arXiv subjects

Armen E. Allahverdyan

Publications and source records attributed to Armen E. Allahverdyan.

At least 55 records · Page 3Linked to original sources

arXiv2010

Community Detection with and without Prior Information

We study the problem of graph partitioning, or clustering, in sparse networks with prior information about the clusters. Specifically, we assume that for a fraction $ρ$ of the nodes their true cluster assignments are known in advance. This can be understood as a semi--supervised version of clustering, in contrast to unsupervised clustering where the only available information is the graph structure. In the unsupervised case, it is known that there is a threshold of the inter--cluster connectivity beyond which clusters cannot be detected. Here we study the impact of the prior information on the detection threshold, and show that even minute [but generic] values of $ρ>0$ shift the threshold downwards to its lowest possible value. For weighted graphs we show that a small semi--supervising can be used for a non-trivial definition of communities.

physics.soc-ph↗

arXiv2010

Optimal refrigerator

We study a refrigerator model which consists of two $n$-level systems interacting via a pulsed external field. Each system couples to its own thermal bath at temperatures $T_h$ and $T_c$, respectively ($θ\equiv T_c/T_h<1$). The refrigerator functions in two steps: thermally isolated interaction between the systems driven by the external field and isothermal relaxation back to equilibrium. There is a complementarity between the power of heat transfer from the cold bath and the efficiency: the latter nullifies when the former is maximized and {\it vice versa}. A reasonable compromise is achieved by optimizing the product of the heat-power and efficiency over the Hamiltonian of the two system. The efficiency is then found to be bounded from below by $ζ_{\rm CA}=\frac{1}{\sqrt{1-θ}}-1$ (an analogue of the Curzon-Ahlborn efficiency), besides being bound from above by the Carnot efficiency $ζ_{\rm C} = \frac{1}{1-θ}-1$. The lower bound is reached in the equilibrium limit $θ\to 1$. The Carnot bound is reached (for a finite power and a finite amount of heat transferred per cycle) for $\ln n\gg 1$. If the above maximization is constrained by assuming homogeneous energy spectra for both systems, the efficiency is bounded from above by $ζ_{\rm CA}$ and converges to it for $n\gg 1$.

cond-mat.stat-mech↗

arXiv2010

Thermodynamics of enhanced heat transfer: a model study

Situations where a spontaneous process of energy or matter transfer is enhanced by an external device are widespread in nature (human sweating system, enzyme catalysis, facilitated diffusion across bio-membranes, industrial heat exchangers). The thermodynamics of such processes remains however open. Here we study enhanced heat transfer by a model junction immersed between two thermal baths at different temperatures $T_h$ and $T_c$ ($T_h>T_c$). The transferred heat power is enhanced via controlling the junction by means of external time-dependent fields. Provided that the spontaneous heat flow process is optimized over the junction Hamiltonian, any enhancement of this spontaneous process does demand consumption and subsequent dissipation of work. The efficiency of enhancement is defined via the increment in the heat power divided over the amount of consumed work. We show that this efficiency is bounded from above by $T_c/(T_h-T_c)$. Formally this is identical to the Carnot bound for the efficiency of ordinary refrigerators which transfer heat from cold to hot. It also shares some (but not all) physical features of the Carnot bound.

cond-mat.stat-mech↗

arXiv2010

Transferring elements of a density matrix

We study restrictions imposed by quantum mechanics on the process of matrix elements transfer. This problem is at the core of quantum measurements and state transfer. Given two systems $\A$ and $\B$ with initial density matrices $λ$ and $r$, respectively, we consider interactions that lead to transferring certain matrix elements of unknown $λ$ into those of the final state ${\widetilde r}$ of $\B$. We find that this process eliminates the memory on the transferred (or certain other) matrix elements from the final state of $\A$. If one diagonal matrix element is transferred, ${\widetilde r}_{aa}=λ_{aa}$, the memory on each non-diagonal element $λ_{a\not=b}$ is completely eliminated from the final density operator of $\A$. Consider the following three quantities $\Re \la_{a\not =b}$, $\Im \la_{a\not =b}$ and $\la_{aa}-\la_{bb}$ (the real and imaginary part of a non-diagonal element and the corresponding difference between diagonal elements). Transferring one of them, e.g., $\Re\tir_{a\not = b}=\Re\la_{a\not = b}$, erases the memory on two others from the final state of $\A$. Generalization of these set-ups to a finite-accuracy transfer brings in a trade-off between the accuracy and the amount of preserved memory. This trade-off is expressed via system-independent uncertainty relations which account for local aspects of the accuracy-disturbance trade-off in quantum measurements.

quant-ph↗

arXiv2009

Thermodynamic efficiency of information and heat flow

A basic task of information processing is information transfer (flow). Here we study a pair of Brownian particles each coupled to a thermal bath at temperature $T_1$ and $T_2$, respectively. The information flow in such a system is defined via the time-shifted mutual information. The information flow nullifies at equilibrium, and its efficiency is defined as the ratio of flow over the total entropy production in the system. For a stationary state the information flows from higher to lower temperatures, and its the efficiency is bound from above by $\frac{{\rm max}[T_1,T_2]}{|T_1-T_2|}$. This upper bound is imposed by the second law and it quantifies the thermodynamic cost for information flow in the present class of systems. It can be reached in the adiabatic situation, where the particles have widely different characteristic times. The efficiency of heat flow|defined as the heat flow over the total amount of dissipated heat|is limited from above by the same factor. There is a complementarity between heat- and information-flow: the setup which is most efficient for the former is the least efficient for the latter and {\it vice versa}. The above bound for the efficiency can be [transiently] overcome in certain non-stationary situations, but the efficiency is still limited from above. We study yet another measure of information-processing [transfer entropy] proposed in literature. Though this measure does not require any thermodynamic cost, the information flow and transfer entropy are shown to be intimately related for stationary states.

cond-mat.stat-mech↗

arXiv2009

Optimal thermal refrigerator

We study a refrigerator model which consists of two $n$-level systems interacting via a pulsed external field. Each system couples to its own thermal bath at temperatures $T_h$ and $T_c$, respectively ($θ\equiv T_c/T_h<1$). The refrigerator functions in two steps: thermally isolated interaction between the systems driven by the external field and isothermal relaxation back to equilibrium. There is a complementarity between the power of heat transfer from the cold bath and the efficiency: the latter nullifies when the former is maximized and {\it vice versa}. A reasonable compromise is achieved by optimizing over the inter-system interaction and intra-system energy levels the product of the heat-power and efficiency. The efficiency is then found to be bounded from below by $ζ_{\rm CA}=\frac{1}{\sqrt{1-θ}}-1$ (an analogue of Curzon-Ahlborn efficiency for refrigerators), besides being bound from above by the Carnot efficiency $ζ_{\rm C} = \frac{1}{1-θ}-1$. The lower bound is reached in the equilibrium limit $θ\to 1$, while the Carnot bound is reached (for a finite power and a finite amount of heat transferred per cycle) in the macroscopic limit $\ln n\gg 1$. The efficiency is exactly equal to $ζ_{\rm CA}$, when the above optimization is constrained by assuming homogeneous energy spectra for both systems.

cond-mat.stat-mech↗

arXiv2009

Kinetics of the helix-coil transition

Based on the Zimm-Bragg model we study cooperative helix-coil transition driven by a finite-speed change of temperature. There is an asymmetry between the coil-to-helix and helix-to-coil transition: the latter is displayed already for finite speeds, and takes shorter time than the former. This hysteresis effect has been observed experimentally, and it is explained here via quantifying system's stability in the vicinity of the critical temperature. A finite-speed cooling induces a non-equilibrium helical phase with the correlation length larger than in equilibrium. In this phase the characteristic length of the coiled domain and the non-equilibrium specific heat can display an anomalous response to temperature changes. Several pertinent experimental results on the kinetics helical biopolymers are discussed in detail.

cond-mat.soft↗

arXiv2009

Replicators in Fine-grained Environment: Adaptation and Polymorphism

Selection in a time-periodic environment is modeled via the two-player replicator dynamics. For sufficiently fast environmental changes, this is reduced to a multi-player replicator dynamics in a constant environment. The two-player terms correspond to the time-averaged payoffs, while the three and four-player terms arise from the adaptation of the morphs to their varying environment. Such multi-player (adaptive) terms can induce a stable polymorphism. The establishment of the polymorphism in partnership games [genetic selection] is accompanied by decreasing mean fitness of the population.

q-bio.PE↗

arXiv2008

Quantum restrictions on transfer of matrix elements

We discuss restrictions imposed by quantum mechanics on the process of matrix elements transfer from the one system to another. This is relevant for various processes of partial state transfer (quantum communication, indirect measurement, polarization transfer, {\it etc}). Given two systems A and B with initial density operators $λ$ and $r$, respectively, we consider most general interactions, which lead to transferring certain matrix elements of unknown $λ$ into those of the final state ${\widetilde r}$ of B. We find that this process leads to eliminating the memory on the transferred (or certain other) matrix elements from the final state of A. If one diagonal matrix element is transferred: ${\widetilde r}_{aa}=λ_{aa}$, the memory on each non-diagonal element $λ_{a\not=b}$ is completely eliminated from the final density operator of A. The transfer of a non-diagonal element: ${\widetilde r}_{ab}=λ_{ab}$ eliminates the memory on the diagonal elements $λ_{aa}$ and $λ_{bb}$, while the memory about their sum $λ_{aa}+λ_{bb}$ is kept. Moreover, the memory about $λ_{ab}$ itself is completely eliminated from the final state of A. Generalization of these set-ups to non-ideal transfer brings in a trade-off between the quality of the transfer and the amount of preserved memory. This trade-off is expressed via system-independent uncertainty relations.

quant-ph↗

arXiv2008

Entropy of Hidden Markov Processes via Cycle Expansion

Hidden Markov Processes (HMP) is one of the basic tools of the modern probabilistic modeling. The characterization of their entropy remains however an open problem. Here the entropy of HMP is calculated via the cycle expansion of the zeta-function, a method adopted from the theory of dynamical systems. For a class of HMP this method produces exact results both for the entropy and the moment-generating function. The latter allows to estimate, via the Chernoff bound, the probabilities of large deviations for the HMP. More generally, the method offers a representation of the moment-generating function and of the entropy via convergent series.

cs.IT↗

arXiv2008

Employing feedback in adiabatic quantum dynamics

We study quantum adiabatic dynamics, where the slowly moving field is influenced by system's state (feedback). The information for the feedback is gained from non-disturbating measurements done on an ensemble of identical non-interacting systems. The situation without feedback is governed by the adiabatic theorem: adiabatic energy level populations stay constant, while the adiabatic eigenvectors get a specific phase contribution (Berry phase). However, under feedback the adiabatic theorem does not hold: the adiabatic populations satisfy a closed equation of motion that coincides with the replicator dynamics well-known by its applications in evolutionary game theory. The feedback generates a new gauge-invariant adiabatic phase, which is free of the constraints on the Berry phase (e.g., the new phase is non-zero even for real adiabatic eigenfunctions).

quant-ph↗

arXiv2007

Work extremum principle: Structure and function of quantum heat engines

We consider a class of quantum heat engines consisting of two subsystems interacting via a unitary transformation and coupled to two separate baths at different temperatures $T_h > T_c$. The purpose of the engine is to extract work due to the temperature difference. Its dynamics is not restricted to the near equilibrium regime. The engine structure is determined by maximizing the extracted work under various constraints. When this maximization is carried out at finite power, the engine dynamics is described by well-defined temperatures and satisfies the local version of the second law. In addition, its efficiency is bounded from below by the Curzon-Ahlborn value $1-\sqrt{T_c/T_h}$ and from above by the Carnot value $1-(T_c/T_h)$. The latter is reached|at finite power|for a macroscopic engine, while the former is achieved in the equilibrium limit $T_h\to T_c$. When the work is maximized at a zero power, even a small (few-level) engine extracts work right at the Carnot efficiency.

cond-mat.stat-mech↗

arXiv2007

Relating the thermodynamic arrow of time to the causal arrow

Consider a Hamiltonian system that consists of a slow subsystem S and a fast subsystem F. The autonomous dynamics of S is driven by an effective Hamiltonian, but its thermodynamics is unexpected. We show that a well-defined thermodynamic arrow of time (second law) emerges for S whenever there is a well-defined causal arrow from S to F and the back-action is negligible. This is because the back-action of F on S is described by a non-globally Hamiltonian Born-Oppenheimer term that violates the Liouville theorem, and makes the second law inapplicable to S. If S and F are mixing, under the causal arrow condition they are described by microcanonic distributions P(S) and P(S|F). Their structure supports a causal inference principle proposed recently in machine learning.

cond-mat.stat-mech↗

arXiv2007

The Quantum Measurement Process: Lessons from an Exactly Solvable Model

The measurement of a spin-$\half$ is modeled by coupling it to an apparatus, that consists of an Ising magnetic dot coupled to a phonon bath. Features of quantum measurements are derived from the dynamical solution of the measurement, regarded as a process of quantum statistical mechanics. Schrödinger cat terms involving both the system and the apparatus, die out very quickly, while the registration is a process taking the apparatus from its initially metastable state to one of its stable final states. The occurrence of Born probabilities can be inferred at the macroscopic level, by looking at the pointer alone. Apparent non-unitary behavior of the measurement process is explained by the arisal of small many particle correlations, that characterize relaxation.

quant-ph↗

arXiv2005

Phase transitions and quantum measurements

In a quantum measurement, a coupling $g$ between the system S and the apparatus A triggers the establishment of correlations, which provide statistical information about S. Robust registration requires A to be macroscopic, and a dynamical symmetry breaking of A governed by S allows the absence of any bias. Phase transitions are thus a paradigm for quantum measurement apparatuses, with the order parameter as pointer variable. The coupling $g$ behaves as the source of symmetry breaking. The exact solution of a model where S is a single spin and A a magnetic dot (consisting of $N$ interacting spins and a phonon thermal bath) exhibits the reduction of the state as a relaxation process of the off-diagonal elements of S+A, rapid due to the large size of $N$. The registration of the diagonal elements involves a slower relaxation from the initial paramagnetic state of A to either one of its ferromagnetic states. If $g$ is too weak, the measurement fails due to a ``Buridan's ass'' effect. The probability distribution for the magnetization then develops not one but two narrow peaks at the ferromagnetic values. During its evolution it goes through wide shapes extending between these values.

quant-ph↗

arXiv2004

Dynamics of a quantum measurement

We work out an exactly solvable hamiltonian model which retains all the features of realistic quantum measurements. In order to use an interaction process involving a system and an apparatus as a measurement, it is necessary that the apparatus is macroscopic. This implies to treat it with quantum statistical mechanics. The relevant time scales of the process are exhibited. It begins with a very rapid disappearance of the off-diagonal blocks of the overall density matrix of the tested system and the apparatus. Possible recurrences are hindered by the large size of the latter. On a much larger time scale the apparatus registers the outcome: Correlations are established between the final values of the pointer and the initial diagonal blocks of the density matrix of the tested system. We thus derive Born's rule and von Neumann's reduction of the state from the dynamical process.

quant-ph↗

arXiv2004

The quantum measurement process in an exactly solvable model

An exactly solvable model for a quantum measurement is discussed which is governed by hamiltonian quantum dynamics. The $z$-component $\hat s_z$ of a spin-1/2 is measured with an apparatus, which itself consists of magnet coupled to a bath. The initial state of the magnet is a metastable paramagnet, while the bath starts in a thermal, gibbsian state. Conditions are such that the act of measurement drives the magnet in the up or down ferromagnetic state according to the sign of $s_z$ of the tested spin. The quantum measurement goes in two steps. On a timescale $1/\sqrt{N}$ the off-diagonal elements of the spin's density matrix vanish due to a unitary evolution of the tested spin and the $N$ apparatus spins; on a larger but still short timescale this is made definite by the bath. Then the system is in a `classical' state, having a diagonal density matrix. The registration of that state is a quantum process which can already be understood from classical statistical mechanics. The von Neumann collapse and the Born rule are derived rather than postulated.

cond-mat.stat-mech↗

arXiv2002

Curie-Weiss model of the quantum measurement process

A hamiltonian model is solved, which satisfies all requirements for a realistic ideal quantum measurement. The system S is a spin-$\half$, whose $z$-component is measured through coupling with an apparatus A=M+B, consisting of a magnet $\RM$ formed by a set of $N\gg 1$ spins with quartic infinite-range Ising interactions, and a phonon bath $\RB$ at temperature $T$. Initially A is in a metastable paramagnetic phase. The process involves several time-scales. Without being much affected, A first acts on S, whose state collapses in a very brief time. The mechanism differs from the usual decoherence. Soon after its irreversibility is achieved. Finally the field induced by S on M, which may take two opposite values with probabilities given by Born's rule, drives A into its up or down ferromagnetic phase. The overall final state involves the expected correlations between the result registered in M and the state of S. The measurement is thus accounted for by standard quantum statistical mechanics and its specific features arise from the macroscopic size of the apparatus.

cond-mat.stat-mech↗