Thermalization in a model of enhanced memory capacity
Oleg Kaikov · 2025
Read Thermalization in a model of enhanced memory capacity by Oleg Kaikov (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-01556.txt.
Introduction
Kaikov studies heat spread in a lone quantum set built to hold much data. The set was made of late as a toy for black hole data work. Black holes both hold and work vast sums of facts in that view. The study runs size scans on the lone set with no bath tied. Signs point to heat-like ends for the data part at large size. A new heat path is put forth from those signs.
A broad class of start states drives the path in his claim. For those states, a small-part watch still finds heat-like means. That holds though state sums and hold shares shift in linked ways. Those linked shifts are too big to brush off in the read. The claim sets this set apart from stock heat tales.
Catalog record
| CREATOR | Oleg Kaikov |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-01556
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Thermalization in a model of enhanced memory capacity Thanks: The present work is an improved version of chapter 4 “Thermalization Despite Correlation” of the author’s Ph.D. thesis [ 1 ] , and has neither been previously peer-reviewed nor published elsewhere. Oleg Kaikov Abstract We study thermalization within a quantum system with an enhanced capacity to store information. This system has been recently introduced to provide a prototype model of how a black hole processes and stores information. We perform a numerical finite-size analysis of this isolated quantum system and find indications that its information-carrying subsystem approaches thermality in the large system-size limit. The results lead us to suggest a novel thermalization mechanism. The corresponding distinguishing characteristic is that for a large class of physically meaningful non-equilibrium initial states | in ⟩ \ket{\text{in}} , a few-body observable A ^ \hat{A} thermalizes despite unignorable correlations between the fluctuations of its eigenstate expectation values ⟨ α | A ^ | α ⟩ \bra{\alpha}\hat{A}\ket{\alpha} in the eigenstate basis of the model { | α ⟩ } \left\{\ket{\alpha}\right\} and the fluctuations of the squared magnitudes of the coefficients | C α | 2 = | ⟨ α | in ⟩ | 2 |C_{\alpha}|^{2}=|\braket{\alpha|\text{in}}|^{2} . \shortdate 1 Introduction 1.1 Overview Understanding how isolated quantum many-body systems thermalize is a longstanding area of research [ 2 ] . Specifically, consider such a system prepared in a pure initial state far from equilibrium. The question of interest is, what is the microscopic mechanism by which few-body observables within this system equilibrate under unitary time-evolution towards their typical expectation values given by an appropriate statistical ensemble? A distinguished mechanism for thermalization was established by the Eigenstate Thermalization Hypothesis (ETH) [ 3 , 4 ] . This work has been further developed in [ 5 , 6 , 7 ] among other studies (for reviews see e.g. [ 8 , 9 , 10 , 11 , 12 ] ). Moreover, other mechanisms have been proposed, such as [ 13 , 14 , 15 , 16 , 17 , 18 ] , among others. The primary goal of this work is to test thermalization within a specific prototype system that exhibits an enhanced capacity to store information [ 19 , 20 ] . Furthermore, we would like to establish the microscopic mechanism that explains how this thermalization occurs. To investigate this, we consider finite-size realizations of the system. We find indications that certain few-body observables thermalize in the thermodynamic limit (TDL), i.e. in the limit of large system size. Moreover, our results suggest that this thermalization occurs via a new mechanism. Specifically, a few-body observable \hat{A} can still thermalize despite non-negligible correlations between two quantities: the fluctuations of its eigenstate expectation values A_{\alpha\alpha}=\bra{\alpha}\hat{A}\ket{\alpha} in the eigenstate basis of the model \{\ket{\alpha}\} , and the fluctuations of the squared magnitudes of the coefficients |C_{\alpha}|^{2}=|\braket{\alpha|\text{in}}|^{2} for a non-equilibrium initial state \ket{\text{in}} . The two sets of fluctuations are correlated in such a way that the infinite-time average of the few-body observable is equal to its statistical ensemble average. The model [ 19 , 20 ] that we study was introduced to emulate the information-processing properties of a black hole. Other works, such as [ 21 , 22 , 23 ] , have previously strengthened the connection of this and related systems to the original black hole fast scrambling conjecture [ 24 , 25 ] . However, our principal goal is more direct: We want to study thermalization in the specific realization [ 20 ] of a quantum system that exhibits an enhanced capacity to store information. Indeed, this quantum model is interesting in its own right and deserves an individual analysis. In the remainder of the introduction we review the two aspects that are central for this work: thermalization and the enhanced memory capacity effect. In section 1.2 , we briefly review thermalization in isolated quantum systems, the ETH, two other mechanisms that are relevant for the mechanism we introduce, as well as outline the new mechanism itself. In section 1.3 we recapitulate the phenomenon of enhanced memory capacity as well as the model in [ 19 , 20 ] . In section 2 we test the thermalization of certain few-body observables within the system by considering finite-size realizations of the model and extrapolating to the large-size limit. In section 3 we test whether these few-body observables thermalize via the ETH. In sections 4 and 5 we test other mechanisms in application to the thermalization of these few-body observables within the system. In section 6 we introduce the novel thermalization mechanism, discuss our findings and comment on possible applications. In section 7 we present our conclusions. 1.2 Thermalization 1.2.1 Conditions for thermalization Here we recapitulate thermalization of few-body observables within isolated quantum systems. Consider an isolated quantum many-body system of D degrees of freedom. Let \hat{H} denote the corresponding Hamiltonian. We denote its \mathcal{N} energy eigenstates by \ket{\alpha} and the corresponding energies by E_{\alpha} . Consider also a physically meaningful local few-body observable \hat{A} . Let \hat{A}(t)=\mathrm{e}^{i\hat{H}t}\hat{A}\mathrm{e}^{-i\hat{H}t} be the corresponding Hermitian operator in the Heisenberg representation. Here and throughout the work we set \hbar\equiv 1 . We prepare the system in an initial state \ket{\text{in}} , such that \bra{\text{in}}\hat{A}\ket{\text{in}} is sufficiently far from the expectation value of \hat{A} given by the microcanonical ensemble at the energy of the system \bra{\text{in}}\hat{H}\ket{\text{in}} . We say that the observable \hat{A}(t) thermalizes if [ 10 ] : (i) The infinite-time average of the expectation value of the observable \overline{\braket{\hat{A}(t)}} is equal to its microcanonical average, and (ii) The temporal fluctuations of \braket{\hat{A}(t)} about the microcanonical average are small at most later times. 1.2.2 Conditions of the ETH We now review the ETH and state the corresponding conditions on the matrix elements of the observable in the eigenstate basis of the Hamiltonian. Specifically, the ETH states that an observable \hat{A}(t) will thermalize if [ 3 , 4 , 6 , 7 , 26 , 27 ] : (1) The diagonal matrix elements A_{\alpha\alpha}=\bra{\alpha}\hat{A}\ket{\alpha} vary approximately smoothly with E_{\alpha} and the magnitude of the difference between neighboring values |A_{\alpha+1,\alpha+1}-A_{\alpha\alpha}| is exponentially small in D , and (2) The magnitudes of the off-diagonal matrix elements |A_{\alpha\beta}|=|\bra{\alpha}\hat{A}\ket{\beta}| with \alpha\neq\beta are themselves exponentially small in D . These are commonly expressed as [ 6 , 7 , 28 , 29 , 30 ] A_{\alpha\beta}=\mathcal{A}(E)\delta_{\alpha\beta}+\mathrm{e}^{-S(E)/2}f_{A}(E,\omega)R_{\alpha\beta}\,. (1) The average of the relevant energies is E=(E_{\alpha}+E_{\beta})/2 and their difference is \omega=E_{\alpha}-E_{\beta} . The microcanonical expectation value at E is denoted by \mathcal{A}(E) . The thermodynamic entropy S(E) at energy E is given by the logarithm of the number of degenerate microstates. The function f_{A}(E,\omega) is a real function corresponding to a particular observable \hat{A} , with f_{A}(E,\omega)=f_{A}(E,-\omega) . The functions \mathcal{A}(E) and f_{A}(E,\omega) vary smoothly with their arguments. The entries of the random Hermitian matrix R_{\alpha\beta} have zero mean and unit variance. As usual, the Hamiltonian is assumed to be non-degenerate [ 6 , 31 ] . Last, there is a condition that is commonly assumed to hold implicitly in the literature. Namely that the initial state of the system has a relatively small energy uncertainty [ 4 , 5 , 6 , 7 , 14 , 26 ] . We refer to