Artificial Intelligence
1. Introduction and Scope
Artificial intelligence is about algorithms enabled by constraints exposed by representations
that model targeted thinking, perception, and action. mit's full of mit 's full of them. I 'd like to
give you one more example. it 's something we call, in artificial intelligence, generated test.
and it 's such a simple idea, you need to add to your repertoire of problem solving methods.
the Rumpelstiltskin principle says that once you can name something, you get power over it.
once you have a name for something you can start talking about. and that vocabulary gives
you power. so I can say that if you 're doing a generate and test approach to a problem,, you
better build a generator with certain properties that make generators good.. For example,,
they should not be redundant. they should not give you the same solution twice. and they
should be informable.. They should be informable?. Mit Professor says simple ideas in
artificial intelligence are often the most powerful.. He says that simple ideas can be
powerful, and trivial makes it sound like it's not only simple, but of little worth.. Mit people
miss opportunities because they have a tendency to think that ideas are important unless
they 're complicated..
If we 're engineers, it 's for building smarter programs. if you 're a scientist, there's a
somewhat different motivation., but most this subject is going to be about the other part that
makes it possible for you to build smart programs. some of it will be about what it is that
makes us different from chimpanzees. on Monday, we 're going to talk about this program.
and you can write one yourself. in one day discussing it will be in itself a miniature artificial
intelligence course. so that 's the Dawn age, early dawn age. this was the age of speculation.
in the late dawn age, we began to turn our attention from purely symbolic reasoning to
thinking a little bit about perceptual apparatus.. programs were written that could figure out
the nature of shapes and forms.. forms. the most important thing, perhaps, was what you
look at with me on Wednesday Next. It 's a rulebased expert systems..
There is a question of what age we 're in right now. I like to call at the age of the right way.
and this is an age when we begin to realize that that definition up there is actually
incomplete. much of our intelligence has to do not with thinking, perception, and action
, acting separately, but with loops that tie all those together. the high school idea is that we
evolved through slow, gradual, and continuous improvement. but that does n't seem to be the
way it happened. we humans have been around for maybe 200,000 years. In our present
anatomical form. we probably necked down as a species to a few thousand or maybe even a
few hundred individuals, something which made these accidental changes more capable of
sticking. we 're never going to understand human intelligence until we can understand that.
we think with our eyes.. Eyes. descriptions enable us to tell stories. storytelling enables us to
marshal the resources of our perceptual systems. and that's where we 're going to finish the
subject the semester by trying to understand more about that phenomenon.. mit is about
two things. it 's about skill building and it's about big ideas..
2. Reasoning: Goal Trees and Problem Solving
Today we 're going to be modeling a little bit of human problem solving, the kind that is
required when you do symbolic integration. the kind of problem solving. You 'll see today is
like generating tests,, which you saw last time. this is what a freshman would do when they
see a problem like that. we 'd like to find a way to make it into another problem that's more
likely to be found in the table. we 're going to take the problem we 're given, and convert it
into a problem that 's simpler. and we're going to give that process and name, and we're
calling it problem reduction.. Some of these transformations are extremely simple and
always safe., but some of them are safe. and I wonder if somebody could volunteer a simple
transformation that always is a good thing to do. we 're. We 're going to divide our
transformations into those two categories. so I need another safe one. the architects are
sitting over there. divided not only by nationality, but by course. there are more than this, this
is a sample. and these are the ones we're going to need in order to solve that problem.
This program, by the Way, is a Dawnage program.. It was written by a nearly blind. student by
the name of James Slagle. In 1960.. Slagle was able to write a program that did extremely
well when benchmarked against freshmen.. He anticipated so much of the subsequent 20
years that talking about his program is a miniature introduction to the whole field. [UNK], a
method that often works is n't guaranteed to work. heuristic transformations are sometimes
useful, not always useful. but you ca n't get an a in calculus without knowing some of them..
There are others that we need to have in our repertoire. In order to solve the problem. there 's
1. Introduction and Scope
Artificial intelligence is about algorithms enabled by constraints exposed by representations
that model targeted thinking, perception, and action. mit's full of mit 's full of them. I 'd like to
give you one more example. it 's something we call, in artificial intelligence, generated test.
and it 's such a simple idea, you need to add to your repertoire of problem solving methods.
the Rumpelstiltskin principle says that once you can name something, you get power over it.
once you have a name for something you can start talking about. and that vocabulary gives
you power. so I can say that if you 're doing a generate and test approach to a problem,, you
better build a generator with certain properties that make generators good.. For example,,
they should not be redundant. they should not give you the same solution twice. and they
should be informable.. They should be informable?. Mit Professor says simple ideas in
artificial intelligence are often the most powerful.. He says that simple ideas can be
powerful, and trivial makes it sound like it's not only simple, but of little worth.. Mit people
miss opportunities because they have a tendency to think that ideas are important unless
they 're complicated..
If we 're engineers, it 's for building smarter programs. if you 're a scientist, there's a
somewhat different motivation., but most this subject is going to be about the other part that
makes it possible for you to build smart programs. some of it will be about what it is that
makes us different from chimpanzees. on Monday, we 're going to talk about this program.
and you can write one yourself. in one day discussing it will be in itself a miniature artificial
intelligence course. so that 's the Dawn age, early dawn age. this was the age of speculation.
in the late dawn age, we began to turn our attention from purely symbolic reasoning to
thinking a little bit about perceptual apparatus.. programs were written that could figure out
the nature of shapes and forms.. forms. the most important thing, perhaps, was what you
look at with me on Wednesday Next. It 's a rulebased expert systems..
There is a question of what age we 're in right now. I like to call at the age of the right way.
and this is an age when we begin to realize that that definition up there is actually
incomplete. much of our intelligence has to do not with thinking, perception, and action
, acting separately, but with loops that tie all those together. the high school idea is that we
evolved through slow, gradual, and continuous improvement. but that does n't seem to be the
way it happened. we humans have been around for maybe 200,000 years. In our present
anatomical form. we probably necked down as a species to a few thousand or maybe even a
few hundred individuals, something which made these accidental changes more capable of
sticking. we 're never going to understand human intelligence until we can understand that.
we think with our eyes.. Eyes. descriptions enable us to tell stories. storytelling enables us to
marshal the resources of our perceptual systems. and that's where we 're going to finish the
subject the semester by trying to understand more about that phenomenon.. mit is about
two things. it 's about skill building and it's about big ideas..
2. Reasoning: Goal Trees and Problem Solving
Today we 're going to be modeling a little bit of human problem solving, the kind that is
required when you do symbolic integration. the kind of problem solving. You 'll see today is
like generating tests,, which you saw last time. this is what a freshman would do when they
see a problem like that. we 'd like to find a way to make it into another problem that's more
likely to be found in the table. we 're going to take the problem we 're given, and convert it
into a problem that 's simpler. and we're going to give that process and name, and we're
calling it problem reduction.. Some of these transformations are extremely simple and
always safe., but some of them are safe. and I wonder if somebody could volunteer a simple
transformation that always is a good thing to do. we 're. We 're going to divide our
transformations into those two categories. so I need another safe one. the architects are
sitting over there. divided not only by nationality, but by course. there are more than this, this
is a sample. and these are the ones we're going to need in order to solve that problem.
This program, by the Way, is a Dawnage program.. It was written by a nearly blind. student by
the name of James Slagle. In 1960.. Slagle was able to write a program that did extremely
well when benchmarked against freshmen.. He anticipated so much of the subsequent 20
years that talking about his program is a miniature introduction to the whole field. [UNK], a
method that often works is n't guaranteed to work. heuristic transformations are sometimes
useful, not always useful. but you ca n't get an a in calculus without knowing some of them..
There are others that we need to have in our repertoire. In order to solve the problem. there 's
