Robotics Reference
In-Depth Information
grad. SHRDLU could conduct a conversation in which it pretended to
be a robot manipulating a number of simple objects on a table top.
Person:
Pick up a big red block.
SHRDLU:
OK. (does it)
Person:
Grasp the pyramid.
SHRDLU:
I DON'T UNDERSTANDWHICH PYRA-
MID YOU MEAN.
Person:
Find a block which is taller than the one you
are holding and put it into the box.
SHRDLU:
BY “IT”, I ASSUME YOU MEAN THE
BLOCKWHICH IS TALLER THANTHE
ONE I AM HOLDING.
SHRDLU:
OK. (does it)
Person:
What does the box contain?
SHRDLU:
THE BLUE PYRAMID AND THE BLUE
BLOCK.
Person:
What is the pyramid supported by?
SHRDLU:
THE BOX.
Person:
How many blocks are not in the box?
SHRDLU:
FOUR OF THEM. [2]
SHRDLU's output was more natural than those of earlier systems and
the program also demonstrated some capability of reasoning. Winograd's
breakthrough had been to combine semantic and syntactic analysis, ex-
pressing the meanings of words and sentences as program code in a com-
puter language, in such a way that the execution of programs contain-
ing this code was analogous to a reasoning process based on the coded
meanings.
Even more impressive than Winograd's breakthrough was the LIFER
system developed by Gary Hendrix in 1977. It was designed as a natural
language interface to a database of information about ships in the U.S.
