On Sat, May 11, 2019 at 12:10 PM Stefan Reich via AGI 
<[email protected]<mailto:[email protected]>> wrote:
I think you are stuck in the "conventional theory". The time is right for new 
theories. I'm still on the path to show that there are better ways to actual AI 
than neural networks.

Agreed

________________________________
From: Stefan Reich via AGI <[email protected]>
Sent: Saturday, 01 June 2019 18:42
To: AGI
Subject: Re: [agi] My AGI 2019 paper draft

Thanks for believing that I can do this.

On Sat, May 11, 2019, 18:27 Matt Mahoney 
<[email protected]<mailto:[email protected]>> wrote:
"Conventional" AI was coding all the knowledge in Lisp. Then came rule based 
systems and the AI winter. Neural networks have produced the best results in 
language, vision, and robotics, now that we have enough computing power to 
implement them. Now we have real progress in AI. It will be interesting if you 
come up with something better.

On Sat, May 11, 2019 at 12:10 PM Stefan Reich via AGI 
<[email protected]<mailto:[email protected]>> wrote:
I think you are stuck in the "conventional theory". The time is right for new 
theories. I'm still on the path to show that there are better ways to actual AI 
than neural networks.

On Sat, May 11, 2019, 08:46 YKY (Yan King Yin, 甄景贤) 
<[email protected]<mailto:[email protected]>> wrote:
Also, the control theoretic stuff was removed because I am unable to define the 
reward based on the current state in a differentiable way.  For example, in the 
game of chess, the reward comes only when checkmate occurs (according to the 
game's official rules), but not when you capture a piece of high value (eg  
xQueen).  This problem is known as "sparse reward" vs "dense reward" in 
reinforcement learning:
[Screenshot from 2019-05-07 20-49-22.png]
The actual reward is a delta-function occurring at the end of the game.  
"Classical" control theory is applicable only when the reward is something like 
the dotted line.



--
-- Matt Mahoney, [email protected]<mailto:[email protected]>
Artificial General Intelligence List<https://agi.topicbox.com/latest> / AGI / 
see discussions<https://agi.topicbox.com/groups/agi> + 
participants<https://agi.topicbox.com/groups/agi/members> + delivery 
options<https://agi.topicbox.com/groups/agi/subscription> 
Permalink<https://agi.topicbox.com/groups/agi/T3cad55ae5144b323-M2018f8e48159c8e72dbb1c31>

------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T3cad55ae5144b323-Mf619664bcac5a025737a9e47
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to