h a l f b a k e r yReformatted to fit your screen.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
"The bass line is too weak" means something different in
a
studio from in a boat, as does "feed me" in a restaurant
from
a newsroom and the verb "wipe" in a nursery from a
skate
park. Until the environmental context is known this
represents a hard limit to language disambiguation.
I
propose a companion application to natural
language
receiving AIs that would perform thusly:
It would
provide a constant feed of what type of environment it's
in
and the objects and behaviors of the people around it.
Recognizing a mug it would output "mug". Seeing a man
carrying a box it would output "moving things". Seeing
people shouting with stress in their voices it would
output
"emergency". This feed could be taken in by the
language
receiving AI and help disambiguate whatever is said to
it.
[link]
|
|
It won't always work, I fear. My missus once asked for a corned beef sandwich 'with onion'. 'Funny' I thought, 'maybe it's a local delicacy round her way', and obligingly made her a sandwich containing corned beef with honey on. Not unpleasant, she maintained, but not quite what had been specified. |
|
| |