anneteoh wrote:beenthere wrote:@Asian_Geekette
- Yes, our focus is on English language texts. Bur much of these text-analytics algorithms are based on statistics and probabilities and could be used to build similar systems on other language. As long as these languages have a grammar and large collections of texts to run analyses - the adaptation would be a few PhD theses away.
@AnneTeoh - These days overlord is gender-neutral; over to you lady.
@nakatago - you shall one day be part of the matrix.
Cheers, beenthere, though some feminist linguists will argue that gender bias runs deep in the structure of our language, positioning females as always, inferior to men; or simply overruling them.
Grammar is easiest to mechanise though there are grey areas too. But hey, how about this - Green sleep dreams furiously ( Noam Chomsky)?
There're many important areas that AI will find diff. to access e.g. accent, stress, semantics , intentions and as mentioned elsewhere, cross-cultural pragmatics.
Recently, the police in the UK were able to decode some terrorists' email texts, where party was meant to be mass destruction of the enemies and nikah, or wedding ceremony meant as detonating the bomb. This timely interference saved many innocent lives. There ought to be more research
and work in this electronic language area.
Sorry, what thread's this. Seems like we're all over the place, beenthere.