Johnson-Lindenstrauss lemma
[edit]
The Johnson-Lindenstrauss lemma asserts that a set of n points in any high dimensional Euclidean space can be mapped down into an
dimensional Euclidean space such that the distance between any two points changes by only a factor of
for any
.
Johnson and Lindenstrauss {cite} proved a fundamental mathematical result: any
point set in any Euclidean space can be embedded in
dimensions without distorting the distances between any pair of points by more than a factor of
, for any
. The original proof of Johnson and Lindenstrauss was much simplified by Frankl and Maehara {cite}, using geometric insights and refined approximation techniques.
Suppose we have a set of
-dimensional points
and we map them down to
dimensions, for appropriate constant
. Define
as the linear map, that is if
, then
. For example
could be a
matrix.
The general proof framework
All known proofs of the Johnson-Lindenstrauss lemma proceed according to the following scheme: For given
and an appropriate
, one defines a suitable probability distribution
on the set of all linear maps
. Then one proves the following statement:
Statement: If any
is a random linear mapping drown from the distribution
, then for every vector
we have
Having established this statement for the considered distribution
, the JL result follows
easily: We choose
at random according to F. Then for every
, using linearity
of
and the above Statement with
, we get that
fails to satisfy
with probability at most
. Consequently, the probability
that any of the
pairwise distances is distorted by
by more than
is at most
. Therefore, a random
works with probability at least
.
- S. Dasgupta and A. Gupta, An elementary proof of the Johnson-Lindenstrauss lemma, Tech. Rep. TR-99-06, Intl. Comput. Sci. Inst., Berkeley, CA, 1999.
- W. Johnson and J. Lindenstrauss. Extensions of Lipschitz maps into a Hilbert space. Contemporary Mathematics, 26:189--206, 1984.
Fast monte-carlo algorithms for MM
[edit]
Given an
matrix
and an
matrix
,
we present 2 simple and intuitive algorithms to compute
an approximation P to the product
, with provable
bounds for the norm of the "error matrix"
.
Both algorithms run in
time. In both
algorithms, we randomly pick
columns of A to
form an
matrix S and the corresponding rows of
B to form an
matrix R. After scaling the columns
of S and the rows of R, we multiply them together to
obtain our approximation P . The choice of the probability
distribution we use for picking the columns of
A and the scaling are the crucial features which enable
us to give fairly elementary proofs of the error bounds.
Our rest algorithm can be implemented without storing
the matrices A and B in Random Access Memory,
provided we can make two passes through the matrices
(stored in external memory). The second algorithm has
a smaller bound on the 2-norm of the error matrix, but
requires storage of A and B in RAM. We also present
a fast algorithm that \describes" P as a sum of rank
one matrices if B = A
![](//upload.wikimedia.org/wikipedia/commons/thumb/f/fe/Unbalanced_scales.svg/45px-Unbalanced_scales.svg.png) | This user strives to maintain a policy of neutrality on controversial issues. |
|
|
|
|
|
|
This user has a website, which can be found here.
|
|
![KDE](//upload.wikimedia.org/wikipedia/commons/thumb/8/8d/KDE_logo.svg/43px-KDE_logo.svg.png) | This user contributes with the KDE desktop environment. |
|
C | This user can program in C. |
|