[FRIAM] AI advance

[Steven A Smith]

Fascinating!   I remember the broad discussions at the Cellular Automata 
Conference here in 1984 on the challenges/opportunities of using a CA to 
play GO.

I had an (unpublished of course) variation on Bill Gosper's HashLife 
<https://en.wikipedia.org/wiki/Hashlife> which I hoped might be a good 
basis for a  winning GO system back in those Pre Artificial Life days.

MIne used a less optimal subdivision (he did quad-tree, I used N-1 
Patches).  The purpose was to make the memoization translation invariant 
at all scales, not just binary orders of magnitude.   I was interested 
in general in the problem of using the hash to help analyze the 
computational complexity of a problem under solution based on the growth 
of the hash table.

Through my colleague, Susan Stepney (who some of you know) in York, I 
encouraged her grad student Jenny Owen to take this somewhere.  Alas, 
she chose to work with the Gosper version which I still believe has the 
unfortunate artifact of quad-tree/binary subdivision of the space, 
missing *many* repeated patterns at scales and offsets not aligning with 
the quad-tree.

Now we just need to teach it to play a mean game of "Go back to your 
Golden Towers" in DC?

- Steve


On 1/28/17 7:31 AM, Joseph Spinden wrote:
> Of interest to some:
>
> https://www.wired.com/2016/01/in-a-huge-breakthrough-googles-ai-beats-a-top-player-at-the-game-of-go 
>
>
> -JS
>
>
>
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
> FRIAM-COMIC http://friam-comic.blogspot.com/ by Dr. Strangelove

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20170128/db0dac75/attachment-0002.html>