Logo

How AI impacts chess and the game of World Chess Champion Magnus Carlsen - Shared screen with speaker view
Chat and Q&A
01:00:28
Thank you for joining the AI for Good webinar on “How AI impacts chess and the game of World Chess Champion Magnus Carlsen”! We hope you and your families are healthy and safe. Please feel free to use the chat function to message "all panelists" or "all panelists and attendees" and ask questions in the Q&A tab.
Kseniia Fontaine
01:01:47
Geneva
Karthik Raj C
01:01:57
India , Chennai
Ramon Lopez de Mantaras
01:02:01
Barcelona
Loredana Manushaqa
01:02:06
Dubai here
Jiří Mohl
01:02:09
Czech republic
Iurii Marinskii
01:02:12
Moscow, Russia
VARDAAN MONGIA
01:02:13
india, namastey
Roelof Reinders
01:02:13
Amsterdam
Sanjay Kumar Mishra
01:02:26
Hi I am from India
Jiří Mohl
01:03:33
So I think people not playing chess online are those who heard about AlphaZero but not about Stockfish
Bernard Ancel
01:04:10
I am in the Valais mountains !
Chat and Q&A
01:05:25
We are pleased to have with us Peter Heine Nielsen (Chess Grandmaster; Coach of World Class Champion Magnus Carlsen) and Sebastian J. Kuhnert (CEO of chess24; CEO of ChessX; COO of Play Magnus AS).We are grateful to have Kenneth Cukier moderate the session. He is a Senior Editor at The Economist and a New York Times bestselling author.
Mark Bryan
01:06:08
Hi everyone :) Greetings from Cyprus!
Chat and Q&A
01:09:56
Please feel free to submit your brief questions in the Q&A tab
Ramon Lopez de Mantaras
01:09:56
Deep Mind algorithm is actually not general. It needs to be “tuned” (the least to say) to play each different game
Ramon Lopez de Mantaras
01:10:31
AI is decades far from general AI
Ramon Lopez de Mantaras
01:14:08
Deep Learning “suffers” from the well known problem of catastrophic forgetting. Which means that once trained to solve a task, if they are trained to deal with another task, even if related, it “forgets” completely the previous task. It needs to be retrained again
Ramon Lopez de Mantaras
01:14:51
The solution is called “transfer learning”, but this is still a huge open problem in AI
Mario Fromow
01:15:40
Hello all, greetings from Mexico
sanjai dhoni
01:16:24
Leela chess uses the same ideaology as of alpha zero(AI).But they lack in training the hidden weights more accurate and thus policy making ie Q value is not very accurate for decision making/reinforcement learning.
Chat and Q&A
01:16:51
Thank you for your question. Please submit it in the Q&A tab.
Vladimir Potkin
01:18:53
Is game of chess , where we can evaluate the AI decision and AI can evaluate the human's one. Helps to make AI and human be closer, in future terms?
Chat and Q&A
01:19:03
Thank you for your question. Please submit it in the Q&A tab.
Chat and Q&A
01:19:14
Thank you for your question. Please submit it in the Q&A tab.
VARDAAN MONGIA
01:19:48
but wouldn't that create a bias? since Magnus has learned Kasparov's games while becoming a grandmaster?
Chat and Q&A
01:19:50
Please feel free to submit your brief questions in the Q&A tab
sanjai dhoni
01:21:42
we should focus on making AI helping 100M chess players smarter in this stage rather than trying to go hard on achieving AGI .
Sebastian Kuhnert
01:22:51
we are doing this as well @sanjai
Sebastian Kuhnert
01:24:14
check e.g. www.chessable.com, www.playmagnus.com and Tactics Frenzy, further ones coming up as well
Chat and Q&A
01:25:12
Please note that you can send chat either to "all panelists" or "all panelists and attendees"
Ramon Lopez de Mantaras
01:25:49
When it comes to understanding the games played by AlphaZero, the program’s inability to explain its games and the knowledge acquired in human-understandable terms is a serious limitation.
sanjai dhoni
01:28:01
It can be done by decoding how Magnus, Fabiano or any other top player uses their mental model. It's an inverse approach. But well may help others to understand their games in more depth and better.
Sanjay Kumar Mishra
01:28:43
I agree with Sanjai Dhoni
Ramon Lopez de Mantaras
01:29:58
Magnus can explain, Alphazero cannot
Ramon Lopez de Mantaras
01:30:49
The inverse approach does not work with deep learning.
Sravankumar Kota
01:30:56
This is only for Sebastian.Though, your academic background should be different how you have managed to initiate and run couple of companies simultaneously.
Ramya Devasahayam
01:31:22
Please tell Magnus to speak
Sebastian Kuhnert
01:31:59
Magnus is playing soon
Ramya Devasahayam
01:32:28
i
Ramya Devasahayam
01:33:07
Is Magnus playing any game today
Ramon Lopez de Mantaras
01:34:56
The problem of AI is that everyone anthropomorphises it. Artificial Intelligence has very little to do with human intelligence.
Ramya Devasahayam
01:35:05
Can you tell it again
Sebastian Kuhnert
01:35:34
https://chess24.com/en/watch/live-tournaments/lindores-abbey-rapid-challenge-2020-final-8/1/1/5
Sebastian Kuhnert
01:35:37
playing today
Ramon Lopez de Mantaras
01:38:08
As Daniel Dennet says in his book “From Bacteria to Bach and Back”, AI is “competence without comprehension”
Ramya Devasahayam
01:38:20
Is Magnus learing m
Sravankumar Kota
01:39:30
why the people are emphasizing AI in only chess game among umpteen games are available?
Ramya Devasahayam
01:40:09
Is Magnus learning now from Peter Heine Nielsen
Juhani Kuronen
01:41:48
Lc0 fans are furious. Don't ever think no emotions are attached in engine chess :)
Sebastian Kuhnert
01:42:01
:)
Sebastian Kuhnert
01:42:05
understood, lesson taken
Sanjay Kumar Mishra
01:42:17
:)
sanjai dhoni
01:42:29
Anything AI is creating should be either gamified or it should create an environment of empathy and interest to players.
Sebastian Kuhnert
01:43:15
you can check chessable.com and playmagnus.com that take different roads towards this
Webinar Attendee
01:43:27
ook
Ramon Lopez de Mantaras
01:47:53
When it comes to understanding the games played by AlphaZero, the program’s inability to explain its games and the knowledge acquired in human-understandable terms is a serious limitation.
Jose Mainez Bosch
01:48:10
The use of emotions is in itself is a kind of a game, a potential strategy... all belong to AI in the broader scope. How an ethical approach to AI ought to allow its use, might be the issue...
Ramon Lopez de Mantaras
01:48:24
I recommend the paper: AlphaZero – What’s Missing?by Ivan Bratko. Published in the Informatica Journal 42, 2018.
Juhani Kuronen
01:48:38
Different brain chemistry
Juhani Kuronen
01:53:31
One problem is the female only tournaments in my opnion as that will result them playing the weaker field. You get better by playing the best, not by playing your equals.
Sravankumar Kota
01:54:03
why the people are emphasizing and focusing AI in only chess game among umpteen games are available similarly?
Webinar Attendee
01:56:38
You 06:25 PMcan we develope any dynamically implementable algorithm rather than using trained models to make chess bot!?
Juhani Kuronen
01:59:21
But AI playing StarCraft is kind of "stupid" because human's are limited by hand speed. You cannot "micro" several places same time as a human but AI can do it. It's not fair comparison.
Webinar Attendee
02:00:12
right Juhani
Ramon Lopez de Mantaras
02:01:27
Yes, Juhani is right
Chat and Q&A
02:02:15
Thank you everyone for participating in this webinar. Feel free to stay informed with upcoming webinars and AI for Good Global Summit activities https://aiforgood.itu.int/
Kenneth Cukier
02:02:15
Thank you to the panelists!!!
Mark Bryan
02:02:18
Thanks :)
Kenneth Cukier
02:02:20
Thank you Sebastian
Sebastian Kuhnert
02:02:20
Got to run, thank you!
Kenneth Cukier
02:02:24
And Peter Heiene
Sebastian Kuhnert
02:02:29
Thank you all, pleasure!
Roelof Reinders
02:02:32
Thanks!
Sanjay Kumar Mishra
02:02:35
Thank..!!
Webinar Attendee
02:02:38
thx
Webinar Attendee
02:02:56
can you share whatsapp group link
Paola Morera
02:03:13
Very interesting! thank you!
Webinar Attendee
02:04:12
lol am here also
Saurabh Hirugade
02:04:41
Is it over??
Juhani Kuronen
02:05:03
Yes