Norm of sum of shifted outer products
Clash Royale CLAN TAG#URR8PPP
up vote
2
down vote
favorite
Consider two vectors $u, v in mathbbR^m$ satisfying $| u |_2 = | v |_2 = 1$ and $#mathrmsupp(v) leq n < m$, where $mathrmsupp$ denotes the support of the vector (i.e. locations of nonzero coordinates). Additionally, assume that $mathrmsupp(u) = mathrmsupp(v)$. Assume that $u^i$ denotes the left-circular shift of $u$ by $i$ locations, i.e.
$$
u = (u_1, dots, u_n, 0, dots, 0) Rightarrow u^1 =
(0, u_2, dots, u_n, 0 dots, u_1)
$$
Form the following matrix:
$$
X = u v^top + u^1 v^1^top + dots + u^m-1 v^m-1^top
$$
Question:
I want to obtain a nontrivial bound on the operator and Frobenius norms of this matrix. The only bound that I was able to obtain so far is using the subadditivity of the Frobenius norm and the fact that $| u_i v_i^top |_F = 1, ; forall i.$
$$
| X |_mathrmop leq | X |_F leq m | u_i v_i^top | = m
$$
Any ideas or pointers to references are more than welcome.
linear-algebra matrices reference-request spheres matrix-norms
add a comment |Â
up vote
2
down vote
favorite
Consider two vectors $u, v in mathbbR^m$ satisfying $| u |_2 = | v |_2 = 1$ and $#mathrmsupp(v) leq n < m$, where $mathrmsupp$ denotes the support of the vector (i.e. locations of nonzero coordinates). Additionally, assume that $mathrmsupp(u) = mathrmsupp(v)$. Assume that $u^i$ denotes the left-circular shift of $u$ by $i$ locations, i.e.
$$
u = (u_1, dots, u_n, 0, dots, 0) Rightarrow u^1 =
(0, u_2, dots, u_n, 0 dots, u_1)
$$
Form the following matrix:
$$
X = u v^top + u^1 v^1^top + dots + u^m-1 v^m-1^top
$$
Question:
I want to obtain a nontrivial bound on the operator and Frobenius norms of this matrix. The only bound that I was able to obtain so far is using the subadditivity of the Frobenius norm and the fact that $| u_i v_i^top |_F = 1, ; forall i.$
$$
| X |_mathrmop leq | X |_F leq m | u_i v_i^top | = m
$$
Any ideas or pointers to references are more than welcome.
linear-algebra matrices reference-request spheres matrix-norms
1
I'd expect the worst case to be constant $u$ and $v$ on a contiguous support. That would yield a bound of roughly $sqrtnm$ for the Frobenius norm.
â joriki
Jul 24 at 4:08
add a comment |Â
up vote
2
down vote
favorite
up vote
2
down vote
favorite
Consider two vectors $u, v in mathbbR^m$ satisfying $| u |_2 = | v |_2 = 1$ and $#mathrmsupp(v) leq n < m$, where $mathrmsupp$ denotes the support of the vector (i.e. locations of nonzero coordinates). Additionally, assume that $mathrmsupp(u) = mathrmsupp(v)$. Assume that $u^i$ denotes the left-circular shift of $u$ by $i$ locations, i.e.
$$
u = (u_1, dots, u_n, 0, dots, 0) Rightarrow u^1 =
(0, u_2, dots, u_n, 0 dots, u_1)
$$
Form the following matrix:
$$
X = u v^top + u^1 v^1^top + dots + u^m-1 v^m-1^top
$$
Question:
I want to obtain a nontrivial bound on the operator and Frobenius norms of this matrix. The only bound that I was able to obtain so far is using the subadditivity of the Frobenius norm and the fact that $| u_i v_i^top |_F = 1, ; forall i.$
$$
| X |_mathrmop leq | X |_F leq m | u_i v_i^top | = m
$$
Any ideas or pointers to references are more than welcome.
linear-algebra matrices reference-request spheres matrix-norms
Consider two vectors $u, v in mathbbR^m$ satisfying $| u |_2 = | v |_2 = 1$ and $#mathrmsupp(v) leq n < m$, where $mathrmsupp$ denotes the support of the vector (i.e. locations of nonzero coordinates). Additionally, assume that $mathrmsupp(u) = mathrmsupp(v)$. Assume that $u^i$ denotes the left-circular shift of $u$ by $i$ locations, i.e.
$$
u = (u_1, dots, u_n, 0, dots, 0) Rightarrow u^1 =
(0, u_2, dots, u_n, 0 dots, u_1)
$$
Form the following matrix:
$$
X = u v^top + u^1 v^1^top + dots + u^m-1 v^m-1^top
$$
Question:
I want to obtain a nontrivial bound on the operator and Frobenius norms of this matrix. The only bound that I was able to obtain so far is using the subadditivity of the Frobenius norm and the fact that $| u_i v_i^top |_F = 1, ; forall i.$
$$
| X |_mathrmop leq | X |_F leq m | u_i v_i^top | = m
$$
Any ideas or pointers to references are more than welcome.
linear-algebra matrices reference-request spheres matrix-norms
asked Jul 23 at 23:22
VHarisop
804421
804421
1
I'd expect the worst case to be constant $u$ and $v$ on a contiguous support. That would yield a bound of roughly $sqrtnm$ for the Frobenius norm.
â joriki
Jul 24 at 4:08
add a comment |Â
1
I'd expect the worst case to be constant $u$ and $v$ on a contiguous support. That would yield a bound of roughly $sqrtnm$ for the Frobenius norm.
â joriki
Jul 24 at 4:08
1
1
I'd expect the worst case to be constant $u$ and $v$ on a contiguous support. That would yield a bound of roughly $sqrtnm$ for the Frobenius norm.
â joriki
Jul 24 at 4:08
I'd expect the worst case to be constant $u$ and $v$ on a contiguous support. That would yield a bound of roughly $sqrtnm$ for the Frobenius norm.
â joriki
Jul 24 at 4:08
add a comment |Â
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2860859%2fnorm-of-sum-of-shifted-outer-products%23new-answer', 'question_page');
);
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
1
I'd expect the worst case to be constant $u$ and $v$ on a contiguous support. That would yield a bound of roughly $sqrtnm$ for the Frobenius norm.
â joriki
Jul 24 at 4:08