video_id
stringclasses 7
values | text
stringlengths 2
29.3k
|
|---|---|
5DUQ3-Y_gX4
|
I'm not asking for df.
|
5DUQ3-Y_gX4
|
Of course, they're related.
|
5DUQ3-Y_gX4
|
And let's see if we can
move this out of the way.
|
5DUQ3-Y_gX4
|
Good.
|
5DUQ3-Y_gX4
|
So what was I saying?
|
5DUQ3-Y_gX4
|
Yes.
|
5DUQ3-Y_gX4
|
So now, what I want
to do is basically use
|
5DUQ3-Y_gX4
|
this as the definition of a
derivative, a more general
|
5DUQ3-Y_gX4
|
definition of a derivative.
|
5DUQ3-Y_gX4
|
So the linear algebra
notion is that we have
|
5DUQ3-Y_gX4
|
what's called a linear operator.
|
5DUQ3-Y_gX4
|
So basically, the
change in the output df
|
5DUQ3-Y_gX4
|
is going to be a
linear operator.
|
5DUQ3-Y_gX4
|
Let me write that
as times, which
|
5DUQ3-Y_gX4
|
I'm going to call f prime of x
times the change in the input.
|
5DUQ3-Y_gX4
|
This is our-- whoops.
|
5DUQ3-Y_gX4
|
Actually, but my input
is in red, right?
|
5DUQ3-Y_gX4
|
That's my color code.
|
5DUQ3-Y_gX4
|
My color scheme is input
is red and output is blue.
|
5DUQ3-Y_gX4
|
So this is our dx.
|
5DUQ3-Y_gX4
|
So I'm going to interpret
this more generally.
|
5DUQ3-Y_gX4
|
If x is going to be some kind
of vector or matrix or whatever,
|
5DUQ3-Y_gX4
|
this is just going to be a
linear operation on this.
|
5DUQ3-Y_gX4
|
And of course, for numbers,
a linear operation,
|
5DUQ3-Y_gX4
|
this is just a number.
|
5DUQ3-Y_gX4
|
If dx is a number, the only
linear operation you can do
|
5DUQ3-Y_gX4
|
is multiply by a number.
|
5DUQ3-Y_gX4
|
So let me just remind you if
you haven't taken linear algebra
|
5DUQ3-Y_gX4
|
for a while, a review.
|
5DUQ3-Y_gX4
|
Let's talk about what
a linear operator is.
|
5DUQ3-Y_gX4
|
So suppose we have
some given vectors, v,
|
5DUQ3-Y_gX4
|
in some vector space.
|
5DUQ3-Y_gX4
|
That would be capital
V. And remember,
|
5DUQ3-Y_gX4
|
a vector space is anything
where you can basically
|
5DUQ3-Y_gX4
|
add, subtract, and
multiply by scalars.
|
5DUQ3-Y_gX4
|
We have a plus or minus
and times scalar operations
|
5DUQ3-Y_gX4
|
that stay in our vector space.
|
5DUQ3-Y_gX4
|
That's what the informal
definition of a vector space
|
5DUQ3-Y_gX4
|
is.
|
5DUQ3-Y_gX4
|
You can write out
axioms and so forth,
|
5DUQ3-Y_gX4
|
but that's basically
what it means.
|
5DUQ3-Y_gX4
|
A linear operator,
this is what we're
|
5DUQ3-Y_gX4
|
going to mean by linearization
in the derivative.
|
5DUQ3-Y_gX4
|
It has to be linear.
|
5DUQ3-Y_gX4
|
What does it mean to
be linear in general?
|
5DUQ3-Y_gX4
|
So we're going to call this--
|
5DUQ3-Y_gX4
|
I'm going to denote this
by, let's say, L of--
|
5DUQ3-Y_gX4
|
let me denote it by square
brackets or just by Lv.
|
5DUQ3-Y_gX4
|
When it's clear enough,
I'll just write it
|
5DUQ3-Y_gX4
|
as if it were a multiplication.
|
5DUQ3-Y_gX4
|
Often, that'll be clear enough.
|
5DUQ3-Y_gX4
|
This is really acting on--
|
5DUQ3-Y_gX4
|
when I write Lv,
it's not necessarily
|
5DUQ3-Y_gX4
|
an ordinary multiplication.
|
5DUQ3-Y_gX4
|
This is just going
to be acting on v.
|
5DUQ3-Y_gX4
|
So a linear operator is a rule
that basically takes a vector
|
5DUQ3-Y_gX4
|
and gives you a vector out maybe
in a different vector space.
|
5DUQ3-Y_gX4
|
So this is L takes
a vector in, v in,
|
5DUQ3-Y_gX4
|
and it gives you a vector--
|
5DUQ3-Y_gX4
|
that's a terrible L--
|
5DUQ3-Y_gX4
|
Lv out maybe in a
different vector space.
|
5DUQ3-Y_gX4
|
And linearity means
what you think it means.
|
5DUQ3-Y_gX4
|
It means if you
take, for example,
|
5DUQ3-Y_gX4
|
L of v1 plus v2, if you take
the sum of the inputs, that's
|
5DUQ3-Y_gX4
|
the same thing as L
of v1 plus L of v2.
|
5DUQ3-Y_gX4
|
So if you add inputs,
that's the same thing
|
5DUQ3-Y_gX4
|
as adding the outputs or if
you multiply by a scalar.
|
5DUQ3-Y_gX4
|
And so as usual
in linear algebra,
|
5DUQ3-Y_gX4
|
Greek letters are going to
denote scalars as you would.
|
5DUQ3-Y_gX4
|
So that's equal to--
|
5DUQ3-Y_gX4
|
you can pull out the scalar.
|
5DUQ3-Y_gX4
|
And so the nice thing
about linear operators
|
5DUQ3-Y_gX4
|
is we can define them on lots
of kinds of vector spaces.
|
5DUQ3-Y_gX4
|
So let's just do a
couple of examples
|
5DUQ3-Y_gX4
|
just to make sure we're
on the same page here.
|
5DUQ3-Y_gX4
|
And so, for example,
you could just
|
5DUQ3-Y_gX4
|
have L is multiplication
by a scalar.
|
5DUQ3-Y_gX4
|
So you can have just L
of v is just alpha v.
|
5DUQ3-Y_gX4
|
That's a perfectly
good linear operation.
|
5DUQ3-Y_gX4
|
And if your vector space,
if your vs are scalars,
|
5DUQ3-Y_gX4
|
this is the only option.
|
5DUQ3-Y_gX4
|
If vs are, say,
real numbers, that's
|
5DUQ3-Y_gX4
|
a perfectly good vector space.
|
5DUQ3-Y_gX4
|
Another one that you're
very familiar with
|
5DUQ3-Y_gX4
|
is if L is a
multiplication by a matrix.
|
5DUQ3-Y_gX4
|
ALAN EDELMAN: Steven, maybe I'll
just point out sometimes people
|
5DUQ3-Y_gX4
|
like to ask me.
|
5DUQ3-Y_gX4
|
Wait, I thought that the linear
operators on scalars are,
|
5DUQ3-Y_gX4
|
I think in high school notation,
y equals mx plus b, right?
|
5DUQ3-Y_gX4
|
It's scalar times--
|
5DUQ3-Y_gX4
|
STEVEN G. JOHNSON: Yeah, yeah.
|
5DUQ3-Y_gX4
|
ALAN EDELMAN: --plus an
offset that may not be 0.
|
5DUQ3-Y_gX4
|
So what's going on here?
|
5DUQ3-Y_gX4
|
Is that linear or not linear?
|
5DUQ3-Y_gX4
|
STEVEN G. JOHNSON: Yeah.
|
5DUQ3-Y_gX4
|
So what about-- yeah,
so let's do that.
|
5DUQ3-Y_gX4
|
Let me call it a
different thing.
|
5DUQ3-Y_gX4
|
What letter should I use?
|
5DUQ3-Y_gX4
|
O, let's use O. Ov equals 2v
plus 1 for v is real numbers.
|
5DUQ3-Y_gX4
|
ALAN EDELMAN: Is that
linear or not linear?
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.