Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers...
Transcript of Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers...
![Page 1: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/1.jpg)
Uncertainty Principles for Fourier Multipliers
Michael Northington V
School of MathematicsGeorgia Tech
6/6/2018
With Shahaf Nitzan (Ga Tech) and Alex Powell (Vanderbilt)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 2: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/2.jpg)
Table of Contents
Exponentials in Weighted Spaces
Restrictions on Fourier Multipliers
Applications to Balian-Low Type Theorems
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 3: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/3.jpg)
Exponentials in Weighted Spaces
Let E = E(d) ={e2πik·x
}k∈Zd .
I E is an orthonormal basis for L2(Td) = L2(Rd/Zd).
I For a weight w satisfying w(x) > 0 almost everywhere,consider L2
w(Td) with norm,
‖g‖2L2w(Td)
=
∫Td|g(x)|2w(x)dx.
I Question 1: What basis properties does E have in L2w(Td)
and can these be characterized in terms of w?
I Question 2: Why do we care about this setting?
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 4: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/4.jpg)
Exponentials in Weighted Spaces
Let E = E(d) ={e2πik·x
}k∈Zd .
I E is an orthonormal basis for L2(Td) = L2(Rd/Zd).
I For a weight w satisfying w(x) > 0 almost everywhere,consider L2
w(Td) with norm,
‖g‖2L2w(Td)
=
∫Td|g(x)|2w(x)dx.
I Question 1: What basis properties does E have in L2w(Td)
and can these be characterized in terms of w?
I Question 2: Why do we care about this setting?
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 5: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/5.jpg)
Exponentials in Weighted Spaces
Let E = E(d) ={e2πik·x
}k∈Zd .
I E is an orthonormal basis for L2(Td) = L2(Rd/Zd).
I For a weight w satisfying w(x) > 0 almost everywhere,consider L2
w(Td) with norm,
‖g‖2L2w(Td)
=
∫Td|g(x)|2w(x)dx.
I Question 1: What basis properties does E have in L2w(Td)
and can these be characterized in terms of w?
I Question 2: Why do we care about this setting?
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 6: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/6.jpg)
Exponentials in Weighted Spaces
Let E = E(d) ={e2πik·x
}k∈Zd .
I E is an orthonormal basis for L2(Td) = L2(Rd/Zd).
I For a weight w satisfying w(x) > 0 almost everywhere,consider L2
w(Td) with norm,
‖g‖2L2w(Td)
=
∫Td|g(x)|2w(x)dx.
I Question 1: What basis properties does E have in L2w(Td)
and can these be characterized in terms of w?
I Question 2: Why do we care about this setting?
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 7: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/7.jpg)
Exponentials in Weighted Spaces
Let E = E(d) ={e2πik·x
}k∈Zd .
I E is an orthonormal basis for L2(Td) = L2(Rd/Zd).
I For a weight w satisfying w(x) > 0 almost everywhere,consider L2
w(Td) with norm,
‖g‖2L2w(Td)
=
∫Td|g(x)|2w(x)dx.
I Question 1: What basis properties does E have in L2w(Td)
and can these be characterized in terms of w?
I Question 2: Why do we care about this setting?
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 8: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/8.jpg)
Example 1: Gabor Systems and the Zak Transform
I Gabor System: For g ∈ L2(R),
G(g) := {e2πimxg(x− n)}m,n∈Z = {MmTng}m,n∈Z
I Zak Transform: Zg(x, y) :=∑
k∈Z g(x− k)e2πiky
I Converts TF-shifts to exponentials:
Z(MmTng) = e2πi(mx−ny)Zg
Z(G(g)) = {e2πi(mx−ny)Z(g)}n,m∈Z
I Leads to an isometric isomorphism:I L2(R)→ L2
w(T2), for w = |Zg|2I G(g)→ E = E(2).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 9: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/9.jpg)
Example 1: Gabor Systems and the Zak Transform
I Gabor System: For g ∈ L2(R),
G(g) := {e2πimxg(x− n)}m,n∈Z = {MmTng}m,n∈Z
I Zak Transform: Zg(x, y) :=∑
k∈Z g(x− k)e2πiky
I Converts TF-shifts to exponentials:
Z(MmTng) = e2πi(mx−ny)Zg
Z(G(g)) = {e2πi(mx−ny)Z(g)}n,m∈Z
I Leads to an isometric isomorphism:I L2(R)→ L2
w(T2), for w = |Zg|2I G(g)→ E = E(2).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 10: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/10.jpg)
Example 1: Gabor Systems and the Zak Transform
I Gabor System: For g ∈ L2(R),
G(g) := {e2πimxg(x− n)}m,n∈Z = {MmTng}m,n∈Z
I Zak Transform: Zg(x, y) :=∑
k∈Z g(x− k)e2πiky
I Converts TF-shifts to exponentials:
Z(MmTng) = e2πi(mx−ny)Zg
Z(G(g)) = {e2πi(mx−ny)Z(g)}n,m∈Z
I Leads to an isometric isomorphism:I L2(R)→ L2
w(T2), for w = |Zg|2I G(g)→ E = E(2).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 11: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/11.jpg)
Example 1: Gabor Systems and the Zak Transform
I Gabor System: For g ∈ L2(R),
G(g) := {e2πimxg(x− n)}m,n∈Z = {MmTng}m,n∈Z
I Zak Transform: Zg(x, y) :=∑
k∈Z g(x− k)e2πiky
I Converts TF-shifts to exponentials:
Z(MmTng) = e2πi(mx−ny)Zg
Z(G(g)) = {e2πi(mx−ny)Z(g)}n,m∈Z
I Leads to an isometric isomorphism:I L2(R)→ L2
w(T2), for w = |Zg|2I G(g)→ E = E(2).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 12: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/12.jpg)
Example 2: Shift-Invariant Spaces and Periodization
I Integer Translates: For f ∈ L2(Rd), T (f) = {f(· − l)}l∈Zd
I Shift-Invariant Space: V (f) = span(T (f))L2(Rd)
I Periodization: P f(ξ) =∑
k∈Zd |f(ξ − k)|2
I If h ∈ V (f), there exists a Zd-periodic m, so that h = mf .I Leads to an isometric isomorphism:
I V (f)→ L2w(Td), for w = P f
I h→ mI T (f)→ E = E(d)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 13: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/13.jpg)
Example 2: Shift-Invariant Spaces and Periodization
I Integer Translates: For f ∈ L2(Rd), T (f) = {f(· − l)}l∈Zd
I Shift-Invariant Space: V (f) = span(T (f))L2(Rd)
I Periodization: P f(ξ) =∑
k∈Zd |f(ξ − k)|2
I If h ∈ V (f), there exists a Zd-periodic m, so that h = mf .I Leads to an isometric isomorphism:
I V (f)→ L2w(Td), for w = P f
I h→ mI T (f)→ E = E(d)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 14: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/14.jpg)
Example 2: Shift-Invariant Spaces and Periodization
I Integer Translates: For f ∈ L2(Rd), T (f) = {f(· − l)}l∈Zd
I Shift-Invariant Space: V (f) = span(T (f))L2(Rd)
I Periodization: P f(ξ) =∑
k∈Zd |f(ξ − k)|2
I If h ∈ V (f), there exists a Zd-periodic m, so that h = mf .I Leads to an isometric isomorphism:
I V (f)→ L2w(Td), for w = P f
I h→ mI T (f)→ E = E(d)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 15: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/15.jpg)
Example 2: Shift-Invariant Spaces and Periodization
I Integer Translates: For f ∈ L2(Rd), T (f) = {f(· − l)}l∈Zd
I Shift-Invariant Space: V (f) = span(T (f))L2(Rd)
I Periodization: P f(ξ) =∑
k∈Zd |f(ξ − k)|2
I If h ∈ V (f), there exists a Zd-periodic m, so that h = mf .
I Leads to an isometric isomorphism:I V (f)→ L2
w(Td), for w = P fI h→ mI T (f)→ E = E(d)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 16: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/16.jpg)
Example 2: Shift-Invariant Spaces and Periodization
I Integer Translates: For f ∈ L2(Rd), T (f) = {f(· − l)}l∈Zd
I Shift-Invariant Space: V (f) = span(T (f))L2(Rd)
I Periodization: P f(ξ) =∑
k∈Zd |f(ξ − k)|2
I If h ∈ V (f), there exists a Zd-periodic m, so that h = mf .I Leads to an isometric isomorphism:
I V (f)→ L2w(Td), for w = P f
I h→ mI T (f)→ E = E(d)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 17: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/17.jpg)
Spanning and Independence Properties
Complete
Frame
Let H be a Hilbert space, and H = {hn}∞n=1 ⊂ H.
I H is complete if span H = H.
I H is a frame if it’s complete, and there existconstants 0 < A ≤ B <∞ with
∀h ∈ H, A‖h‖2H ≤∞∑n=1
|〈h, hn〉|2 ≤ B‖h‖2H.
I Every frame is complete, with the additionalbonus that there exist a choice of coefficientssuch that h =
∑cnhn with
‖cn‖l2 � ‖h‖H.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 18: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/18.jpg)
Spanning and Independence Properties
Complete
Frame
Let H be a Hilbert space, and H = {hn}∞n=1 ⊂ H.
I H is complete if span H = H.
I H is a frame if it’s complete, and there existconstants 0 < A ≤ B <∞ with
∀h ∈ H, A‖h‖2H ≤∞∑n=1
|〈h, hn〉|2 ≤ B‖h‖2H.
I Every frame is complete, with the additionalbonus that there exist a choice of coefficientssuch that h =
∑cnhn with
‖cn‖l2 � ‖h‖H.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 19: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/19.jpg)
Spanning and Independence Properties
Complete
Frame
Let H be a Hilbert space, and H = {hn}∞n=1 ⊂ H.
I H is complete if span H = H.
I H is a frame if it’s complete, and there existconstants 0 < A ≤ B <∞ with
∀h ∈ H, A‖h‖2H ≤∞∑n=1
|〈h, hn〉|2 ≤ B‖h‖2H.
I Every frame is complete, with the additionalbonus that there exist a choice of coefficientssuch that h =
∑cnhn with
‖cn‖l2 � ‖h‖H.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 20: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/20.jpg)
Spanning and Independence Properties
CompleteExact
FrameRiesz Basis
I H is a minimal system if for each n,
hn /∈ span{hm : m 6= n}.
I H is exact if it is complete and minimal.
I H is a Riesz basis if there is an orthonormalbasis {en}∞n=1 and a bounded invertibleoperator T on H such that
Ten = hn.
I Riesz basis =⇒ frame; Riesz basis ⇐⇒minimal frame.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 21: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/21.jpg)
Spanning and Independence Properties
CompleteExact
FrameRiesz Basis
I H is a minimal system if for each n,
hn /∈ span{hm : m 6= n}.
I H is exact if it is complete and minimal.
I H is a Riesz basis if there is an orthonormalbasis {en}∞n=1 and a bounded invertibleoperator T on H such that
Ten = hn.
I Riesz basis =⇒ frame; Riesz basis ⇐⇒minimal frame.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 22: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/22.jpg)
Spanning and Independence Properties
CompleteExact
FrameRiesz Basis
I H is a minimal system if for each n,
hn /∈ span{hm : m 6= n}.
I H is exact if it is complete and minimal.
I H is a Riesz basis if there is an orthonormalbasis {en}∞n=1 and a bounded invertibleoperator T on H such that
Ten = hn.
I Riesz basis =⇒ frame; Riesz basis ⇐⇒minimal frame.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 23: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/23.jpg)
Spanning and Independence Properties
CompleteExact
FrameRiesz Basis
I H is a minimal system if for each n,
hn /∈ span{hm : m 6= n}.
I H is exact if it is complete and minimal.
I H is a Riesz basis if there is an orthonormalbasis {en}∞n=1 and a bounded invertibleoperator T on H such that
Ten = hn.
I Riesz basis =⇒ frame; Riesz basis ⇐⇒minimal frame.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 24: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/24.jpg)
(Cq)-systems (Olevskii, Nitzan ’07)
I Fix 2 ≤ q ≤ ∞. {hn}∞n=1 ⊂ H is a (Cq)-system if for eachh ∈ H, h can be approximated to arbitrary accuracy by afinite sum
∑anhn such that
‖an‖lq ≤ C‖h‖H.
I Equivalently, {hn}∞n=1 is a (Cq)-system if and only if
‖h‖H ≤ C
( ∞∑n=1
|〈h, hn〉|q′
)1/q′
I (Cq) stands for completeness with lq control of coefficients.
I Bessel (C2)-system ⇐⇒ frame
I (Cq)-system =⇒ (Cq′)-system for all q′ ≥ q
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 25: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/25.jpg)
(Cq)-systems (Olevskii, Nitzan ’07)
I Fix 2 ≤ q ≤ ∞. {hn}∞n=1 ⊂ H is a (Cq)-system if for eachh ∈ H, h can be approximated to arbitrary accuracy by afinite sum
∑anhn such that
‖an‖lq ≤ C‖h‖H.
I Equivalently, {hn}∞n=1 is a (Cq)-system if and only if
‖h‖H ≤ C
( ∞∑n=1
|〈h, hn〉|q′
)1/q′
I (Cq) stands for completeness with lq control of coefficients.
I Bessel (C2)-system ⇐⇒ frame
I (Cq)-system =⇒ (Cq′)-system for all q′ ≥ q
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 26: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/26.jpg)
(Cq)-systems (Olevskii, Nitzan ’07)
I Fix 2 ≤ q ≤ ∞. {hn}∞n=1 ⊂ H is a (Cq)-system if for eachh ∈ H, h can be approximated to arbitrary accuracy by afinite sum
∑anhn such that
‖an‖lq ≤ C‖h‖H.
I Equivalently, {hn}∞n=1 is a (Cq)-system if and only if
‖h‖H ≤ C
( ∞∑n=1
|〈h, hn〉|q′
)1/q′
I (Cq) stands for completeness with lq control of coefficients.
I Bessel (C2)-system ⇐⇒ frame
I (Cq)-system =⇒ (Cq′)-system for all q′ ≥ q
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 27: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/27.jpg)
(Cq)-systems (Olevskii, Nitzan ’07)
I Fix 2 ≤ q ≤ ∞. {hn}∞n=1 ⊂ H is a (Cq)-system if for eachh ∈ H, h can be approximated to arbitrary accuracy by afinite sum
∑anhn such that
‖an‖lq ≤ C‖h‖H.
I Equivalently, {hn}∞n=1 is a (Cq)-system if and only if
‖h‖H ≤ C
( ∞∑n=1
|〈h, hn〉|q′
)1/q′
I (Cq) stands for completeness with lq control of coefficients.
I Bessel (C2)-system ⇐⇒ frame
I (Cq)-system =⇒ (Cq′)-system for all q′ ≥ q
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 28: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/28.jpg)
(Cq)-systems (Olevskii, Nitzan ’07)
I Fix 2 ≤ q ≤ ∞. {hn}∞n=1 ⊂ H is a (Cq)-system if for eachh ∈ H, h can be approximated to arbitrary accuracy by afinite sum
∑anhn such that
‖an‖lq ≤ C‖h‖H.
I Equivalently, {hn}∞n=1 is a (Cq)-system if and only if
‖h‖H ≤ C
( ∞∑n=1
|〈h, hn〉|q′
)1/q′
I (Cq) stands for completeness with lq control of coefficients.
I Bessel (C2)-system ⇐⇒ frame
I (Cq)-system =⇒ (Cq′)-system for all q′ ≥ q
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 29: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/29.jpg)
(Cq)-systems
CompleteExact
(C2)-systemFrame
Riesz Basis
(C∞)-system
q ↗
Exact
Riesz Basis
q ↗
For Exact E in L2w(Td):
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 30: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/30.jpg)
(Cq)-systems
CompleteExact
(C2)-systemFrame
Riesz Basis
(C∞)-system
q ↗
Exact
Riesz Basis
q ↗
For Exact E in L2w(Td):
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 31: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/31.jpg)
Back to Question 1
What is known about basis properties of E in L2w(Td)?
Property Characterization
Riesz Basis∃0 < A ≤ B <∞ such thatA ≤ |w(x)| ≤ B, for a.e. x
FrameFor S = {x : w(x) > 0},
w ∈ L∞(Td), w−1 ∈ L∞(S)
Minimal System w−1 ∈ L1(Td)Exact (Cq)-system w−1/2 ∈Mq
2
I The first 3 are well known: de Boor, DeVore, Ron (’92), Ron,Shen (’95), Bownik (’00)
I Nitzan, Olsen (’11) gave necessary and sufficient conditionssimilar to the fourth characterization
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 32: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/32.jpg)
Back to Question 1
What is known about basis properties of E in L2w(Td)?
Property Characterization
Riesz Basis∃0 < A ≤ B <∞ such thatA ≤ |w(x)| ≤ B, for a.e. x
FrameFor S = {x : w(x) > 0},
w ∈ L∞(Td), w−1 ∈ L∞(S)
Minimal System w−1 ∈ L1(Td)Exact (Cq)-system w−1/2 ∈Mq
2
I The first 3 are well known: de Boor, DeVore, Ron (’92), Ron,Shen (’95), Bownik (’00)
I Nitzan, Olsen (’11) gave necessary and sufficient conditionssimilar to the fourth characterization
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 33: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/33.jpg)
Back to Question 1
What is known about basis properties of E in L2w(Td)?
Property Characterization
Riesz Basis∃0 < A ≤ B <∞ such thatA ≤ |w(x)| ≤ B, for a.e. x
FrameFor S = {x : w(x) > 0},
w ∈ L∞(Td), w−1 ∈ L∞(S)
Minimal System w−1 ∈ L1(Td)
Exact (Cq)-system w−1/2 ∈Mq2
I The first 3 are well known: de Boor, DeVore, Ron (’92), Ron,Shen (’95), Bownik (’00)
I Nitzan, Olsen (’11) gave necessary and sufficient conditionssimilar to the fourth characterization
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 34: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/34.jpg)
Back to Question 1
What is known about basis properties of E in L2w(Td)?
Property Characterization
Riesz Basis∃0 < A ≤ B <∞ such thatA ≤ |w(x)| ≤ B, for a.e. x
FrameFor S = {x : w(x) > 0},
w ∈ L∞(Td), w−1 ∈ L∞(S)
Minimal System w−1 ∈ L1(Td)Exact (Cq)-system w−1/2 ∈Mq
2
I The first 3 are well known: de Boor, DeVore, Ron (’92), Ron,Shen (’95), Bownik (’00)
I Nitzan, Olsen (’11) gave necessary and sufficient conditionssimilar to the fourth characterization
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 35: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/35.jpg)
Back to Question 1
What is known about basis properties of E in L2w(Td)?
Property Characterization
Riesz Basis∃0 < A ≤ B <∞ such thatA ≤ |w(x)| ≤ B, for a.e. x
FrameFor S = {x : w(x) > 0},
w ∈ L∞(Td), w−1 ∈ L∞(S)
Minimal System w−1 ∈ L1(Td)Exact (Cq)-system w−1/2 ∈Mq
2
I The first 3 are well known: de Boor, DeVore, Ron (’92), Ron,Shen (’95), Bownik (’00)
I Nitzan, Olsen (’11) gave necessary and sufficient conditionssimilar to the fourth characterization
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 36: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/36.jpg)
Back to Question 1
What is known about basis properties of E in L2w(Td)?
Property Characterization
Riesz Basis∃0 < A ≤ B <∞ such thatA ≤ |w(x)| ≤ B, for a.e. x
FrameFor S = {x : w(x) > 0},
w ∈ L∞(Td), w−1 ∈ L∞(S)
Minimal System w−1 ∈ L1(Td)Exact (Cq)-system w−1/2 ∈Mq
2
I The first 3 are well known: de Boor, DeVore, Ron (’92), Ron,Shen (’95), Bownik (’00)
I Nitzan, Olsen (’11) gave necessary and sufficient conditionssimilar to the fourth characterization
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 37: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/37.jpg)
What is Mq2?
c−1
c0
c1
c2
−1 0 2
−1 −0.5 0 0.5 1
−2
−1
0
1
2
c
−1 −0.5 0 0.5 1
−2
−1
0
1
2
h = uc
a−1
a0
a1
a2
−1 0 1 2
h
Tuc
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 38: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/38.jpg)
What is Mq2?
c−1
c0
c1
c2
−1 0 2
−1 −0.5 0 0.5 1
−2
−1
0
1
2
c
−1 −0.5 0 0.5 1
−2
−1
0
1
2
h = uc
a−1
a0
a1
a2
−1 0 1 2
h
Tuc
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 39: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/39.jpg)
What is Mq2?
c−1
c0
c1
c2
−1 0 2
−1 −0.5 0 0.5 1
−2
−1
0
1
2
c
−1 −0.5 0 0.5 1
−2
−1
0
1
2
h = uc
a−1
a0
a1
a2
−1 0 1 2
h
Tuc
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 40: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/40.jpg)
What is Mq2?
c−1
c0
c1
c2
−1 0 2
−1 −0.5 0 0.5 1
−2
−1
0
1
2
c
−1 −0.5 0 0.5 1
−2
−1
0
1
2
h = uc
a−1
a0
a1
a2
−1 0 1 2
h
Tuc
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 41: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/41.jpg)
What is Mq2?
c−1
c0
c1
c2
−1 0 2
−1 −0.5 0 0.5 1
−2
−1
0
1
2
c
−1 −0.5 0 0.5 1
−2
−1
0
1
2
h = uc
a−1
a0
a1
a2
−1 0 1 2
h
Tuc
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 42: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/42.jpg)
What is Mq2?
I For a periodic function u and a finite sequence c = {cn}n∈Zd ,
define Tu by Tuc = uc.
I Then, u ∈Mq2, if for all such c,
‖Tuc‖lq(Zd) ≤ C‖c‖l2(Zd)
I Properties and Special Cases:I If q < 2, Mq
2 = {0}I M2
2 = L∞(Td) (Agrees with Riesz basis characterization)I M∞
2 = L2(Td) (Agrees with minimal system characterization)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 43: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/43.jpg)
What is Mq2?
I For a periodic function u and a finite sequence c = {cn}n∈Zd ,
define Tu by Tuc = uc.
I Then, u ∈Mq2, if for all such c,
‖Tuc‖lq(Zd) ≤ C‖c‖l2(Zd)
I Properties and Special Cases:I If q < 2, Mq
2 = {0}I M2
2 = L∞(Td) (Agrees with Riesz basis characterization)I M∞
2 = L2(Td) (Agrees with minimal system characterization)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 44: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/44.jpg)
What is Mq2?
I For a periodic function u and a finite sequence c = {cn}n∈Zd ,
define Tu by Tuc = uc.
I Then, u ∈Mq2, if for all such c,
‖Tuc‖lq(Zd) ≤ C‖c‖l2(Zd)
I Properties and Special Cases:I If q < 2, Mq
2 = {0}
I M22 = L∞(Td) (Agrees with Riesz basis characterization)
I M∞2 = L2(Td) (Agrees with minimal system characterization)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 45: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/45.jpg)
What is Mq2?
I For a periodic function u and a finite sequence c = {cn}n∈Zd ,
define Tu by Tuc = uc.
I Then, u ∈Mq2, if for all such c,
‖Tuc‖lq(Zd) ≤ C‖c‖l2(Zd)
I Properties and Special Cases:I If q < 2, Mq
2 = {0}I M2
2 = L∞(Td) (Agrees with Riesz basis characterization)
I M∞2 = L2(Td) (Agrees with minimal system characterization)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 46: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/46.jpg)
What is Mq2?
I For a periodic function u and a finite sequence c = {cn}n∈Zd ,
define Tu by Tuc = uc.
I Then, u ∈Mq2, if for all such c,
‖Tuc‖lq(Zd) ≤ C‖c‖l2(Zd)
I Properties and Special Cases:I If q < 2, Mq
2 = {0}I M2
2 = L∞(Td) (Agrees with Riesz basis characterization)I M∞
2 = L2(Td) (Agrees with minimal system characterization)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 47: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/47.jpg)
Table of Contents
Exponentials in Weighted Spaces
Restrictions on Fourier Multipliers
Applications to Balian-Low Type Theorems
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 48: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/48.jpg)
What can prevent u from being a bounded multiplier?
Based on connections with uncertainty principles (to be describedlater in the talk) we wish to study what properties would prevent afunction u from being in Mq
2.
If u ∈ L∞(Td) =M22, then u ∈Mq
2 for all q ≥ 2, but if u isunbounded it will fail to be in Mq
2 for small q.
We will assume that w = 1/u is smooth in the sense of Sobolevspaces, and that w has a zero or a set of zeros, and we will try todetermine when the level of smoothness or the size of the zero setbecomes too large to allow u ∈Mq
2.
Sobolev Space:
Hs(Td) = {f ∈ L2(Td) :∑k∈Zd|k|2s|f(k)|2 <∞}
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 49: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/49.jpg)
What can prevent u from being a bounded multiplier?
Based on connections with uncertainty principles (to be describedlater in the talk) we wish to study what properties would prevent afunction u from being in Mq
2.
If u ∈ L∞(Td) =M22, then u ∈Mq
2 for all q ≥ 2, but if u isunbounded it will fail to be in Mq
2 for small q.
We will assume that w = 1/u is smooth in the sense of Sobolevspaces, and that w has a zero or a set of zeros, and we will try todetermine when the level of smoothness or the size of the zero setbecomes too large to allow u ∈Mq
2.
Sobolev Space:
Hs(Td) = {f ∈ L2(Td) :∑k∈Zd|k|2s|f(k)|2 <∞}
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 50: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/50.jpg)
What can prevent u from being a bounded multiplier?
Based on connections with uncertainty principles (to be describedlater in the talk) we wish to study what properties would prevent afunction u from being in Mq
2.
If u ∈ L∞(Td) =M22, then u ∈Mq
2 for all q ≥ 2, but if u isunbounded it will fail to be in Mq
2 for small q.
We will assume that w = 1/u is smooth in the sense of Sobolevspaces, and that w has a zero or a set of zeros, and we will try todetermine when the level of smoothness or the size of the zero setbecomes too large to allow u ∈Mq
2.
Sobolev Space:
Hs(Td) = {f ∈ L2(Td) :∑k∈Zd|k|2s|f(k)|2 <∞}
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 51: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/51.jpg)
What can prevent u from being a bounded multiplier?
Based on connections with uncertainty principles (to be describedlater in the talk) we wish to study what properties would prevent afunction u from being in Mq
2.
If u ∈ L∞(Td) =M22, then u ∈Mq
2 for all q ≥ 2, but if u isunbounded it will fail to be in Mq
2 for small q.
We will assume that w = 1/u is smooth in the sense of Sobolevspaces, and that w has a zero or a set of zeros, and we will try todetermine when the level of smoothness or the size of the zero setbecomes too large to allow u ∈Mq
2.
Sobolev Space:
Hs(Td) = {f ∈ L2(Td) :∑k∈Zd|k|2s|f(k)|2 <∞}
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 52: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/52.jpg)
Results with a single zero
Theorem (Nitzan, M.N., Powell)
Let d2 ≤ s ≤ d, and suppose w ∈ Hs(Td) and w has a zero.
1. If d2 ≤ s <d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q ≤ dd−s . Conversely, for any q >
dd−s there exists
w ∈ Hs(Td) such that w has a zero and u = 1w ∈M
q2.
2. If s = d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2 . Conversely, there exists w ∈ C∞(Td) with a
zero, such that u = 1/w ∈Mq2 for any q > 2d
d−2 .
I Proof relies on Sobolev Embedding Theorem in Holder spaces.
I For s > d2 + 1, we can’t say more than the bound in part 2
unless we require a zero of a larger order.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 53: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/53.jpg)
Results with a single zero
Theorem (Nitzan, M.N., Powell)
Let d2 ≤ s ≤ d, and suppose w ∈ Hs(Td) and w has a zero.
1. If d2 ≤ s <d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q ≤ dd−s .
Conversely, for any q > dd−s there exists
w ∈ Hs(Td) such that w has a zero and u = 1w ∈M
q2.
2. If s = d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2 . Conversely, there exists w ∈ C∞(Td) with a
zero, such that u = 1/w ∈Mq2 for any q > 2d
d−2 .
I Proof relies on Sobolev Embedding Theorem in Holder spaces.
I For s > d2 + 1, we can’t say more than the bound in part 2
unless we require a zero of a larger order.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 54: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/54.jpg)
Results with a single zero
Theorem (Nitzan, M.N., Powell)
Let d2 ≤ s ≤ d, and suppose w ∈ Hs(Td) and w has a zero.
1. If d2 ≤ s <d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q ≤ dd−s . Conversely, for any q >
dd−s there exists
w ∈ Hs(Td) such that w has a zero and u = 1w ∈M
q2.
2. If s = d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2 . Conversely, there exists w ∈ C∞(Td) with a
zero, such that u = 1/w ∈Mq2 for any q > 2d
d−2 .
I Proof relies on Sobolev Embedding Theorem in Holder spaces.
I For s > d2 + 1, we can’t say more than the bound in part 2
unless we require a zero of a larger order.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 55: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/55.jpg)
Results with a single zero
Theorem (Nitzan, M.N., Powell)
Let d2 ≤ s ≤ d, and suppose w ∈ Hs(Td) and w has a zero.
1. If d2 ≤ s <d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q ≤ dd−s . Conversely, for any q >
dd−s there exists
w ∈ Hs(Td) such that w has a zero and u = 1w ∈M
q2.
2. If s = d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2 .
Conversely, there exists w ∈ C∞(Td) with a
zero, such that u = 1/w ∈Mq2 for any q > 2d
d−2 .
I Proof relies on Sobolev Embedding Theorem in Holder spaces.
I For s > d2 + 1, we can’t say more than the bound in part 2
unless we require a zero of a larger order.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 56: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/56.jpg)
Results with a single zero
Theorem (Nitzan, M.N., Powell)
Let d2 ≤ s ≤ d, and suppose w ∈ Hs(Td) and w has a zero.
1. If d2 ≤ s <d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q ≤ dd−s . Conversely, for any q >
dd−s there exists
w ∈ Hs(Td) such that w has a zero and u = 1w ∈M
q2.
2. If s = d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2 . Conversely, there exists w ∈ C∞(Td) with a
zero, such that u = 1/w ∈Mq2 for any q > 2d
d−2 .
I Proof relies on Sobolev Embedding Theorem in Holder spaces.
I For s > d2 + 1, we can’t say more than the bound in part 2
unless we require a zero of a larger order.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 57: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/57.jpg)
Results with a single zero
Theorem (Nitzan, M.N., Powell)
Let d2 ≤ s ≤ d, and suppose w ∈ Hs(Td) and w has a zero.
1. If d2 ≤ s <d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q ≤ dd−s . Conversely, for any q >
dd−s there exists
w ∈ Hs(Td) such that w has a zero and u = 1w ∈M
q2.
2. If s = d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2 . Conversely, there exists w ∈ C∞(Td) with a
zero, such that u = 1/w ∈Mq2 for any q > 2d
d−2 .
I Proof relies on Sobolev Embedding Theorem in Holder spaces.
I For s > d2 + 1, we can’t say more than the bound in part 2
unless we require a zero of a larger order.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 58: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/58.jpg)
Results with a single zero
Theorem (Nitzan, M.N., Powell)
Let d2 ≤ s ≤ d, and suppose w ∈ Hs(Td) and w has a zero.
1. If d2 ≤ s <d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q ≤ dd−s . Conversely, for any q >
dd−s there exists
w ∈ Hs(Td) such that w has a zero and u = 1w ∈M
q2.
2. If s = d2 + 1, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2 . Conversely, there exists w ∈ C∞(Td) with a
zero, such that u = 1/w ∈Mq2 for any q > 2d
d−2 .
I Proof relies on Sobolev Embedding Theorem in Holder spaces.
I For s > d2 + 1, we can’t say more than the bound in part 2
unless we require a zero of a larger order.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 59: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/59.jpg)
Zero Sets of Larger Hausdorff Dimension
I We are also interested in finding a similar result when the zeroset of w = 1
u has a larger Hausdorff dimension.
I In this case, our functions may not be continuous, so wedefine our zero set as
Σ(w) =
{x ∈ Td : lim sup
τ→0
1
|Bτ |
∫Bτ (x)
|w(y)|dy = 0
}.
I A similar question was studied by Jiang, Lin (’03) andSchikorra (’13) with the Fourier multiplier condition replacedwith an integrability condition.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 60: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/60.jpg)
Zero Sets of Larger Hausdorff Dimension
I We are also interested in finding a similar result when the zeroset of w = 1
u has a larger Hausdorff dimension.
I In this case, our functions may not be continuous, so wedefine our zero set as
Σ(w) =
{x ∈ Td : lim sup
τ→0
1
|Bτ |
∫Bτ (x)
|w(y)|dy = 0
}.
I A similar question was studied by Jiang, Lin (’03) andSchikorra (’13) with the Fourier multiplier condition replacedwith an integrability condition.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 61: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/61.jpg)
Zero Sets of Larger Hausdorff Dimension
I We are also interested in finding a similar result when the zeroset of w = 1
u has a larger Hausdorff dimension.
I In this case, our functions may not be continuous, so wedefine our zero set as
Σ(w) =
{x ∈ Td : lim sup
τ→0
1
|Bτ |
∫Bτ (x)
|w(y)|dy = 0
}.
I A similar question was studied by Jiang, Lin (’03) andSchikorra (’13) with the Fourier multiplier condition replacedwith an integrability condition.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 62: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/62.jpg)
Zero Sets of Larger Hausdorff Dimension
Theorem (Nitzan, M.N., Powell)
Let 0 ≤ σ ≤ d and d−σ2 ≤ s ≤ d− σ. Suppose w ∈W
s,2(Td) andHσ(Σ(w)) > 0.
1. If d−σ2 ≤ s < min(d− σ, d2 + 1), then u = 1w /∈Mq
2 for any q
satisfying 2 ≤ q ≤ dd−s−σ/2 .
2. If s = d2 + 1 ≤ d− σ, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2−σ .
3. If s = d− σ < d2 + 1, then u = 1
w /∈Mq2 for any q.
I Part 3 is sharp, but parts 1 and 2 likely are not.I Based on the results of Jiang, Lin (’03) and Schikorra (’13), I
(we?) conjecture that part 1 holds with 2 ≤ q ≤ d−σd−σ−s .
I Proof uses a version of Poincare Inequality from Jiang, Lin(’03) and Schikorra (’13).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 63: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/63.jpg)
Zero Sets of Larger Hausdorff Dimension
Theorem (Nitzan, M.N., Powell)
Let 0 ≤ σ ≤ d and d−σ2 ≤ s ≤ d− σ. Suppose w ∈W
s,2(Td) andHσ(Σ(w)) > 0.
1. If d−σ2 ≤ s < min(d− σ, d2 + 1), then u = 1w /∈Mq
2 for any q
satisfying 2 ≤ q ≤ dd−s−σ/2 .
2. If s = d2 + 1 ≤ d− σ, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2−σ .
3. If s = d− σ < d2 + 1, then u = 1
w /∈Mq2 for any q.
I Part 3 is sharp, but parts 1 and 2 likely are not.I Based on the results of Jiang, Lin (’03) and Schikorra (’13), I
(we?) conjecture that part 1 holds with 2 ≤ q ≤ d−σd−σ−s .
I Proof uses a version of Poincare Inequality from Jiang, Lin(’03) and Schikorra (’13).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 64: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/64.jpg)
Zero Sets of Larger Hausdorff Dimension
Theorem (Nitzan, M.N., Powell)
Let 0 ≤ σ ≤ d and d−σ2 ≤ s ≤ d− σ. Suppose w ∈W
s,2(Td) andHσ(Σ(w)) > 0.
1. If d−σ2 ≤ s < min(d− σ, d2 + 1), then u = 1w /∈Mq
2 for any q
satisfying 2 ≤ q ≤ dd−s−σ/2 .
2. If s = d2 + 1 ≤ d− σ, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2−σ .
3. If s = d− σ < d2 + 1, then u = 1
w /∈Mq2 for any q.
I Part 3 is sharp, but parts 1 and 2 likely are not.I Based on the results of Jiang, Lin (’03) and Schikorra (’13), I
(we?) conjecture that part 1 holds with 2 ≤ q ≤ d−σd−σ−s .
I Proof uses a version of Poincare Inequality from Jiang, Lin(’03) and Schikorra (’13).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 65: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/65.jpg)
Zero Sets of Larger Hausdorff Dimension
Theorem (Nitzan, M.N., Powell)
Let 0 ≤ σ ≤ d and d−σ2 ≤ s ≤ d− σ. Suppose w ∈W
s,2(Td) andHσ(Σ(w)) > 0.
1. If d−σ2 ≤ s < min(d− σ, d2 + 1), then u = 1w /∈Mq
2 for any q
satisfying 2 ≤ q ≤ dd−s−σ/2 .
2. If s = d2 + 1 ≤ d− σ, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2−σ .
3. If s = d− σ < d2 + 1, then u = 1
w /∈Mq2 for any q.
I Part 3 is sharp, but parts 1 and 2 likely are not.I Based on the results of Jiang, Lin (’03) and Schikorra (’13), I
(we?) conjecture that part 1 holds with 2 ≤ q ≤ d−σd−σ−s .
I Proof uses a version of Poincare Inequality from Jiang, Lin(’03) and Schikorra (’13).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 66: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/66.jpg)
Zero Sets of Larger Hausdorff Dimension
Theorem (Nitzan, M.N., Powell)
Let 0 ≤ σ ≤ d and d−σ2 ≤ s ≤ d− σ. Suppose w ∈W
s,2(Td) andHσ(Σ(w)) > 0.
1. If d−σ2 ≤ s < min(d− σ, d2 + 1), then u = 1w /∈Mq
2 for any q
satisfying 2 ≤ q ≤ dd−s−σ/2 .
2. If s = d2 + 1 ≤ d− σ, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2−σ .
3. If s = d− σ < d2 + 1, then u = 1
w /∈Mq2 for any q.
I Part 3 is sharp, but parts 1 and 2 likely are not.
I Based on the results of Jiang, Lin (’03) and Schikorra (’13), I(we?) conjecture that part 1 holds with 2 ≤ q ≤ d−σ
d−σ−s .I Proof uses a version of Poincare Inequality from Jiang, Lin
(’03) and Schikorra (’13).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 67: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/67.jpg)
Zero Sets of Larger Hausdorff Dimension
Theorem (Nitzan, M.N., Powell)
Let 0 ≤ σ ≤ d and d−σ2 ≤ s ≤ d− σ. Suppose w ∈W
s,2(Td) andHσ(Σ(w)) > 0.
1. If d−σ2 ≤ s < min(d− σ, d2 + 1), then u = 1w /∈Mq
2 for any q
satisfying 2 ≤ q ≤ dd−s−σ/2 .
2. If s = d2 + 1 ≤ d− σ, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2−σ .
3. If s = d− σ < d2 + 1, then u = 1
w /∈Mq2 for any q.
I Part 3 is sharp, but parts 1 and 2 likely are not.I Based on the results of Jiang, Lin (’03) and Schikorra (’13), I
(we?) conjecture that part 1 holds with 2 ≤ q ≤ d−σd−σ−s .
I Proof uses a version of Poincare Inequality from Jiang, Lin(’03) and Schikorra (’13).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 68: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/68.jpg)
Zero Sets of Larger Hausdorff Dimension
Theorem (Nitzan, M.N., Powell)
Let 0 ≤ σ ≤ d and d−σ2 ≤ s ≤ d− σ. Suppose w ∈W
s,2(Td) andHσ(Σ(w)) > 0.
1. If d−σ2 ≤ s < min(d− σ, d2 + 1), then u = 1w /∈Mq
2 for any q
satisfying 2 ≤ q ≤ dd−s−σ/2 .
2. If s = d2 + 1 ≤ d− σ, then u = 1
w /∈Mq2 for any q satisfying
2 ≤ q < 2dd−2−σ .
3. If s = d− σ < d2 + 1, then u = 1
w /∈Mq2 for any q.
I Part 3 is sharp, but parts 1 and 2 likely are not.I Based on the results of Jiang, Lin (’03) and Schikorra (’13), I
(we?) conjecture that part 1 holds with 2 ≤ q ≤ d−σd−σ−s .
I Proof uses a version of Poincare Inequality from Jiang, Lin(’03) and Schikorra (’13).
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 69: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/69.jpg)
Extensions
We have a few variations of these multiplier results
I Multipliers in Mqp for certain ranges of p and q.
I Matrix-weights where W (x) is a K ×K matrix
I Nonsymmetric verisons where the Sobolev smoothness isdifferent in different axis directions.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 70: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/70.jpg)
Extensions
We have a few variations of these multiplier results
I Multipliers in Mqp for certain ranges of p and q.
I Matrix-weights where W (x) is a K ×K matrix
I Nonsymmetric verisons where the Sobolev smoothness isdifferent in different axis directions.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 71: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/71.jpg)
Extensions
We have a few variations of these multiplier results
I Multipliers in Mqp for certain ranges of p and q.
I Matrix-weights where W (x) is a K ×K matrix
I Nonsymmetric verisons where the Sobolev smoothness isdifferent in different axis directions.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 72: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/72.jpg)
Extensions
We have a few variations of these multiplier results
I Multipliers in Mqp for certain ranges of p and q.
I Matrix-weights where W (x) is a K ×K matrix
I Nonsymmetric verisons where the Sobolev smoothness isdifferent in different axis directions.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 73: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/73.jpg)
Table of Contents
Exponentials in Weighted Spaces
Restrictions on Fourier Multipliers
Applications to Balian-Low Type Theorems
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 74: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/74.jpg)
The Balian-Low Theorem
Theorem (Battle (’88) Daubechies, Coifman, Semmes (’90))
Let f ∈ L2(R). If G(f) = {e2πimxf(x− n)}m,n∈Z is a Riesz basisfor L2(R), then(∫
R|x|2|f(x)|2dx
)(∫R|ξ|2|f(ξ)|2dξ
)=∞.
I Conclusion rephrased: “either f /∈ H1(R) or f /∈ H1(R).”
I Assume f, f ∈ H1(R). Smoothness passed to Zf ∈ H1loc(R2).
I Quasiperiodicity of Zf forces it to have a (essential) zero.
I The Riesz basis property forces |Zf | ≥ A > 0, which givescontradiction.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 75: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/75.jpg)
The Balian-Low Theorem
Theorem (Battle (’88) Daubechies, Coifman, Semmes (’90))
Let f ∈ L2(R). If G(f) = {e2πimxf(x− n)}m,n∈Z is a Riesz basisfor L2(R), then(∫
R|x|2|f(x)|2dx
)(∫R|ξ|2|f(ξ)|2dξ
)=∞.
I Conclusion rephrased: “either f /∈ H1(R) or f /∈ H1(R).”
I Assume f, f ∈ H1(R). Smoothness passed to Zf ∈ H1loc(R2).
I Quasiperiodicity of Zf forces it to have a (essential) zero.
I The Riesz basis property forces |Zf | ≥ A > 0, which givescontradiction.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 76: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/76.jpg)
The Balian-Low Theorem
Theorem (Battle (’88) Daubechies, Coifman, Semmes (’90))
Let f ∈ L2(R). If G(f) = {e2πimxf(x− n)}m,n∈Z is a Riesz basisfor L2(R), then(∫
R|x|2|f(x)|2dx
)(∫R|ξ|2|f(ξ)|2dξ
)=∞.
I Conclusion rephrased: “either f /∈ H1(R) or f /∈ H1(R).”
I Assume f, f ∈ H1(R). Smoothness passed to Zf ∈ H1loc(R2).
I Quasiperiodicity of Zf forces it to have a (essential) zero.
I The Riesz basis property forces |Zf | ≥ A > 0, which givescontradiction.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 77: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/77.jpg)
The Balian-Low Theorem
Theorem (Battle (’88) Daubechies, Coifman, Semmes (’90))
Let f ∈ L2(R). If G(f) = {e2πimxf(x− n)}m,n∈Z is a Riesz basisfor L2(R), then(∫
R|x|2|f(x)|2dx
)(∫R|ξ|2|f(ξ)|2dξ
)=∞.
I Conclusion rephrased: “either f /∈ H1(R) or f /∈ H1(R).”
I Assume f, f ∈ H1(R). Smoothness passed to Zf ∈ H1loc(R2).
I Quasiperiodicity of Zf forces it to have a (essential) zero.
I The Riesz basis property forces |Zf | ≥ A > 0, which givescontradiction.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 78: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/78.jpg)
The Balian-Low Theorem
Theorem (Battle (’88) Daubechies, Coifman, Semmes (’90))
Let f ∈ L2(R). If G(f) = {e2πimxf(x− n)}m,n∈Z is a Riesz basisfor L2(R), then(∫
R|x|2|f(x)|2dx
)(∫R|ξ|2|f(ξ)|2dξ
)=∞.
I Conclusion rephrased: “either f /∈ H1(R) or f /∈ H1(R).”
I Assume f, f ∈ H1(R). Smoothness passed to Zf ∈ H1loc(R2).
I Quasiperiodicity of Zf forces it to have a (essential) zero.
I The Riesz basis property forces |Zf | ≥ A > 0, which givescontradiction.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 79: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/79.jpg)
Sharp (Cq)-system BLT
Theorem (Nitzan, M.N, Powell)
Fix q > 2. If G(f, 1, 1) = {e2πimxf(x− n)}m,n∈Z is an exact(Cq)-system for L2(R), then(∫
R|x|4(1−1/q)|f(x)|2dx
)(∫R|ξ|4(1−1/q)|f(ξ)|2dξ
)=∞. (1)
Equivalently, either f /∈ H2(1−1/q)(R) or f /∈ H2(1−1/q)(R).
I Follows from single zero multiplier result.
I Nitzan, Olsen (’11) proved similar result, with an additional εon the weight, as well as non-symmetric versions.
I The q =∞ case gives the BLT for exact systems (originallydue to Daubechies, Janssen (’93)) and nonsymmetric versionswere given by Heil and Powell (’09)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 80: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/80.jpg)
Sharp (Cq)-system BLT
Theorem (Nitzan, M.N, Powell)
Fix q > 2. If G(f, 1, 1) = {e2πimxf(x− n)}m,n∈Z is an exact(Cq)-system for L2(R), then(∫
R|x|4(1−1/q)|f(x)|2dx
)(∫R|ξ|4(1−1/q)|f(ξ)|2dξ
)=∞. (1)
Equivalently, either f /∈ H2(1−1/q)(R) or f /∈ H2(1−1/q)(R).
I Follows from single zero multiplier result.
I Nitzan, Olsen (’11) proved similar result, with an additional εon the weight, as well as non-symmetric versions.
I The q =∞ case gives the BLT for exact systems (originallydue to Daubechies, Janssen (’93)) and nonsymmetric versionswere given by Heil and Powell (’09)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 81: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/81.jpg)
Sharp (Cq)-system BLT
Theorem (Nitzan, M.N, Powell)
Fix q > 2. If G(f, 1, 1) = {e2πimxf(x− n)}m,n∈Z is an exact(Cq)-system for L2(R), then(∫
R|x|4(1−1/q)|f(x)|2dx
)(∫R|ξ|4(1−1/q)|f(ξ)|2dξ
)=∞. (1)
Equivalently, either f /∈ H2(1−1/q)(R) or f /∈ H2(1−1/q)(R).
I Follows from single zero multiplier result.
I Nitzan, Olsen (’11) proved similar result, with an additional εon the weight, as well as non-symmetric versions.
I The q =∞ case gives the BLT for exact systems (originallydue to Daubechies, Janssen (’93)) and nonsymmetric versionswere given by Heil and Powell (’09)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 82: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/82.jpg)
Sharp (Cq)-system BLT
Theorem (Nitzan, M.N, Powell)
Fix q > 2. If G(f, 1, 1) = {e2πimxf(x− n)}m,n∈Z is an exact(Cq)-system for L2(R), then(∫
R|x|4(1−1/q)|f(x)|2dx
)(∫R|ξ|4(1−1/q)|f(ξ)|2dξ
)=∞. (1)
Equivalently, either f /∈ H2(1−1/q)(R) or f /∈ H2(1−1/q)(R).
I Follows from single zero multiplier result.
I Nitzan, Olsen (’11) proved similar result, with an additional εon the weight, as well as non-symmetric versions.
I The q =∞ case gives the BLT for exact systems (originallydue to Daubechies, Janssen (’93)) and nonsymmetric versionswere given by Heil and Powell (’09)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 83: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/83.jpg)
Shift-Invariant Spaces with Extra Invariance
For a given shift-invariant space V = V (f) ⊂ L2(Rd),
I For Γ ⊂ Rd, V is Γ-invariant if TγV ⊂ V for all γ ∈ Γ.
I For any lattice Γ ⊃ Zd, there exists f ∈ L2(Rd) such thatV (f) is precisely Γ-invariant
I d = 1 by Aldroubi, Cabrelli, Heil, Kornelson, Molter (2010)I d > 1 by Anastasio, Cabrelli, Paternostro (2011)
I Aldroubi, Sun, Wang (2011), and Tessera, Wang (2014),showed that Balian-Low type results exist for shift-invariantspaces with extra-invariance.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 84: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/84.jpg)
Shift-Invariant Spaces with Extra Invariance
For a given shift-invariant space V = V (f) ⊂ L2(Rd),
I For Γ ⊂ Rd, V is Γ-invariant if TγV ⊂ V for all γ ∈ Γ.
I For any lattice Γ ⊃ Zd, there exists f ∈ L2(Rd) such thatV (f) is precisely Γ-invariant
I d = 1 by Aldroubi, Cabrelli, Heil, Kornelson, Molter (2010)I d > 1 by Anastasio, Cabrelli, Paternostro (2011)
I Aldroubi, Sun, Wang (2011), and Tessera, Wang (2014),showed that Balian-Low type results exist for shift-invariantspaces with extra-invariance.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 85: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/85.jpg)
Shift-Invariant Spaces with Extra Invariance
For a given shift-invariant space V = V (f) ⊂ L2(Rd),
I For Γ ⊂ Rd, V is Γ-invariant if TγV ⊂ V for all γ ∈ Γ.
I For any lattice Γ ⊃ Zd, there exists f ∈ L2(Rd) such thatV (f) is precisely Γ-invariant
I d = 1 by Aldroubi, Cabrelli, Heil, Kornelson, Molter (2010)I d > 1 by Anastasio, Cabrelli, Paternostro (2011)
I Aldroubi, Sun, Wang (2011), and Tessera, Wang (2014),showed that Balian-Low type results exist for shift-invariantspaces with extra-invariance.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 86: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/86.jpg)
(Cq)-system SIS BLT
Theorem (Nitzan, M.N., Powell)
Fix 2 ≤ q ≤ ∞. Suppose that f ∈ L2(R) is nonzero and V (f) is1NZ-invariant. If T (f) is a minimal (Cq)-system in V (f), then∫
R|x|2(1−1/q) |f(x)|2dx =∞.
Equivalently, f /∈ H1−1/q(R).
I If T (f) is a minimal system for V (f), then T (f) is a(C∞)-system. Thus, the q =∞ case gives us a result forminimal systems.
I (Hardin, M.N., Powell) In the q = 2 case, the result holds inhigher dimensions, and without assuming minimality. (i.e.,frames and not necessarily Riesz bases)
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 87: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/87.jpg)
Minimal (Cq)-result Higher Dimensions
Theorem
Fix q such that 2 ≤ q ≤ ∞, and let s = min(d(12 −1q ) + 1
2 , 1). Let
0 6= f ∈ L2(Rd), and suppose V (f) is invariant under somenon-integer shift. If T (f) is a minimal (Cq)-system for V (f) then∫
Rd|x|2s|f(x)|2dx =∞.
I Can be extended to finitely many generators, requires amatrix-weight version of the Fourier multiplier results.
I Probably the sharp s is 1− 1/q in all dimensions.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 88: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/88.jpg)
Minimal (Cq)-result Higher Dimensions
Theorem
Fix q such that 2 ≤ q ≤ ∞, and let s = min(d(12 −1q ) + 1
2 , 1). Let
0 6= f ∈ L2(Rd), and suppose V (f) is invariant under somenon-integer shift. If T (f) is a minimal (Cq)-system for V (f) then∫
Rd|x|2s|f(x)|2dx =∞.
I Can be extended to finitely many generators, requires amatrix-weight version of the Fourier multiplier results.
I Probably the sharp s is 1− 1/q in all dimensions.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 89: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/89.jpg)
Minimal (Cq)-result Higher Dimensions
Theorem
Fix q such that 2 ≤ q ≤ ∞, and let s = min(d(12 −1q ) + 1
2 , 1). Let
0 6= f ∈ L2(Rd), and suppose V (f) is invariant under somenon-integer shift. If T (f) is a minimal (Cq)-system for V (f) then∫
Rd|x|2s|f(x)|2dx =∞.
I Can be extended to finitely many generators, requires amatrix-weight version of the Fourier multiplier results.
I Probably the sharp s is 1− 1/q in all dimensions.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 90: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/90.jpg)
Where does the zero come from?
I Extra-invarance can be characterized in terms of P f .(Aldroubi, Cabrelli, Heil, Kornelson, Molter (2010), Anastasio,Cabrelli, Paternostro (2011))
I The condition is somewhat technical, so lets look at anexample of f ∈ L2(R2) and V (f) having 1
2Z2-invariance.
P (x) =∑k∈Z2
|f(x− k)|2
=∑k∈Z2
|f(x− 2k)|2 +∑k∈Z2
|f(x− 2k + e1)|2
+∑k∈Z2
|f(x− 2k + e2)|2 +∑k∈Z2
|f(x− 2k + e1 + e2)|2
= P2(x) + P2(x+ e1) + P2(x+ e2) + P2(x+ e1 + e2).
I V (f) is 12Z
2-invariant iff P2(x) and it’s shifts have disjointsupport.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 91: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/91.jpg)
Where does the zero come from?
I Extra-invarance can be characterized in terms of P f .(Aldroubi, Cabrelli, Heil, Kornelson, Molter (2010), Anastasio,Cabrelli, Paternostro (2011))
I The condition is somewhat technical, so lets look at anexample of f ∈ L2(R2) and V (f) having 1
2Z2-invariance.
P (x) =∑k∈Z2
|f(x− k)|2
=∑k∈Z2
|f(x− 2k)|2 +∑k∈Z2
|f(x− 2k + e1)|2
+∑k∈Z2
|f(x− 2k + e2)|2 +∑k∈Z2
|f(x− 2k + e1 + e2)|2
= P2(x) + P2(x+ e1) + P2(x+ e2) + P2(x+ e1 + e2).
I V (f) is 12Z
2-invariant iff P2(x) and it’s shifts have disjointsupport.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 92: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/92.jpg)
Where does the zero come from?
I Extra-invarance can be characterized in terms of P f .(Aldroubi, Cabrelli, Heil, Kornelson, Molter (2010), Anastasio,Cabrelli, Paternostro (2011))
I The condition is somewhat technical, so lets look at anexample of f ∈ L2(R2) and V (f) having 1
2Z2-invariance.
P (x) =∑k∈Z2
|f(x− k)|2
=∑k∈Z2
|f(x− 2k)|2 +∑k∈Z2
|f(x− 2k + e1)|2
+∑k∈Z2
|f(x− 2k + e2)|2 +∑k∈Z2
|f(x− 2k + e1 + e2)|2
= P2(x) + P2(x+ e1) + P2(x+ e2) + P2(x+ e1 + e2).
I V (f) is 12Z
2-invariant iff P2(x) and it’s shifts have disjointsupport.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 93: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/93.jpg)
Where does the zero come from?
I Extra-invarance can be characterized in terms of P f .(Aldroubi, Cabrelli, Heil, Kornelson, Molter (2010), Anastasio,Cabrelli, Paternostro (2011))
I The condition is somewhat technical, so lets look at anexample of f ∈ L2(R2) and V (f) having 1
2Z2-invariance.
P (x) =∑k∈Z2
|f(x− k)|2
=∑k∈Z2
|f(x− 2k)|2 +∑k∈Z2
|f(x− 2k + e1)|2
+∑k∈Z2
|f(x− 2k + e2)|2 +∑k∈Z2
|f(x− 2k + e1 + e2)|2
= P2(x) + P2(x+ e1) + P2(x+ e2) + P2(x+ e1 + e2).
I V (f) is 12Z
2-invariant iff P2(x) and it’s shifts have disjointsupport.
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018
![Page 94: Uncertainty Principles for Fourier Multipliers...Uncertainty Principles for Fourier Multipliers Michael Northington V School of Mathematics Georgia Tech 6/6/2018 With Shahaf Nitzan](https://reader035.fdocuments.in/reader035/viewer/2022071110/5fe4f25c4026ff54850a39e2/html5/thumbnails/94.jpg)
Thanks
Thanks!!!
Uncertainty Principles for Fourier Multipliers M. Northington V ([email protected]) June 6, 2018