YGOrganization | [Master Duel] Upcoming Products and more on the AI project


I, for one, welcome our new A.I. overlords.

-New Structure Deck: Blue-Eyes Max
“A Deck that uses a tenacious “Blue-Eyes” monster as an ace-in-the-hole. Many “Blue-Eyes” support cards like “The White Stone of Ancients” and “Blue-Eyes Jet Dragon” have effects related to your Graveyard. Make sure to keep a close eye on the status on your Graveyard. Once preparations are set and you call upon powerful monsters like “Blue-Eyes Chaos MAX Dragon,” your complete and total victory is a stone’s throw away!

Included cards:
Blue-Eyes White Dragon (UR)
Blue-Eyes Alternative White Dragon (UR)
Blue-Eyes Chaos MAX Dragon (UR)

Available on February 7th (time zones applying)

-New Selection Pack: Ride into the Future
Included cards:
Diabellstar the Black Witch (UR)
Snake-Eyes Flamberge Dragon (UR)

Cosmic Quasar Dragon (UR)
T.G. Glaive Blaster (UR)

Emperor Charles the Great (UR)

Available on February 7th (time zones applying)

-New Selection Pack: Name of Champions
Included cards:
Lady Labrynth of the Silver Castle (UR)

Destiny HERO – Destroyer Phoenix Enforcer (UR)

Baronne de Fleur (UR)

Dingirsu, the Orcust of the Evening Star (UR)

Accesscode Talker (UR)

Available on February 7th (time zones applying)

Furthermore, regarding the A.I. system, a short presentation was livestreamed during day 2 of the event (starts at around 47 minutes).

Ken’ichi Kataoka (Master Duel/Duel Links Senior Producer)
Mr. Hattori (Programmer)

Q: Can you give us a simple explanation as to what this A.I. project is about?
Mr. Hattori: This project is a system that utilizes Python for Master Duel. Simply put, by having knowledge of A.I. and the Yu-Gi-Oh! OCG, anyone can create a Duel A.I. that can play Master Duel.
Q: “Anyone” meaning, even someone outside Konami?
Mr. Hattori: Yes, that is correct. First, we would like to take a look at a Duel between two A.I.s created by [one or more people] outside of the Master Duel production team.

The graphic on the screen basically explains that a special interface is created for the A.I. to directly interact with the game (top picture).
The text in the middle of both picture explains that the operations from the A.I. and the probability evaluations and more create a feedback system.
The bottom picture explains that it is also possible to use a console simulator to run Duels at high speed with a text-only interface.

The two squares on the right basically explain that information from the Duel and board is organized and distributed to the A.I., which then responds with its inputs and requests. The A.I. agent (written in Python) is an A.I. framework / trained model.

Q: This looks very different from the usual screen that we see in Master Duel, could you give us a rundown of what we’re looking at?
Mr. Hattori: Yes. The first thing I would like to point out is at the bottom-left, we see a percentage. This percentage is the winning odds from the perspective of the moves of the A.I. In this bottom-left box, we can see the total estimated winning odds, but in the bottom right, we can see the top 5 moves. So we’ll be looking at the A.I. choosing the #1 move. Right now, the winning odds are low, but soon enough, there will be a move that heavily increases the winning odds, so let’s pay close attention to that.

(The A.I. uses The Melody of Awakening Dragon to search for Blue-Eyes Chaos MAX Dragon and is now able to Ritual Summon it, increasing the winning odds to about 80%)

Q: From my perspective, this almost looked like two CPUs Dueling each other, but it’s not just that, is it?
Mr. Hattori: Correct, these are not CPUs, but rather external programs operating Master Duel. In other words, machine learning is being used to teach the A.I.. The data of the users is not being utilized at all, and the A.I. is learning from scratch how to Duel by Dueling over and over, without even teaching it the rules.
Q: By “over and over”, around how many times would you say?
Mr. Hattori: It has trained Dueling the Solo Mode CPU over 100000 times.
Q: So even without teaching it the rules, you just make it Duel over 100000 times, and the result is this current A.I.?
Mr. Hattori: Yes.
“It has probably played more Duels than humans!”, “That’s some though training!”

Q: There is so much information on the screen, that looks so cool!
Mr. Hattori: Yes, we thought it would be interesting for whoever is watching to display what the A.I. was thinking on the screen.

Ken’ichi: We are game developers, however, we are not A.I. specialists. That’s why we thought it would be interesting to see what kind of A.I. could be created by a company that specializes in this field, and what kind of Duels would be like between these A.I.s, and that’s why we created this project. “Who can create the strongest A.I.?” Of course, it’s also possible for A.I. to Duel human players, so we would also like to see what kind of Duels would happen between “the world’s #1 A.I.” and “the champion of the World Championship”.
Q: So an outsider, someone not from Konami, can participate in this as well?
Ken’ichi: Of course, this is not a system that we will just give away for everyone to create their own A.I. At this moment, [protocols/rules/terms of service] are being created, and we are currently moving forward so that within this year, positions for A.I. Duel creators will be posted at the official Master Duel website.
Mr. Hattori: First, we would expect the participation of groups/organizations with A.I. production skills. If possible, competitions between universities or such.
“That would be so cool, with each university teaching their own A.I. and then pitting it against other universities!”
Ken’inchi: Yes, the A.I.s created by universities are very interesting.
“Yes, they have their unique quirks according to who makes them”

Ken’ichi: So far, this conversation has been mostly for A.I. enthusiasts, but there will also be some benefits for all Duelists.
Mr. Hattori: We’ve shown this before in the previous video, but please look at this:

Mr. Hattori: By putting to use the winning odds estimation function of the A.I., we can create something similar to the well known advantage meter from go/shogi.
Q: I see, so as it can be seen in this image, the turn player is in an advantageous position, and that is what is being displayed, right?
Mr. Hattori: That is correct.
“Right, so in just one glance, one can easily tell that the player holding Mystical Space Typhoon is favored to win!”
“For us that are commenting on tournaments, you can just look at the screen and figure out who’s winning”
“[…]and you might also have a case where someone has like 5% chance of winning, then they draw a new card and boom! They are suddenly able to reverse the odds!”

Q: Is there something else?
Mr. Hattori: Yes, we are also excited about a function that lets you predict what face-down cards could be. If we could look at the next slide…

Mr. Hattori: Here we can see the A.I. trying to predict what the face-down card could be. [Vaguely explains that the prediction could be off, the A.I. could hold back, move aggressively, etc.]“So the A.I. is thinking: That face-down card might just be Ring of Destruction, right”
“It probably got hit by RoD way too many times”
“Whom was it Dueling with those 100000 times?”
“That’s how the A.I. develops its quirks, depending on who it Duels”
“It almost feels like the world of VRAINS, where you can research someone’s Dueling data”

Ken’ichi: The concept of Master Duel is that both those participating and those watching can enjoy the Duels. That’s why we are also trying to improve “what is being shown”. We believe that conveying in the simplest way possible the best parts of Master Duel to those in the audience is a fundamental part of increasing the pool of competing players. That is why we keep evolving Master Duel for Duelists, and those cheering on those Duelists. Stay tuned for the future of Yu-Gi-Oh digital games!

A live demonstration then begins. Both Ken’ichi and Mr. Hattori quickly mention that while the A.I. has Dueled over 100000 times, this is still considered “young” and “underdeveloped” in terms of A.I., so it may still make some unexpected moves. Mr. Hattori highlights that when the A.I. is about ot attack and the opponent controls a face-down card, it’s making a decision based on potential face-down cards, and once it decides that it’s not a threat (Eternal Soul instead of Mirror Force), it proceeds with the attack. He also highlights that the A.I. grasps some understanding of the opponent’s Deck in order to make such predictions. Mr. Hattori is asked if it’s possible to teach the A.I. about the existence of certain Decks so that it learns how to deal with them, for example, taking a currently strong Deck, like “Horus”, and teaching it to the A.I., and Mr. Hattori responds it is possible to teach the A.I. that way. The A.I. uses Ring of Destruction during the End Phase, and when Mr. Hattori is asked about whether this particular use of the card was learned by the A.I., he explains that through thousands of Duels, the A.I. has a high valuation of moves that lead to victory. While the A.I. does not know the rules of Yu-Gi-Oh, it understands that this particular move leads to positive results.

(The Duel unfortunately ends prematurely due to time constrains)

Related articles

Recent articles