On 7/7/13 3:07 AM, Walter Bright wrote:
On 7/7/2013 1:30 AM, Andrei Alexandrescu wrote:
On 7/7/13 1:26 AM, Walter Bright wrote:
On 7/6/2013 11:11 PM, TommiT wrote:
I can see machine translation that is based on statistical
correlation with a
sufficiently large corpus of human translations, but I don't see much
hope for
actual understanding of non-literal speech in the foreseeable future,
and I'm
actually rather glad of that.
You haven't read Ray Kurzweil's latest books then or you just don't
think he's
right?
Spend a little quality time with Siri. I did, and discovered it was
hardly any better than Eliza, which is a few lines of BASIC written in
the 1970's.
Ow come on.
All Siri does is recognize a set of stock patterns, just like Eliza.
Step out of that, even slightly, and it reverts to a default, again,
just like Eliza.
Of course, Siri had a much larger set of patterns it recognized, but
with a bit of experimentation you quickly figure out what those stock
patterns are. There's nothing resembling human understanding there.
But that applies to humans, too - they just have a much larger set of
patterns they recognize. But they don't overlap perfectly for all
humans. Try to ask your mailman whether a hash table is better than a
singly-linked list for a symbol table.
Andrei
