Chapter 357: Island Artificial Intelligence
In actual quantum computer programming, some unexpected problems may be encountered.
In the following time, Chang Xiaoping took Xu You to familiarize himself with the new quantum optimization device architecture.
Through practical operations, Xu You became more familiar with the entire framework and further deepened his ideas.
"Chang Gong, my current idea is to test the computing and programming application capabilities of the island quantum computer by developing an artificial intelligence system that plays chess."
In fact, before coming here, Xu You already had such a preliminary idea.
After understanding the device architecture in detail, Xu You was basically certain that this idea was feasible.
"Artificial intelligence for chess? What kind of chess is it?"
The chess games that we play most often include Chinese chess, international chess, Gobang, Go and so on.
Chinese chess is the most common chess game in my country and can be seen everywhere on the streets and in parks.
In terms of popularity, there is definitely no other chess game that can compare to Chinese chess.
But when it comes to testing computer performance, the most critical factor is not popularity but complexity.
In this regard, Go is definitely the most complex chess game.
The total number of changes is approximately 10 to the power of 172.
Even the most advanced electronic computers cannot complete such a huge calculation before the universe is destroyed.
Although quantum computers are much faster than electronic computers in terms of computing speed.
It is also impossible to complete this nearly infinite calculation within a limited time.
However, the inability to exhaust all situations does not mean that it is impossible to create an excellent Go artificial intelligence.
AlphaGo is a good example in this regard.
Thinking of this, Xu You replied:
"Of course it's Go. If you want to test the capabilities of a quantum computer, you must choose the most complex board game."
"But... can our quantum computers really be programmed to this degree?"
When Chang Xiaoping heard Xu You's idea, he did not show any optimism.
Chang Xiaoping, who is familiar with the operating environment of quantum computers, knows clearly that no quantum computer in the world currently has such capabilities.
In fact, there is still a big gap when the distance reaches this level.
Afterwards, Chang Xiaoping explained to Xu You the various difficulties that might be encountered in this research based on his own experience.
"Our quantum computers have great advantages when dealing with exhaustive methods. But many practical problems cannot be solved by exhaustive methods alone. With the new quantum optimization device framework, we do have some subversive ideas for using quantum computers to solve traditional problems. However, solving problems like Go still seems a bit difficult."
Xu You was not unaware of these difficulties, but he had already prepared for them.
"I understand, Chang Gong. I think we can still give it a try first. If this direction doesn't work, then we can start with something simpler."
Xu You’s goal is to make the system created by the Island quantum computer have the ability to conduct deep learning on its own, just like AlphaGo.
Once this is achieved, it will also have more important applications in other fields including medicine, energy, etc.
"Okay, Professor Xu, we will definitely fully support your research."
Since Xu You insisted on his own idea, Chang Xiaoping would definitely not have any objections.
After all, even Han Shubin, an academician and the head of the team, will do his best to support Xu You.
Not to mention a software engineer like Chang Xiaoping.
After finalizing the plan, Xu You officially started this work.
Although we are only using quantum computers to create an artificial intelligence software for Go.
However, necessary Go knowledge is a must.
Fortunately, Xu You had taken an elective Go course at Jijing University before, and his Go skills improved greatly during the three games he played against Nie Jie.
In order to enhance your learning experience and exercise your brain power.
Xu You also plays blind Go with himself in his spare time, and his Go skills are constantly improving.
Even though Xu You has not formally tested his Go skills, at least among amateur Go enthusiasts, Xu You is definitely considered to be quite good, and he has the ability to compete with some professional players.
After a detailed understanding of the new quantum device optimization framework and a thorough analysis of the game of Go.
Xu You identified several important parts of this Go quantum artificial intelligence system.
The first part is the chess moving system.
Determine the current board situation, sample and predict next moves.
The second part is the evaluation system.
Based on the given situation, determine which of the black and white pieces has the advantage, and estimate the approximate chance of winning.
The third part is the quality system.
Because in actual competitions, the game time is limited.
It is impossible to calculate every step completely before making a move.
Quality planning must be carried out according to the length of time.
When necessary, you can sacrifice the quality of playing chess to ensure that you do not exceed the time limit.
This is because quantum computers have significant limitations and new quantum device optimization frameworks are still in their infancy.
The programming process was not easy at first.
Xu You needs to make repeated adjustments to ensure that the program is compatible with quantum computers.
In order to be able to frequently use the island quantum computer to verify the correctness of the program.
Whenever Xu You has time, he goes to the Island Quantum Laboratory of HUST.
Later, Xu You even asked Han Shubin to help him arrange a dormitory at Huazhong University of Science and Technology.
If I start my research too late, I will just live in HUST.
During his research day by day, Xu You solved one problem after another.
After continuous improvement, Xu You finally solved all three major problems and completed the entire quantum programming work.
Xu You named this artificial intelligence system HaidaoGo.
Completing the quantum programming work does not mean that HaidaoGo has already acquired strong chess skills.
It also requires continuous deep learning to improve chess skills.
To this end, Xu You imported a large number of chess game records into HaidaoGo, and let it improve by constantly reviewing the games by playing against itself.
Because the calculation speed of the island quantum computer is very fast, a game of chess can be completed almost instantly.
This allows it to accumulate a large amount of game materials in a short period of time and continuously improve its chess skills.
In just one day, HaidaoGo's chess skills improved by leaps and bounds.
In order to test HaidaoGo's current strength, Xu You tried to play a few games with it.
