AlphaZero: DeepMind's New Chess AI | Two Minute Papers #216

The paper "Mastering Chess and Shogi by Self-Play with a
General Reinforcement Learning Algorithm" is available here:

Our Patreon page with the details:

One-time payments:
Bitcoin: 13hhmJnLEzwXgmgJN7RB6bWVdT7WkrFAHh
Ethereum: 0x002BB163DfE89B7aD0712846F1a1E53ba6136b5A

Recommendations:

We would like to thank our generous Patreon supporters who make Two Minute Papers possible:
Andrew Melnychuk, Brian Gilman, Christian Ahlin, Christoph Jadanowski, Dave Rushton-Smith, Dennis Abts, Emmanuel, Eric Haddad, Esa Turkulainen, Evan Breznyik, Frank Goertzen, Kaben Gabriel Nanlohy, Malek Cellier, Marten Rauschenberg, Michael Albrecht, Michael Jensen, Michael Orenstein, Raul Araújo da Silva, Robin Graham, Steef, Steve Messina, Sunil Kim, Torsten Reil.

Credits:

Károly Zsolnai-Fehér's links:

[NicolaiCzempin]

For some reason people seem to latch onto the "only 4 TPUs used", both in AlphaZero and in AlphaGo Zero.

Please clarify that this is only for the fully trained network, i. e. just to get the next move out of a playing AlphaZero.

During training, according to the paper 5000 first-generation TPUs and 64 second-generation TPUs were used.

Karol did clarify that this cannot yet be done on "commodity hardware", but the way things are presented both here and elsewhere, the 4 TPU figure is what sticks in people's minds.

[fcarriedo]

Chills down the spine is the correct reaction.

[NeverInterpreter]

There's still debates going on that Stockfish was handicapped by not allowing it to use opening and endgame database. Also they say hardware they both ran at was incomparable. But certainly as an avid chess player I was really excited about this news. Unlike most of the engines the play of AlphaZero was way more human-like and easy to understand. Those ten games they published are super interesting and educational. Would love to see the other ones.

[MrTurbo_]

first we make games so we have something to do in our free time, then we make programs that play these games for us

[DamianReloaded]

Next time AlphaZero applied to stock market becomes the richest "person" in the world in 4 hours. ^_^

[francescomartella]

They could play tens of thousands of games, publish them and feed chess enthusiasts for the next 3.000 years

[bruceli]

1:05 the sudden jump in Elo Rating after a period of stagnation from 17hrs to 27hrs is interesting & scary!!! I wonder what caused the jump

[meegul]

>Two Minute Papers
>6 Minute video

Not complaining at all! Not worth rebranding the channel over such a small gripe, but I find it funny that as the channel has grown, the videos seem to get longer and longer. Again, I have no problem with this and enjoy some of the added depth and explanations (and would hope many videos are 5+ minutes in the future!), but it's a bit funny considering the name of the channel.

Love the work, Karoly, especially your course you made available on Rendering/Ray tracing from the university of Vienna. I'm not fully through it yet, but it's been a pleasure to work through in my free time.

[andrewxx]

They only released 10 of the games. We want the other 90!!!!

[capnrob]

Interesting thing about alpha zero it analyzes far fewer moves compared to stockfish, and still plays at such a high level.

[DerrickBest]

ChessNetwork's analysis it's a must watch!

[cyleyoakum]

Sir, I watch a lot of YouTube on a lot of different subjects. And I mean all the time. But your channel is my absolute favorite. I get psyched every time I see you have posted something. Thank you so much for doing this. Not only are you exciting and have fantastic info you keep it short and to the point. Please, keep up the good work. My family and I are cheering you on!

[mrfans]

Very interesting, thanks.  Nice to hear Jerry and Daniel getting a shoutout, I have enjoyed their analysis of the AlphaZero games, I'll check out the other guy now.  :-)

[julioandresgomez]

Alpha Zero progress in chess skyrocketted and then the learning curve flattened after a few hours. It hit some sort of ceiling, yet it still betters every once in a while. Like Bruce Lee said, there are no limits, only plateaus.

[miscelanea]

God! I love your videos. I use them as means to select my next paper reading as there are tons of really interesting knowledge. What you say is totally true: 'What a time to be alive' :D

[Thornstream]

It should be noted that this was Stockfish 8. There are now stronger versions released.

[satoshinakamoto]

so much to learn so little time ..and energy. i wish i never had to sleep

[thejaywalker]

What is weird is that Stockfish doesn't recommend some of the moves it made if you analyze it yourself.

[artman]

If it was to play dark chess (where you can only see the enemy pieces that can be captured), would it need even more generalized algorithm?

[musicalnostalgician]

Some points that most people are not aware of:
1. AlphaZero played against StockFish 8. At that time, StockFish 9 was already in the market.
2. AlphaZero was powered by a supercomputer (StockFish wasn't).
3. The 60 second rule is pure crap as the StockFish algorithm thinks accordingly at a specific moment of the game. If they played in the classical time format (90 min per side for the whole game), then StockFish might have won.