We explore how the expectation values 〈ψ|A|ψ〉 of a largely arbitrary observable A are distributed when normalized vectors |ψ〉 are randomly sampled from a high-dimensional Hilbert space. Our analytical results predict that the distribution exhibits a very narrow peak of approximately Gaussian shape, while the tails significantly deviate from a Gaussian behavior. In the important special case that the eigenvalues of A satisfy Wigner's semicircle law, the expectation-value distribution for asymptotically large dimensions is explicitly obtained in terms of a large deviation function, which exhibits two symmetric nonanalyticities akin to critical points in thermodynamics.
We explore how the expectation values 〈ψ|A|ψ〉 of a largely arbitrary observable A are distributed when normalized vectors |ψ〉 are randomly sampled from a high-dimensional Hilbert space. Our analytical results predict that the distribution exhibits a very narrow peak of approximately Gaussian shape, while the tails significantly deviate from a Gaussian behavior. In the important special case that the eigenvalues of A satisfy Wigner's semicircle law, the expectation-value distribution for asymptotically large dimensions is explicitly obtained in terms of a large deviation function, which exhibits two symmetric nonanalyticities akin to critical points in thermodynamics.
Abstract According to the concept of typicality, an ensemble average can be accurately approximated by an expectation value with respect to a single pure state drawn at random from a high-dimensional Hilbert space. This random-vector approximation, or trace estimator, provides a powerful approach to, e.g. thermodynamic quantities for systems with large Hilbert-space sizes, which usually cannot be treated exactly, analytically or numerically. Here, we discuss the finite-size scaling of the accuracy of such trace estimators from two perspectives. First, we study the full probability distribution of random-vector expectation values and, second, the full temperature dependence of the standard deviation. With the help of numerical examples, we find pronounced Gaussian probability distributions and the expected decrease of the standard deviation with system size, at least above certain system-specific temperatures. Below and in particular for temperatures smaller than the excitation gap, simple rules are not available.
We study the eigenstates of quantum systems with large Hilbert spaces, via their distribution of wave-function amplitudes in a real-space basis. For single-particle "quantum billiards," these real-space amplitudes are known to have Gaussian distribution for chaotic systems. In this work, we formulate and address the corresponding question for many-body lattice quantum systems. For integrable many-body systems, we examine the deviation from Gaussianity and provide evidence that the distribution generically tends toward power-law behavior in the limit of large sizes. We relate the deviation from Gaussianity to the entanglement content of many-body eigenstates. For integrable billiards, we find several cases where the distribution has power-law tails.
Random quantum states are presently of interest in the fields of quantum information theory and quantum chaos. Moreover, a detailed study of their properties can shed light on some foundational issues of the quantum statistical mechanics such as the emergence of well-defined thermal properties from the pure quantum mechanical description of large many-body systems. When dealing with an ensemble of pure quantum states, two questions naturally arise, what is the probability density function on the parameters which specify the state of the system in a given ensemble, and does there exist a "most typical" value of a function of interest in the considered ensemble? Here, two different ensembles are considered, the random pure state ensemble (RPSE) and the fixed expectation energy ensemble (FEEE). By means of a suitable parametrization of the wave function in terms of populations and phases, we focus on the probability distribution of the populations in these ensembles. A comparison is made between the distribution induced by the inherent geometry of the Hilbert Space and an approximate distribution derived by means of the minimization of the informational functional. While the latter can be analytically handled, the exact geometrical distribution is sampled by a Metropolis-Hastings algorithm. The analysis is made for an ensemble of wave functions describing an ideal system composed of n spins of 1/2 and reveals the salient differences between the geometrical and the approximate distributions. The analytical approximations are proven to be useful tools in order to obtain an ensemble-averaged quantity. In particular, we focus on the distribution of the Shannon entropy by providing an explanation of the emergence of a typical value of this quantity in the ensembles.
We numerically study the extreme-value statistics of the Schmidt eigenvalues of reduced density matrices obtained from the ergodic eigenstates. We start by exploring the extreme value statistics of the ultrametric random matrices and then the related Quantum Sun Model, which is also a toy model of avalanche theory. It is expected that these ergodic eigenstates are purely random and thus possess random matrix theory-like features, and the corresponding eigenvalue density should follow the universal Marchenko-Pastur law. Nonetheless, we find deviations, specifically near the tail in both cases. Similarly, the distribution of maximum eigenvalue, after appropriate centering and scaling, should follow the Tracy-Widom distribution. However, our results show that, for both the ultrametric random matrix and the Quantum Sun model, it can be better described using the extreme value distribution. As the extreme value distribution is associated with uncorrelated or weakly correlated random variables, the results hence indicate that the Schmidt eigenvalues exhibit much weaker correlations compared to the strong correlations typically observed in Wishart matrices. Similar deviations are observed for the case of minimum Schmidt eigenvalues as well . Despite the spectral statistics, such as nearest neighbor spacing ratios, aligning with the random matrix theory predictions, our findings reveal that randomness is still not fully achieved. This suggests that deviations in extreme-value statistics offer a stringent test to probe the randomness of ergodic eigenstates and can provide deeper insights into the underlying structure and correlations in ergodic systems.
Abstract Typicality is a well-established and very general property of quantum many-body systems, referring to the phenomenon that the expectation values of any given observable are practically indistinguishable for the overwhelming majority of all pure states (normalized vectors) in a sufficiently high-dimensional Hilbert (sub-)space. Here, we provide very simple and general arguments that analogous typicality properties of pure states (phase space points) in classical many-body systems are still expected to hold true for macroscopic observables, but not any more for microscopic (few-body) observables.
Abstract We consider a closed macroscopic quantum system in a pure state ψt ψ t evolving unitarily and take for granted that different macro states correspond to mutually orthogonal subspaces Hν H ν (macro spaces) of Hilbert space, each of which has large dimension. We extend previous work on the question what the evolution of ψt ψ t looks like macroscopically, specifically on how much of ψt ψ t lies in each Hν H ν . Previous bounds concerned the absolute error for typical ψ0 ψ 0 and/or t and are valid for arbitrary Hamiltonians H; now, we provide bounds on the relative error, which means much tighter bounds, with probability close to 1 by modeling H as a random matrix, more precisely as a random band matrix (i.e., where only entries near the main diagonal are significantly nonzero) in a basis aligned with the macro spaces. We exploit particularly that the eigenvectors of H are delocalized in this basis. Our main mathematical results confirm the two phenomena of generalized normal typicality (a type of long-time behavior) and dynamical typicality (a type of similarity within the ensemble of ψ0 ψ 0 from an initial macro space). They are based on an extension we prove of a no-gaps delocalization result for random matrices by Rudelson and Vershynin (Geom Funct Anal 26:1716–1776, 2016).