5,532 reputation
1425
bio website regularize.wordpress.com
location
age
visits member for 3 years, 9 months
seen Aug 24 at 19:01

Jul
25
comment Proximal operator, scaling by a matrix
What I suspected was, that the proximal operator is not really the object you are interested in. But I may also be wrong...
Jul
25
comment Proximal operator, scaling by a matrix
Are you sure that you need to do this? If you are about to solve a larger optimization problem you may consider the possibility of "dualizing the term with the linear operator" or introducing a new varaible $Y=PX$ and add this equality as a constraint. Both ways make your problem larger but simpler in some sense.
Jul
24
answered Coupling methods
Jul
24
comment Coupling methods
Actually, the Wasserstein distance also makes sense for two probability measures defined on different metric spaces.
Jul
5
revised Best textbook for Geometric Measure Theory
fixed typo, extended a bit
Jul
5
answered Best textbook for Geometric Measure Theory
Jun
26
awarded  Popular Question
Jun
16
answered Why am I getting a contradiction?
Jun
14
comment Using the identity theorem to prove existence of non-identically zero function
Funny title, by the way…
Jun
13
awarded  Good Question
Jun
5
awarded  Notable Question
Jun
1
comment Why isn't math completely solved by expert systems?
Why do people still write novels by themselves if computers can produce all meaningful sequences of word?
Jun
1
answered Why does nobody use the deduction symbols $A \Rightarrow B \Rightarrow C \Rightarrow D$ in their thesis/dissertation?
May
16
comment the delta function written as the integral of a complex number
@Leucippus In this case not since the relation involves both forward and inverse Fourier transform.
May
16
comment the delta function written as the integral of a complex number
I would say the equation "holds in the weak sense" (sense of distributions to be precise), i.e. "multiply by some $f(t)$ and integrate".
May
15
comment random nonsingular matrices using matlab
Also rand(n) and randn(n) will give nonsingular matrices with probability 1 (even the condition numbers will be not very large.).
May
10
comment Standards for publishing research on Global Optimization
Not sure if on topic here… Anyway: Why don't you use quasi-Newton right away? Moreover: Sounds like a standard approach (first use a globally convergent, slow method and then switch to a Newton-type thing). The content that you describe sound like barely publishable. At least do a thorough literature search on global optimization methods.
May
8
comment Using the Modulation property of the Fourier Transform
Convolving with a delta is indeed a pretty simple task!
Apr
26
revised Is it possible to uniquely solve $\mathbf{A}\vec{x}=\vec{b}$ for $\mathbf{A}$ (given $\vec{b}$ and $\vec{x}$)?
edited body
Apr
25
revised Is it possible to uniquely solve $\mathbf{A}\vec{x}=\vec{b}$ for $\mathbf{A}$ (given $\vec{b}$ and $\vec{x}$)?
added 116 characters in body