it is a abused dish, I am very happy. For professional players, I am very happy to fight against a master who is better than myself, even if I am slammed by the opponent, because I can improve.

But then he can no longer be ranked. Any human player defeated by Neve will not be ranked second.

After Nafu reaches the top in China, "she" will enter the Korean battle net and fight against the real world's top interstellar professional players.

In this short day, Luo Sheng didn't know that this ID had suddenly become popular in the domestic interstellar circle.

The great players who became "victims" from all walks of life started posting on the social forum asking which god the Nov that suddenly popped out was, and actually blasted the masters one by one.

GG before the hands are returned.

And the "victims" all gathered together to "complain" how they were blown up. At the same time, some videos of the battle with Nave were also uploaded to Bluestar Video, and more and more interstellar players came here. Watch the video.

The ID called "Nov" went viral in the domestic interstellar circle almost overnight.

The topics in the circle are all discussed around this mysterious and powerful human player, and they are also trying to find the real person with this ID.

...