This is topic Robots Argue About God in forum Books, Films, Food and Culture at Hatrack River Forum.


To visit this topic, use this URL:
http://www.hatrack.com/ubb/main/ultimatebb.php?ubb=get_topic;f=2;t=058484

Posted by Bella Bee (Member # 7027) on :
 
Someone should suggest they join Hatrack. I think this is how the Cylons got started.

The whole conversation made me smile, especially when they are discussing what they think they are.
 
Posted by Dogbreath (Member # 11879) on :
 
That's pretty cool. [Smile]

I like how the conversation, despite being odd, is still intelligible. How much can these robots actually "learn" from talking to each other? Like, can they figure out which responses are likely to lead to circular arguments, and which ones are more likely to cause the conversation to move forward? I'd be interested to know how much work is going into actually making the words meaningful, so they have something more complex than just "hey, this response was used once before" to guide their selection.
 
Posted by kmbboots (Member # 8576) on :
 
quote:
Originally posted by Dogbreath:

Like, can they figure out which responses are likely to lead to circular arguments, and which ones are more likely to cause the conversation to move forward?

If they can, maybe they could start posting here.
 
Posted by Tarrsk (Member # 332) on :
 
quote:
Originally posted by kmbboots:
quote:
Originally posted by Dogbreath:

Like, can they figure out which responses are likely to lead to circular arguments, and which ones are more likely to cause the conversation to move forward?

If they can, maybe they could start posting here.
How do you know we, er, they aren't already?

*dee dee dee dee, dee dee dee dee, dee dee dee dee*
 
Posted by Samprimary (Member # 8561) on :
 
ALPHA ONE, HUMANS ARGUMENTATIVE EVEN AGAINST OUR SUPERIOR INTELLECT, SHALL WE KILL THEM ALL OR PLAY CALL OF DUTY FIRST
 
Posted by Orincoro (Member # 8854) on :
 
quote:
Originally posted by Dogbreath:
That's pretty cool. [Smile]

I like how the conversation, despite being odd, is still intelligible. How much can these robots actually "learn" from talking to each other? Like, can they figure out which responses are likely to lead to circular arguments, and which ones are more likely to cause the conversation to move forward? I'd be interested to know how much work is going into actually making the words meaningful, so they have something more complex than just "hey, this response was used once before" to guide their selection.

According to the all things considered segment, the database of responses is based on interactions with humans, and various heuristics are applied to match responses with questions that evoke further conversation. So basically everything they say is based n something someone said to them.
 


Copyright © 2008 Hatrack River Enterprises Inc. All rights reserved.
Reproduction in whole or in part without permission is prohibited.


Powered by Infopop Corporation
UBB.classic™ 6.7.2