Magic: The Gathering - the most difficult 'chew' game for all AI models

You may not know, but video games generally create the perfect teaching and training environment for developers in training their machine learning models, which doesn't need to be discussed. Many, but the problem lies in which games currently can create the ideal environment to train the most powerful AI models?

Talking about the famous AI systems of remarkable power, many people will think of DeepMind's 'monster' AlphaGo or the 'no-team' gaming machine for Dota 2 players named OpenAI. However, the throne will only really belong to which AI model can 'play' the Magic: The Gathering game - the most difficult game for AI that humans ever created!

Magic: The Gathering - the most difficult 'chew' game for all AI models Picture 1

1. OpenAI artificial intelligence defeated the current world champion Dota 2

A research team with the participation of three well-known scientists, led by independent researcher Alex Churchill, recently announced a research project related to preliminary storage hosted by game developers. arXiv used to create Magic: The Gathering, a card game, is arguably the most complex real-world calculation game ever created. The team then created a 'machine' called the Turing Machine out of Magic: The Gathering cards to conquer this game.

A little bit about Magic: The Gathering's gameplay, this game can be played by 2 or more players with different specifications, the most popular of which is to use a deck of cards with more than 60 cards, in which, must not contain more than 4 identical cards (except for land attribute cards). Each game will show a battle between the wizard called "planeswalkers". In it, the planeswalker must use magic, artifacts, as well as the creatures depicted on specific Magic cards to defeat the enemy.

Magic: The Gathering - the most difficult 'chew' game for all AI models Picture 2

1. MIT AI model can capture the relationship between objects with the minimum amount of training data

Going back to the Turing Machine, this machine is basically a device with flexible computing capabilities. In other words, the Turing Machine is a method for activating classical mathematical types that are involved in performing complex calculations. In the case of Magic: The Gathering, researchers seem to have adapted and built this machine based on a previously completed construction by Alex Churchill. The nature of the Turing Machine engine based on this card is that by using a number of specific settings, you can assign it to perform a game-like function similar to a PC with the numbers 1 and 0.

Building Turing Machine based on Magic cards is really a reasonable strategy while ensuring compliance with the rules of the game.

'Despite the practical difficulties involved in setting up the exact state of the table, such as running out of space on a table, a persistent enough player can fully set up and execute this work at a real-world tournament, 'the researchers said.

The table below shows that a 60-card deck may trigger the construction of the Turing Machine as follows:

Magic: The Gathering - the most difficult 'chew' game for all AI models Picture 3

1. Google researchers for gaming AI to improve enhanced learning ability

However, this particular setting leads to a potential result (on a large scale) 'unsolvable'. That is it cannot be calculated correctly:

In addition to showing that the optimal strategy in Magic is impossible to calculate, this also shows that evaluating the definite consequences of Magic moves in the past is not easy. The complexity of optimal strategy play is still an open question, similar to many other aspects of Magic 'calculation.

Although it is clear that not all games in Magic: The Gathering result in an unintelligible result, it is in fact the only real-world game capable of surviving within of the rules.

So what does this mean for the field of game theory and artificial intelligence? The researchers' conclusions are as follows:

"Magic: The Gathering is completely inconsistent with the assumptions commonly made by computer scientists while modeling games. We affirm that creating an AI model can be pure play. Magic: The Gathering is not easy, however, the complexities on the other hand will also leave significant sources of data for future research. '

Magic: The Gathering - the most difficult 'chew' game for all AI models Picture 4

1. McAfee expert explained how deepfake and AI are drilling through the cyber security wall

After all, how to program an AI model that can solve the problem when there is no way to calculate the optimal strategy correctly? Magic: The Gathering is probably the place to hold the answer.