|
||
 |
Viczy says | |||
| In the context of the Sapir-Whorf hypothesis, would it make sense to write some "rules of grammar" for some sort of Turing machine, according to which it would interpret its environment, rather than defining how an AI would interpret its surroundings in a behaviourist manner? Just an idea. |
||||
| Total Topic Karma: 1 | - More by this Author | |||
 |
p0ss says |
|
||||
| a link on the Sapir-Whorf hypothesis would have been nice, for future reference, click on (formatting text) to find out how to insert links, equations and alter text styles. | ||||||
| - 14 February, 2007 | ||||||
|
p0ss says |
|
||||
as for your question, a Turing machine would need to be programmed with the rules of grammar that apply to the language it was communicating in, if it wished to communicate fluently in multiple languages it would need to be programmed with the relevent culturally specific linguistic modifiers, otherwise it would sound like a foreigner  A current attempt at solving this problem is by building AI's that are just instructions on how to learn things, not fully programmed turing machines. This method allows the AI to learn via interaction with humans, thereby giving it the "rules of grammar" you speak of. an examples of this methodology can be found here. |
||||||
| - 14 February, 2007 | ||||||
|
Viczy says |
|
||||
| Awesome; thanks for the comment, the link tip was helpful. | ||||||
| - 15 February, 2007 | ||||||
|
|
