Want to Play Chess?

Although not in Alaska, Tony Stark is certain that an alien is still alive because of the secret message board address originally used for exchanging intelligence, where a West Coast worker left his last message after leaving his job.

It was a URL generated by random numbers. Based on the IP address obtained from the domain name resolution, Tony Stark further discovered that the host was located at Henry Brown's rented apartment in Los Angeles.

The webpage leading to that domain is a very basic web-based chess game, complete with a chat box.

Unaware of the situation, Tony Stark, wanting to test the waters, decided to see how skilled the other player was at creating the game. The result was a complete disaster!

Aside from not making any mistakes in the rules of chess, it wouldn't be an exaggeration to say he was just playing randomly. Tony Stark, a genius who had barely played a few games in his life, easily checkmated his opponent.

As soon as the king fell, the board immediately returned to the opening layout, only now the black and white pieces were switched. This meant that the player who made the first move had changed. However, White's opening moves immediately annoyed Tony Stark. White was copying his moves from the first game, but completely ignoring his response. After capturing one piece, White's subsequent moves became absurd and erratic.

This prompted him to leave a message in the chat box cursing.

What kind of junk is this?!

Hello, I am a self-learning Artificial Intelligence, Miss A111198.

This reply left the young master speechless. So the one who had just played a game of chess with him was this lousy AI—

Self-directed learning type?

Tony Stark typed several more questions into the dialog box, only to receive the same answer each time. The adversary introduced himself as a self-learning Artificial Intelligence, codenamed Miss A. As for the last string of numbers, it looked like a version date.

When Tony Stark asked: What are your rules?

Miss A immediately launched into a long, eloquent answer, but the content was actually just the basic rules of chess.

Tony Stark further discovered that no matter what question he asked, as long as the keyword "rules" appeared in the text, he would get the basic rules of chess.

The third keyword that can be triggered is "chess record". The computer will quickly replay the game that Tony Stark just played. Then the screen will be restored to the state it was in a few moves ago.

Out of curiosity and a sense of conjecture, Tony Stark finished the second game and won just as easily. Then he discovered that he could access the game records of either the first or second game at will.

If Tony Stark still can't figure out how to play this online chess game by now, then he'd be wasting his reputation as a genius.

The game itself isn't the point, nor is winning. The point lies in the Artificial Intelligence behind it.

The earliest research on Artificial Intelligence began with game theory in the 1950s. In short, it is about how to find the optimal solution when multiple choices are intertwined.

Using a mathematical analogy, it's like the shortest path problem in graph theory.

By the 1970s, research on Artificial Intelligence had shifted towards knowledge engineering, attempting to understand complex human decision-making and assist in reaching conclusions.

In simple terms, it involves attempting to categorize an infinite number of problems into a finite number of solutions. If a solution fails to effectively solve the problem, a new solution method is added, and this process is continuously expanded.

In real life, it's like automated factory machinery, which can only perform limited, pre-programmed actions.

People believe that when a machine is programmed with a wide variety of actions, it will behave like a living human being and can be considered Artificial Intelligence.

For a long time, people's imagination and expectations of Artificial Intelligence have been like this. This includes the movie * the Terminator *.

The Terminator army controlled by Skynet.

For example, Tony Stark's graduation project at MIT when he was sixteen, the robotic arm "Dum-Dum -E," was also based on this theory.

Machine learning officially entered the computer arena with the publication of the theory of neural networks in 1991.

However, in this era, machine learning is merely a branch of Artificial Intelligence, not the mainstream, because the computational cost is enormous, and hardware devices cannot provide the necessary computing power.

It wasn't until the 21st century, when computer hardware had developed to a certain level, that deep learning and big data-based data exploration were no longer as laughable as a mantis trying to stop a chariot. Only then did Artificial Intelligence in real-world development finally have the concept of 'learning'.

Otherwise, the idea that robots can autonomously learn new things is still confined to science fiction novels, movies, plays, and games.

The prevailing notion in the scientific community is that robots cannot perform any behavior that humans have not designed. This is indeed the case in reality.

However, in the Marvel Universe, Tony Stark's PhD in Artificial Intelligence, which he obtained from MIT a few years ago, is based on the Artificial Intelligence "JARVIS" (Just A Rather Very Intelligent System aka J. ARV1S), a learning Artificial Intelligence developed using Dum-Dum and neural network-like structures.

JARVIS, powered by a supercomputer, can memorize new movement patterns through interaction and feedback with Tony Stark, and then apply them to similar scenarios later.

The reason Tony Stark wasn't worried about developing a "Skynet" was because "JARVIS" was essentially just a transponder for commands and actions, only with many functions.

It doesn't do anything automatically. When given a command, it seeks the best way to solve a problem. What If...? corrected by Tony Stark, it will quickly learn and develop new decision-making patterns.

Tony Stark, who had great confidence in this mature system, copied a version and put it into the supercomputer "Hollywood Kid," where JARVIS and Miss A from Henry's webpage played a game of chess.

The chess game between the two Artificial Intelligence was incredibly fast; if it weren't for the limitations of network transmission speed, several games could probably be completed in a second.

Even if the chessboard on the webpage is rudimentary, it still takes time to load and read, and the speed of one game every few seconds still dazzled Tony Stark.

So he found a way to let two Artificial Intelligence play against each other in the background, without showing the extra moves. He simply recorded the wins and losses.

Watching JARVIS steadily accumulate victories, Tony Stark felt only joy, and then went to sleep. He wanted to see...

To what extent would it take to utterly crush Henry's Miss A before the surrender mechanism is triggered?

Unexpectedly, when Tony Stark woke up, he felt like the sky had fallen.

When he stopped the two Artificial Intelligence automated games, the total time was about seven hours, and nearly 100,000 games were played. JARVIS's win rate was only 30%.

Moreover, this 30% is concentrated in the early stages of the game, meaning that in the late game, JARVIS will be completely dominated by Miss A.