Player Artificial Intelligences That Are Impossible To Beat

Artificial intelligence is the player who made the World Chess Champion Gary Kasparov give up in 1996 and overtake the best players of Starcraft II in the past years.

The usage areas of artificial intelligence technology continue to expand. We met a long time ago with artificial intelligence that can take pictures for us and respond to incoming messages for us, but such artificial intelligences does not make our life easier at higher levels, does not surprise us much, as it does not offer incredible features.

The reason is very simple. We can already perform these tasks ourselves, so artificial intelligences such as Siri and Google Assistant do not attract our attention. Of course, this also has a game dimension. Artificial intelligence developed for playing games attracts our attention because it offers us a very different challenge than Siri. long-term efforts in the hope of defeating him we give.

You’ve heard of artificial intelligence technologies that are impossible to beat, that beat even the most successful players. What makes these so special is that they too just like you thinking about the next movehave the potential to make the same decision as you.

Deep Blue vs. World Chess Champion Garry Kasparov

The unbeatable artificial intelligence technology rush was IBM’s Deep Blue, which was first developed for chess tournaments. Deep Blue, which could not provide the performance IBM wanted at first, when it faced World Chess Champion Garry Kasparov in 1996-1997. lost in three of the five matches and tied in two left.

After this defeat, IBM continued to develop Deep Blue and a new competition was held a year later. Next is Deep Blue, because he doesn’t want to waste time thinking too much He wanted to make a quick move, but this move was pretty terrible for the course of the game.

Deep Blue gary kasparov

Karparov, who did not realize the mistake made by Deep Blue, said that the reason behind the artificial intelligence making this move was, interpreted the next moves as advanced anticipation and afraid of defeat decision to withdraw from the race got it.

Although Deep Blue is successful, technology developing over time, IBM shelving this artificial intelligence caused. Deep Blue, one of the first invincibles of its kind with his chess moves, is currently on display at the Computer History Museum in the USA.

Watson, also developed by IBM, crushes everyone in the question and answer competition.

ibm watson

Two executives of IBM at a lunch Jeopardy! The most successful name of the general culture competition Talking about Ken Jennings’ achievements, he decided to develop an artificial intelligence technology that could beat him in this competition.

Named after the founder of IBM, Thomas J. Watson, this artificial intelligence began to be designed in 2005 and made its reputation all over the world with the competition held in 2011. announced.

ibm watson

Watson answered Ken Jennings and Brad Rutter in a question and answer contest and answered the questions one by one. A prize of $77k won. Afterwards, Rush D. Hold Jr., one of the important names of the US congress, could not slow down. and Watson, who competed against Bill Cassidy, managed to beat this duo.

Watson did not end up like Deep Blue. IBM later named Watson as “Watson Text to Speech” and “Watson Natural Language Classifier”. decision to convert to some advertising products took.

DeepMind has achieved advanced performance in all Atari games


DeepMind is not an artificial intelligence designed just for a single game like Watson and Deep Blue. On the contrary, a company called DeepMind Technologies laid the first foundations for what is today called general artificial intelligence. To beat DeepMind any Atari game designed it.

DeepMind was taught to use the information on the screen to get the highest possible score from the games. Designed in 2013, this artificial intelligence managed to achieve advanced performance in all games in Atari, and not only broke records to date, but also managed to beat human opponents in games such as Quake without any problems. DeepMind company was acquired by Google in 2014 after this success of gamer artificial intelligence.

AlphaGo versus Lee Sedol, the world’s most successful Go player

alphago lee sedol

The team that designed Deep Blue against Gary Kasparov in 1996, together with the most advanced technologies in 2016. can rival Lee Sedol, the world’s most successful Go player designed an artificial intelligence.

The work of this artificial intelligence called AlphaGo was quite difficult compared to its previous examples, because different strategies have to be produced for each situation in the game of Go, and different paths have to be followed. Opinions against AlphaGo, which has millions of reference options, were that it could not compete against Lee Sedol. In fact, Lee Ha-jin, Secretary General of the International Go Federation, It will not be possible to design an AI player equivalent to Lee Sedol. had stated.

alphago lee sedol

As expected, AlphaGo started the first set of their match against Lee Sedol with a win. Sedol, a professional Go player, Only got the fourth set against AlphaGo and stated that he wanted to start the 5th set with black pieces to confuse the AI. That didn’t work either, and AlphaGo took the last set. Defeated Sedol 4-1.

Even though Sedol is number one on Go in 2019 stating that artificial intelligence is an invincible entity He announced that he was leaving his professional Go life.

The good has the good; AlphaZero Masters takes AlphaGo from its top spot

alpha zero masters

Designed in 2017, AlphaZero Masters was an enhanced version of AlphaGo and its purpose is to play in chess, Go and Shogi (a version of chess played in Japan) to first place was to play.

AlphaZero not only defeated its brother AlphaGo, but also overthrew the computer games developed for it one by one. This AI is in chess after 9 hours of training, Shogi after 2 hours of training, and finally Go after 34 hours of training. to fully master succeeded.

AlphaStar joins esports

alphastar starcraft II

After chess, Go and Shobi, another area where artificial intelligence made a place for themselves was e-sports. Designed to take the lead in Blizzard’s strategy game Starcraft II, AlphaStar was soon launched. Starcraft players Proven better than 99.8%.

AlphaStar is one of the most complex artificial intelligence gamers ever designed. After all, his goal is not to take steps in chess by predicting the next moves, but Rank high in a challenging game like Starcraft II play.

Omron’s table tennis player “Forpheus” debuted at CES 2019

omron forpheus

We are used to the player artificial intelligence making moves on the computer, but at CES 2019, Omron introduced it. Table tennis robot named Forpheusbrought a whole new dimension to the event.

Forpheus, which has a 5-axis arm and a racket fixed on this arm, can move flexibly and quickly thanks to its image recognition and advanced technology, and can think and make unpredictable moves just like its opponent.

omron forpheus

Although the working principle of Forpheus seems simple, this player robot processes dozens of data into the system after each move. These data, which are processed very quickly through sensors and cameras, by analyzing the places where the ball falls and is met Allows Forpheus to make moves like his opponent.

Forpheus isn’t completely invincible yet, the developers at Omron are soon launching this gamer AI that they introduced in 2019. 48% closer to a professional table tennis player has succeeded.