arXiv subjects

Michał Studziński

Publications and source records attributed to Michał Studziński.

48 records · Page 3Linked to original sources

arXiv2015

Towards fully quantum second laws of thermodynamics: limitations on the evolution of quantum coherences

The second law of thermodynamics places a limitation into which states a system can evolve into. For systems in contact with a heat bath, it can be combined with the law of energy conservation, and it says that a system can only evolve into another if the free energy goes down. Recently, it's been shown that there are actually many second laws, and that it is only for large macroscopic systems that they all become equivalent to the ordinary one. These additional second laws also hold for quantum systems, and are, in fact, often more relevant in this regime. They place a restriction on how the probabilities of energy levels can evolve. Here, we consider additional restrictions on how the coherences between energy levels can evolve. Coherences can only go down, and we provide a set of restrictions which limit the extent to which they can be maintained. We find that coherences over energy levels must decay at rates that are suitably adapted to the transition rates between energy levels. We show that the limitations are matched in the case of a single qubit, in which case we obtain the full characterization of state-to-state transformations. For higher dimensions, we conjecture that more severe constraints exist. We also introduce a new class of thermodynamical operations which allow for greater manipulation of coherences and study its power with respect to a class of operations known as thermal operations.

quant-ph↗

arXiv2015

Separability Properties for a Class of Block Matrices

It is shown that, for the block matrices belonging to $M(nd,\mathbb{C})$ with commuting and normal block entries of dimension $d$, the separability of such a block matrices is equivalent to its semi-positive definity. The separability decomposition of lenght equal to the dimension of the block matrix (which is smaller then Carathéodory theorem implies) is given. The separability decomposition depends only on eigenvalues of block entries in the first part and on eigenvectors of the block entries in the second part of the tensor product. It is shown that semi-positive definity of considered block matrices is equivalent to semi-positive definity $d$ smaller matrices of dimension $n$.

quant-ph↗

arXiv2015

Using non-positive maps to characterize entanglement witnesses

In this paper we present a new method for entanglement witnesses construction. We show that to construct such an object we can deal with maps which are not positive on the whole domain, but only on a certain sub-domain. In our approach crucial role play such maps which are surjective between sets $\mathcal{P}_{k}^d$ of $k \leq d$ rank projectors and the set $\mathcal{P}_1^d$ of rank one projectors acting in the $d$ dimensional space. We argue that our method can be used to check whether a given observable is an entanglement witness. In the second part of this paper we show that inverse reduction map satisfies this requirement and using it we can obtain a bunch of new entanglement witnesses.

quant-ph↗

arXiv2015

Construction and properties of a class of private states in arbitrary dimensions

We present a construction of quantum states in dimension $d$ that has at least 1 dit of ideal key, called private dits (pdits), which covers most of the known examples of private bits (pbits) $d=2$. We examine properties of this class of states, focusing mostly on its distance to the set of separable states $\mathcal{SEP}$, showing that for a fixed dimension of key part $d_k$ the distance increases with $d_s$. We provide explicit examples of PPT states (in $d$ dimensions) which are nearly as far from separable ones as possible. Precisely, the distance from the set of $\mathcal{SEP}$ is $2 - ε$, where $d$ scales with $ε$ as $d \propto 1/ε^3$, as opposed to $d \propto 2^{(log(4/ε))^2}$ obtained in [Badzicag et al., Phys. Rev. A 90, 012301 (2014)]. We do not use boosting (taking many copies of pdits to boost the distance) as in Badzicag et al. paper.

quant-ph↗

arXiv2014

Explicit constructions of unitary transformations between equivalent irreducible representations

Irreducible representations (irreps) of a finite group $G$ are equivalent if there exists a similarity transformation between them. In this paper, we describe an explicit algorithm for constructing this transformation between a pair of equivalent irreps, assuming we are given an algorithm to compute the matrix elements of these irreps. Along the way, we derive a generalization of the classical orthogonality relations for matrix elements of irreps of finite groups. We give an explicit form of such unitary matrices for the important case of conjugated Young-Yamanouchi representations, when our group $G$ is symmetric group $S(N)$.

math.RT↗

arXiv2014

Group representation approach to 1 - N universal quantum cloning machines

In this work, we revisit the problem of finding an admissible region of fidelities obtained after an application of an arbitrary $1 \rightarrow N$ universal quantum cloner which has been recently solved in [A. Kay et al., Quant. Inf. Comput 13, 880 (2013)] from the side of cloning machines. Using group-theory formalism, we show that the allowed region for fidelities can be alternatively expressed in terms of overlaps of pure states with recently found irreducible representations of the commutant $U \otimes U \otimes \ldots \otimes U \otimes U^*$, which gives the characterization of the allowed region where states being cloned are figure of merit. Additionally, it is sufficient to take pure states with real coefficients only, which makes calculations simpler. To obtain the allowed region, we make a convex hull of possible ranges of fidelities related to a given irrep. Subsequently, two cases: $1 \rightarrow 2$ and $1 \rightarrow 3$ cloners, are studied for different dimensions of states as illustrative examples.

quant-ph↗

arXiv2014

Structure and properties of the algebra of partially transposed permutation operators

We consider the structure of algebra of operators, acting in $n-$fold tensor product space, which are partially transposed on the last term. Using purely algebraical methods we show that this algebra is semi-simple and then, considering its regular representation, we derive basic properties of the algebra. In particular, we describe all irreducible representations of the algebra of partially transposed operators and derive expressions for matrix elements of the representations. It appears that there are two types of irreducible representations of the algebra. The first one is strictly connected with the representations of the group $S(n-1)$ induced by irreducible representations of the group $S(n-2)$. The second type is structurally connected with irreducible representations of the group $S(n-1)$.

quant-ph↗

arXiv2013

Commutant structuture of Ux...xUxU* transformations

In this paper we have found irreducible representations (irreps) of the algebra of partially transposed permutation operators on last subsystem. We give here direct method of irreps construction for an arbitrary number of subsystems n and local dimension d. Our method is inspired by representation theory of symmetric group S(n), theory of Brauer Algebras and Walled Brauer Algebras.

quant-ph↗

arXiv2013

Local random quantum circuits are approximate polynomial-designs - numerical results

We numerically investigate the statement that local random quantum circuits acting on n qubits composed of polynomially many nearest neighbour two-qubit gates form an approximate unitary poly(n)-design [F.G.S.L. Brandao et al., arXiv:1208.0692]. Using a group theory formalism, spectral gaps that give a ratio of convergence to a given t-design are evaluated for a different number of qubits n (up to 20) and degrees t (t=2,3,4 and 5), improving previously known results for n=2 in the case of t=2 and 3. Their values lead to a conclusion that the previously used lower bound that bounds spectral gaps values may give very little information about the real situation and in most cases, only tells that a gap is closed. We compare our results to the another lower bounding technique, again showing that its results may not be tight.

quant-ph↗

arXiv2012

Region of fidelities for a 1 -> N universal qubit quantum cloner

We analyze a region of fidelities for qubit which is obtained after an application of a 1 -> N universal quantum cloner. We express the allowed region for fidelities in terms of overlaps of pure states with irreps of S(n) (n = N+1) showing that the pure states can be taken with real coefficients only. Subsequently, the case n = 4, corresponding to a 1 -> 3 cloner is studied in more detail as an illustrative example. To obtain the main result, we make a convex hull of possible ranges of fidelities related to a given irrep. The formalism allows to construct the state giving rise to a given N-tuple of fidelities.

quant-ph↗

arXiv2012

Darboux Points and Integrability Analysis of Hamiltonian Systems with Homogeneous Rational Potentials

We study the integrability in the Liouville sense of natural Hamiltonian systems with a homogeneous rational potential $V(\vq)$. Strong necessary conditions for the integrability of such systems were obtained by an analysis of differential Galois group of variational equations along certain particular solutions. These conditions have the form of arithmetic restrictions putted on eigenvalues of Hessian $V"(\vd)$ calculated at a non-zero solution $\vd$ of equation $\grad V(\vd)=\vd$. Such solutions are called proper Darboux points. It was recently proved that for generic polynomial homogeneous potentials there exist universal relations between eigenvalues of Hessians of the potential taken at all proper Darboux points. The question about the existence of such relations for rational potentials seems to be hard. One of the reason of this fact is the presence of indeterminacy points of the potential and its gradient. Nevertheless, for two degrees of freedom we prove that such relation exists. This result is important because it allows to show that the set of admissible values for eigenvalues of Hessian at a proper Darboux point for potentials satisfying necessary conditions for the integrability is finite. In turn, it gives a tool for classification of integrable rational potentials. Also, quite recently, it was shoved that for polynomial homogeneous potentials additional necessary conditions for the integrability can be deduced from the existence of improper Darboux points, i.e., points $\vd$ which are non-zero solution of equation $\grad V(\vd)=\vzero$. These new conditions have also the form of arithmetic restrictions imposed on eigenvalues of $V"(\vd)$. In this paper we prove that for rational potentials improper Darboux points give the same necessary conditions for the integrability.

nlin.SI↗

arXiv2011

Distillation of entanglement by projection on permutationally invariant subspaces

We consider distillation of entanglement from two qubit states which are mixtures of three mutually orthogonal states: two pure entangled states and one pure product state. We distill entanglement from such states by projecting n copies of the state on permutationally invariant subspace and then applying one-way hashing protocol. We find analytical expressions for the rate of the protocol. We also generalize this method to higher dimensional systems. To get analytical expression for two qubit case, we faced a mathematical problem of diagonalizing a family of matrices enjoying some symmetries w.r.t. to symmetric group. We have solved this problem in two ways: (i) directly, by use of Schur-Weyl decomposition and Young symmetrizers (ii) showing that the problem is equivalent to a problem of diagonalizing adjacency matrices in a particular instance of a so called algebraic association scheme.

quant-ph↗