Re: Shanahan: Robotics and Common Sense

From: Button David (drb198@ecs.soton.ac.uk)
Date: Thu May 24 2001 - 11:22:46 BST


>Sloss:
>In this particular situation I believe that the intelligence shown is
>not true intelligence. While the robot can navigate its world, it
>still cant decide why it wants to travel to the particular location.
>This theory will be very useful in fulfilling part of the soft AI goal,
>namely useful devices that appear to reason about their tasks.

Button:
While I personally agree with this comment, the robots 'learning'
capabilities must be acknowledged. As with the comment made by Sloss,
this sort of theory would be very useful where the robot is not used as
a candidate for 'intelligence' it would be useful in many areas defined
by weak AI.

The idea that this is not 'true intelligence' is interesting. What does
Sloss mean by this. Surely the robot in some ways can be considered as
thinking (I personally don't think the robot's thinking), but it still
has the ability to make decisions based on it's environment (simple, but
still decisions). Also, surely the idea of the robot is still mainly
computation. I agree that the robot has sensorimotor capabilities and
therefore, in furtherance of Searle's Chinese Room Argument, the robot
has some means of grounding real world data. However, beyond this the
robot is simply following rules and inessence an algorithm.

>Sloss:
>The addition of a range sensor would allow the robot to avoid actually
>crashing into the object before it does any damage, and a camera could
>allow the robot to build up the shapes of objects more quickly.

Button:
This is an interesting suggestion. However, the introduction of these
'devices' to aid the robot would surely cause additional problems. In
terms of the range sensor, would a soft object, or moveable object be
sensed by it? Wouldn't this mean that a robot must collide to evaluate
any objects? In reference to the camera, this just brings in some more
problems. In particular the knowledge of shapes. The robot would
therefore need a knowledge base consisting of shape data. This shape
recognition system would also have to be able to learn - the same type
of object may be of different shape.

David Button - drb198@ecs.soton.ac.uk



This archive was generated by hypermail 2.1.4 : Tue Sep 24 2002 - 18:37:30 BST