Jarek Duda

general recent conversations

Discussion on answer by Roger V.: Wha…

Imported from a comment discussion on physics.stackexchange.co...

Jarek Duda

Jul 17 16:41

Sure in QC we measure and take averages. While measuring some qubits, we can also postselect the remaining qubits, mathematically realized by multiplying them by |0><0|, what enforces them to 0. Physically it is realized by measuring and discarding the non-satisfying ones, but "the system is dissipative and decohering in both temporal directions" suggests alternative realization: we can enforce initial state 0 by waiting before, so what if symmetrically waiting after?

Jarek Duda

Jul 17 16:41

|0> is state preparation enforcing initial state - its consequences are measurable e.g. in quantum computers. And so would be of <0|, which becomes |0> if evolving toward -t. Mathematically it would act as postselection, but with higher success rate - no need to discard cases which are not satisfying.

Jarek Duda

Jul 17 16:41

But doesn't "the system is dissipative and decohering in both temporal directions" mean thermalization acts in both temporal directions? No matter initial conditions, in very low temperature just waiting leads to the ground state: |0> if applied before, so shouldn't give <0| if applied after?

Jarek Duda

Jul 17 16:41

QFT, GR work on 4D scenarios: eternalism/block universe philosophy of time. We could travel through it with e.g. Euler-Lagrange, Dirac, Lindbladian toward +t or -t, but need to remember this is like "watching a movie" of already found 4D solution. If in this movie there is waiting thermalization time in low temperature, toward +t it leads to the ground state, but what about toward -t? Going back to the question in title: "What Lindbladian-like equation should we use to evolve quantum system toward -t ?"

Jarek Duda

Jul 17 16:41

Believing in CPT symmetry, to get <0| we need to prepare conditions such that evolving toward -t we get analogous conditions as for state preparation |0>. Thermalization in very low temperature reservoir dissipates any energy, erasing all the details - no matter initial conditions, it leads to ~ground state. Evolving toward -t looks the same - waiting after unitary evolution, no matter initial conditions, shouldn't it enforce the final state to the ground state?

Jarek Duda

Jul 17 16:41

Sure, if being able to enforce/prepare both initial and final state in classical computer, we could also attack NP problems - by preparing conditions that physics solving through symmetric formulations, like classical least action principle, solves the given instance of problem. Symmetric preparations are e.g.: wait before/after for thermalization, push/pull e.g. superfluid, apply V(t) -> V(-t) reversed shape impulse.

Jarek Duda

Jul 17 16:41

Thermalization/state preparation are viewed as non-unitary. In photonics QC state preparation is performed by injecting photons, what in hydrodynamical analog could be also reversed by connecting microfluidic chip to a pump: both push and pull the liquid through the chip - acting as preparation and postparation if it is superfluid-based quantum computer (4 in th.if.uj.edu.pl/~dudaj/2WQCposter.pdf ).

Jarek Duda

Jul 17 16:41

Sure classical mechanics is symmetric, reversible. The problem is thermodynamics/state preparation - naively asymmetric. But e.g. CPT symmetry, "the system is dissipative and decohering in both temporal directions" suggest they should be also symmetric - if we can prepare/enforce initial state by waiting before, why cannot we symmetrically prepare/enforce final state by waiting after?

Jarek Duda

Jul 17 16:41

There is no square in S-matrix, Born rule comes from symmetry - one amplitude from past, second from future, we need to multiply them - also e.g. in en.wikipedia.org/wiki/Two-state_vector_formalism , or en.wikipedia.org/wiki/Maximal_entropy_random_walk like in Ising. Anyway, if "the system is dissipative and decohering in both temporal directions", shouldn't we be able to influence both e.g. by symmetrically waiting thermalization time before or after?

Jarek Duda

Jul 17 16:41

Isn't asymmetry between them also artificial? E.g. there are symmetric formulations like <psi_f | U | psi_i> in en.wikipedia.org/wiki/S-matrix#Interaction_picture . The big question is if we could build quantum computers on such more symmetric formulation: influencing both initial state e.g. by waiting before in low temperature reservoir, but also symmetrically the final state by waiting thermalization time after?

Jarek Duda

Jul 17 16:41

Yes, we can prepare/enforce initial qubits to |0> by just waiting before in low temperature reservoir - could we symmetrically enforce chosen final qubits <0| by waiting thermalization time after? This way evolving toward -t, should there be still dissipation/thermalization to the ground state? If so, having both pre- and post-paration, here is solving NP problems with such 2WQC: arxiv.org/pdf/2408.05812

Jarek Duda

Jul 17 16:41

I also agree with this fundamental symmetry, but what about this proposed test/realization - we can enforce/prepare |0> initial state by just WAITING thermalization time BEFORE unitary evolution in low temperature of e.g. superconducting QC. But what if symmetrically WAITING thermalization time AFTER unitary evolution? Both "the system is dissipative and decohering in both temporal directions" and evolving toward -t suggest it should symmetrically enforce the final state - act as postparation <0|? If so, we would get 2WQC in theory solving NP problems ...

Discussion on question by Jarek Duda:…

Imported from a comment discussion on physics.stackexchange.co...

Jarek Duda

Jun 27, 2020 09:32

Hence "|pn with electron seeing pn> - |np with electron seeing np>" is not the proper picture, electron in orbital continuously measures 'pn' or 'np', so $np>-|pn>$ hides dynamics between these two

Jarek Duda

Jun 27, 2020 09:29

researchgate.net/profile/Nir_Nevo_Dinur/publication/… - they include quadrupole moment in deuteron atoms

Jarek Duda

Jun 27, 2020 09:22

So maybe atomic physics could tell something here: do we see quadrupole corrections in deuteron atom?

Jarek Duda

Jun 27, 2020 09:20

Ok, we could say that that we have superposition: |pn with electron seeing pn> - |np with electron seeing np> ... but this way none of them sees quadrupole.

Jarek Duda

Jun 27, 2020 09:11

But is quadrupole moment also aligned with spin? (I think so) $np>-|pn>$ means that measuring them, sometime we would get 'pn', sometimes 'np'. But what measurement means? If there is electron in orbital, it would interact electromagnetically, with slight difference for pn and np - isn't it measurement?

Jarek Duda

Jun 25, 2020 17:53

Thank you for the articles - I will have to read and think. Sure, deuteron has spin/magnetic dipole - if I properly understand, in the same direction as quadrupole moment (?). However, to get quadrupole from l=2, m=0 spherical harmonics, if being able to measure position of proton, we would need to get 1/2-1/2 distribution, requiring dynamics as oscillations between 'pn' and 'np.

Jarek Duda

Jun 25, 2020 16:24

I don't think they are the same. Angular momentum usually means DYNAMICS: rotation/spinning/gyroscope in Larmor precession (can we be sure here?), here of ENTIRE PROTON - spherical harmonics usually say about probability distribution (e.g. of finding the particle in measurement, including electrons in orbitals: journals.aps.org/prb/abstract/10.1103/PhysRevB.80.165404 ), not deformation. Shift of charges is STATIC instead, e.g. quark structure DEFORMS to find the lowest energy configuration: shifting a part of charge from proton to neutron, what would e.g. reduce Coulomb energy.

Jarek Duda

Jun 25, 2020 16:24

When electron approaches proton to become a probability cloud, does it loose dynamics this way? Here, to explain quadrupole, we would like to imagine proton as probability cloud in both sides of neutron - again: doesn't it mean dynamics, especially that angular momentum means dynamics? But most importantly, I still didn't hear a single argument in "angular momentum vs shifts of charges" question - is there any? If not, where certainty in the former comes from?

Jarek Duda

Jun 25, 2020 16:24

In other words: this is the most fundamental question of nuclear physics - of understanding the binding mechanism. Should we be really focused only on one option, especially that it is a bit controversial if thinking about it, and there are possible less magical ways? Why not explore both?

Jarek Duda

Jun 25, 2020 16:24

But I have problem with claiming bilocation for proton, with angular momentum e.g. of photons we can rotate macroscopic objects - this is real angular momentum, dynamics - are there experimental arguments that it is also real angular momentum for deuteron? And such controversial claims are not needed - qudrupole can be get without magic by shift of charges like quarks, why it is not considered? Let me add diagram with such mechanism suggested by soliton particle models.

Jarek Duda

Jun 25, 2020 16:24

Isn't it just cheating allowing us to get any multipole we want from trivial 'pn' charge configuration? So we don't understand why/how it has angular momentu moment, only fit the model? Couldn't we instead e.g. get down into quark configuration, splitting into distant two positive fractional charges (quadrupole) to minimize Coulomb energy - bringing some understanding without cheating? I don't understand why people seem to be certain that this angular momentum mathematical trick is the only proper way?

theory salon

theoretical computer science. highlight reel vzn1.wordpress.co...

Jarek Duda

Jan 15, 2019 05:35

@vzn, I work with ML group for a few years

Jarek Duda

Jan 14, 2019 06:34

@vzn, I honestly don't know standard methods able to handle such basic looking problem

Jarek Duda

Jan 12, 2019 16:15

Any standard tools to for looking basic statistics task? stats.stackexchange.com/questions/386848/…

Jarek Duda

Sep 27, 2018 10:54

Q: Space of change of basis matrices between two similar matrices - how to reduce it with additional tests?

Assume we have two real symmetric $n\times n$ matrices: $A, B$. We can easily test their similarity: $\textrm{Tr}(A^k)=\textrm{Tr}(B^k)$ for $k=1..n$. In this case both can be rotated to the same diagonal matrix $D=\textrm{diag}(\{\lambda_i\})$ made of their eigenvalues - there exist orthogonal $...

Jarek Duda

Sep 27, 2018 10:54

Ok let's try mathoverflow:

Jarek Duda

Sep 27, 2018 07:43

Here are all the files: dropbox.com/sh/e4vyf5njr48cq71/AABjOckfR-d6sdbr83dPMKOFa?dl=0

Jarek Duda

Sep 27, 2018 07:43

The problem is characterizing this restriction, to choose family restricting to permutations only - solving the problem

Jarek Duda

Sep 27, 2018 07:42

So we need complementing tests to restrict the space of possible O matrices, hopefully down to permutations - and finally after a year of (among others) working on this I am able to distinguish the toughest cases this way: SRGs

Jarek Duda

Sep 27, 2018 07:40

Testing similarity is a simple example ensuring orthogonal O between two matrices - there remains question if there is permutation among them

Jarek Duda

Sep 27, 2018 07:38

But generally I posted it in Mathematics as I believe algorithmic treatment, way of thinking is useful for this problem - instead of local vertex descriptions with canonical labelization, use global descriptors/invariants - to get complete description modulo permutation

Jarek Duda

Sep 27, 2018 07:36

@ThomasKlimpel, indeed this general permutation test is for graph with colored edges - I don't know if it can be polynomially converted to standard GI?

Jarek Duda

Sep 25, 2018 16:05

We need operation which is invariant for permutation, but not for other orthogonal operations - and I finally got a simple one which distinguishes at least some SRGs ... the problem is showing that they provide sufficient condition

Jarek Duda

Sep 25, 2018 16:02

Thanks, I will think about. Generally hard examples for standard approach (Weisfeiler-Leman) are different than for what I do - e.g. now trying to restrict the space of possible similarity matrices between them to only permutation.

Jarek Duda

Sep 25, 2018 10:07

@ThomasKlimpel, simplified invariants above - directly for adjacency matrix, distinguish SRGs in tests

Jarek Duda

Sep 25, 2018 10:06

Q: Testing if two matrices differ only by permutation? (generalized graph isomorphism problem)

In graph isomorphism problem, for which Babai's quasi polynomial algorithm is currently under review (stack), we ask if two adjacency matrices: of $\{0,1\}$ coefficients differ only by a permutation. Is general problem: without coefficient restriction more difficult, or we can also assume that ...

linear-algebra matrices matrix-equations graph-isomorphism invariance

Jarek Duda

Sep 25, 2018 07:30

space quantum force? :)

Jarek Duda

Sep 20, 2018 09:33

Q: Testing if two finite sets of points differ only by rotation (in polynomial time)?

Imagine we have two size $m$ sets of points $X=\{x^i\}_{i=1..m}, Y=\{y^i\}_{i=1..m} \subset \mathbb{R}^n$ and we need to answer the question if they differ only by rotation: if there exists othogonal $O$ such that $X=OY$. Can it be answered in polynomial time? While it sounds simple, it isn't - ...

linear-algebra geometry algebraic-geometry graph-isomorphism invariance

Jarek Duda

Sep 20, 2018 06:00

@ThomasKlimpel, I have returned to graph isomorphism problem and now nicely distinguish tested strongly regular graphs with invariants: encode.ru/threads/…

Jarek Duda

Sep 3, 2018 05:27

it lasts more than a year - I have tried everything encode.ru/threads/2648-Published-rANS-patent-by-Storeleap/page3

Jarek Duda

Sep 2, 2018 20:22

2 months ago there was ~25k: old.reddit.com/r/programming/duplicates/8q3kp8/… ... and nothing - they will probably ignore again and continue - clear plagiarism is no longer a problem for them

Jarek Duda

Aug 30, 2018 08:09

if anyone can retweet this: twitter.com/fsf/status/1034181158429515778 ... we need a clear statement that big companies cannot just patent our work behind our back ...

Jarek Duda

Aug 25, 2018 07:19

I have added section abut Bell violation with MERW to 0910.2724 - thread: sciphysicsforums.com/spfbb1/viewtopic.php?f=6&t=360

Jarek Duda

Aug 12, 2018 20:28

... and generally there are video presentations available: emqm17.org/presentations emqm15.org/presentations/speaker-presentations emqm13.org/abstracts

Ten fold

CrossValidated's general room for gossip, grumbles, and idle c...

Jarek Duda

Jan 13, 2019 08:32

Jarek Duda

Jan 13, 2019 08:32

i.sstatic.net/PiouI.png - some pairwise densities - we first need to to model them ... then combine predictions from such multiple variables, also discrete ...

Jarek Duda

Jan 13, 2019 08:29

Looking really basic statistics/ML problem - evaluate (conditional continuous) probability distribution based on variables of various types ... I have some working approach, but need to compare it with something standard - what would be appropriate?

Jarek Duda

Jan 13, 2019 08:27

Q: Credibility evaluation - how to model conditional continuous density from multiple variables of various types?

I recently got dataset for 37000 households with declared income and a few dozens of other variables of various types: continuous, discrete, binary. The task is to automatically (unsupervised) evaluate credibility of declared income based on the remaining variables: evaluate if it agrees with st...

estimation conditional-probability nonparametric pdf density-estimation