Classifying deviation from standard quantum behavior using Kullback–Leibler

divergence

Salman Sajad Wani∗ and Saif Al-Kuwari†

Qatar Center for Quantum Computing,
College of Science and Engineering,
Hamad Bin Khalifa University, Qatar
arXiv:2308.02496v1 [quant-ph] 18 Jul 2023

Xiaoping Shi‡ and Yiting Chen

Irving K. Barber School of Arts and Sciences,
University of British Columbia Okanagan, Kelowna, BC V1V 1V7, Canada

Abrar Ahmed Naqash§ and Seemin Rubab¶

Department Of Physics, National Institute Of Technology Srinagar, Jammu and Kashmir 190006, India

Mir Faizal∗∗
Canadian Quantum Research Center, 204-3002,
32 Ave Vernon, BC V1T 2L7 Canada and
Irving K. Barber School of Arts and Sciences,
University of British Columbia Okanagan, Kelowna, BC V1V 1V7, Canada

S. Kannan††
Department of Quantum Science and Technology, Research School of Physics,
The Australian National University, Canberra, ACT 2601, Australia.

In this letter, we propose a novel statistical method to measure which system is better
suited to probe small deviations from the usual quantum behavior. Such deviations are
motivated by a number of theoretical and phenomenological motivations, and various systems
have been proposed to test them. We propose that measuring deviations from quantum
mechanics for a system would be easier if it has a higher Kullback–Leibler divergence. We
show this explicitly for a non-local Scrödinger equation and argue that it will hold for any
modification to standard quantum behavior. Thus, the results of this letter can be used to
classify a wide range of theoretical and phenomenological models.

∗
Electronic address: sawa54992@hbku.edu.qa
†
Electronic address: smalkuwari@hbku.edu.qa
‡
Electronic address: xiaoping.shi@ubc.ca
§
Electronic address: abrarnaqash phy@nitsri.ac.in
 2

I. INTRODUCTION

Several proposals motivated by physically important phenomena predict a deviation from stan-
dard quantum behavior. For example, it is known that the usual Copenhagen interpretation
makes physical phenomena observer-dependent [1], and to obtain an observer-independent objec-
tive formalism of quantum mechanics, objective-collapse theories (such as the Diosi-Penrose for the
gravitational-induced collapse [2] or the collapse by stochasticity [3]) have been proposed. Even
though these alternatives can actually provide an observer-independent formalism for quantum
mechanical decoherence, they also predict small deviations from standard quantum behavior. The
measurement problem has also been addressed using a deformation of the Heisenberg algebra,
which again predicts a deviation from standard quantum behavior [9]. Several experiments have
been proposed to measure such deviations [4–8], but as these deviations are very small, it is hard
to detect them. In addition, there is an absence of any universal criteria to classify various ex-
periments used to detect such small deviations; thus, it becomes hard to know which experiments
should be performed. As it is hard to perform such ultra-precise experiments, it is crucial to have
such criteria, which would provide important information about the effect of such deviations on
the quantum systems. It is expected that modifications would depend not only on the kind of
deviation but also on the system used to measure such a deviation. Thus, in this letter, we, for
the first time, provide such a criterion to measure how the behavior of different systems changes
by small modifications to standard quantum behavior. This can, in turn, act as a guiding tool
for experiments to know which experiments have a better chance of detecting a specific form of
modification of quantum mechanics.
We point out that the modifications of quantum mechanics are not limited to the measurement
problem. The modification of quantum behavior by an intrinsic minimal length is motivated by
quantum gravity [10]. Such modifications deform the Heisenberg algebra. Even though such
deformation of the Heisenberg algebra was initially motivated by Planck scale physics [11], they
have become important in studying various low-energy effective field theories [12]. In fact, it has
been demonstrated that such deformations occur due to the derivative expansion of effective field
theories [13]. They can be used to analyze purely condensed matter effects like the consequences
of next-to-nearest hopping in graphene and can modify the quantum transport in graphene at

¶
Electronic address: rubab@nitsri.net
∗∗
Electronic address: mirfaizalmir@googlemail.com
††
Electronic address: Kannan.S@anu.edu.au
 3

short distances [14]. The criteria proposed in this paper can also find applications for analyzing
and classifying such effects in condensed matter physics. Apart from such modifications, the
modifications of quantum behavior also occur due to noncommutative modifications of spacetime
[15], which occurs in string theory due to background fields, and polymer quantization [16] which
occurs in loop quantum gravity. Various low-energy experiments have been proposed to study
such deviations from quantum behavior [18–21]. However, like the experiments used to address
measurement theory [4–8], no criteria exist to classify them, and hence we do not know which
system is better suited to test such deviations from the usual quantum behavior. Here, again the
results of this letter can be used, and hence the results of this letter have wide applications which
go beyond the modifications motivated by measurement theory.
We will apply the formalism developed to a physically important example to show how this
formalism works. The specific modification will be based on a non-local modification of the original
Schrödinger evolution [17]. However, the important point is that even though we have chosen a
specific modification, similar results can be obtained for any modification of standard quantum
behavior. Even though this modification is motivated by string field theory [22, 23], such non-
local modifications are not limited to string theory and occur in other approaches to quantum
gravity, such as Causal sets [28]. It has also been observed that such a modification could produce
entanglement in a system which was not entangled [29, 30]. Our method can again be used to
classify different experiments used to detect such behavior [14–17]. Non-locality is also important
as it has been proposed that the non-locality is needed to restore unitarity and resolve the black hole
information paradox [31]. We will demonstrate that if we neglect the temporal modifications, then
spatial non-locality produces corrections similar to the corrections obtained from the deformation
of the Heisenberg algebra [12–14].
To obtain a universal criterion to classify any modification of quantum mechanics, we first note
that such modifications usually depend on a parameter β. We can obtain both the original and
modified (by a parameter β) probability densities from their respective wave functions. Then we
can use Kullback–Leibler divergence, which measures how different a probability distribution is
from another [38, 39]. Even though the Kullback–Leibler divergence is not a statistical distance,
it can provide this important information about the separation between two distributions. If
the Kullback–Leibler divergence between the original and modified probability densities is more
for a given system, then it would be easier to detect such modifications than the modifications
where the Kullback–Leibler divergence is small. So, the Kullback–Leibler divergence can be used
to classify deviations from quantum mechanics, and we should experimentally use systems with
 4

larger Kullback–Leibler divergence. As it is easier to study the modification of quantum mechanics

using the momentum space wave function [32, 33], we will use the Kullback–Leibler divergence of
the two divergences for a modified and original wave function in momentum space.

II. NON-LOCAL DEVIATION FROM QUANTUM MECHANICS

Even though our analysis will be valid for any theory, motivated by either quantum measurement
problem [1–3] or quantum gravity [14–17], which suggests some deviation from standard quantum
behavior, we focus on a specific nonlocal model [17] motivated by string field theory [22, 23].
However, our analysis holds for any such deviation and can be universally used to classify deviations
from standard quantum behavior.
To understand how this non-local behavior could be detected in high precision but low-velocity
experiments, we start from a free complex massive scalar field, ϕ, with m as its mass. We expect
the dynamics of this field will be described by a nonlocal Klein-Gordon equation [17] due to string
field theory [22, 23].

(2 + µ2 ) exp [lk2 (2 + µ2 )] = 0. (1)

where µ = mc/ℏ is the inverse of the reduced Compton wavelength of the field. This specific
deformation occurs due to string field theory. However, even though string field theory produces
non-locality in field theories, the non-locality can also occur from other motivations [24–27], and
need not be limited to string field theory. In fact, in some models, such as those obtained from
Causal sets [28], the non-locality is not even a polynomial function. Thus, we start from a general
functions f of the Klein-Gordon operator 2 = c−2 ∂t2 − ∇2 , and write a general non-local theory as

f (2 + µ2 )ϕ = 0 (2)

As we want to recover the standard large distance limit of the theory, the function f has to satisfy

lim f (2 + µ2 ) → 2 + µ2 (3)
lk →0

where lk is the scale at which non-local effects become important. We assume that f is an analytic
function, and expand it formally as a power series

∞
X
f (z) = bn z n (4)
n=1
 5

mc2
Writting ϕ(x) = e−i ℏ
t
ψ(t, x) and taking the non-relativistic limit (c → ∞), we obtain a non-local
Scr´’odinger equation [17]

f (S)ψ(t, x) = 0, (5)

where S is defined as

∂ ℏ2 2
S = iℏ + ∇ , (6)
∂t 2m

The usual quantum mechanical local Schrödinger equation can be obtained by expanding in power
series, and neglecting all corrections proportional to lk . However, if we consider these corrections,
we can write a free non-local Schrödinger equation as
∞
−2m n−1 2n−2 n
X  
f (S)ψ = Sψ + bn lk S ψ=0 (7)
ℏ2
n=2

with bn as dimensionless coefficients. As the non-local deformation only occurs in the Kinetic part
of the Schrödinger equation, we can use the non-local free Schrödinger equation to motivate a non-
local Schrödinger equation with a potential term. Thus, we can write the non-local Schrödinger
equation in a potential V as

f (S)ψ = V ψ (8)

However, if we now only retain the first order corrections, and define a new constant β =
b2 (−2m/ℏ2 )lk2 , then we observe that we obtain the equation

β 2
Sψ + S ψ =Vψ (9)
m

As we usually deal with stationary states for most tabletop experiments [18–21], we can neglect
the temporal dependence and only keep the non-locality in spatial coordinates only. For such
theories, with special non-locality, we can write the modified Schrödinger equation with a β∇4
correction. It is interesting to observe this is exactly the correction obtained from a deformation
of the Heisenberg algebra [9–12]. Thus, the modification of [xi , pj ] = iℏ to [xi , pj ] = iℏfji (p, β),
where fji (p, β) is a suitable tensorial function of the momentum and the parameter β [9]. This de-
formation was initially motivated by generalizing the usual uncertainty principle to a generalized
uncertainty principle, to incorporate the existence of a minimal length that occurs in quantum
gravity [10]. Here, we have observed that this can also occur from non-local field theories. Even
though various experiments have been proposed to analyze the deviation from the usual quantum
behavior produced from such theories [18–21], here we will for the first time use Kullback–Leibler
 6

divergence [38, 39] to analyze how a specific system can be better suited to detect such deviations.
The Kullback–Leibler divergence measures how different a probability distribution is from another
probability distribution and the more the Kullback–Leibler divergence between the original and
modified probability densities ψ 2 , the easier it would be to experimentally test them. The Kull-
back–Leibler divergence between an original probability density |ψ|2 and a modified probability
density |ψ̃|2 is given by [38, 39].
!
∞
|ψ̃(x)|2
Z
2 2 2
DKL (|ψ̃(x)| ||ψ(x)| ) = |ψ̃(x)| log dx (10)
−∞ |ψ(x)|2

We will perform our analysis in the momentum space, and to define the square of wave function in
the momentum space, usually the integration measure is defined as dp/1 − βp2 [32, 33]. However,
we will absorb the factor 1/1−βp2 in the inner product of the deformed momentum wave functions,
and use it to define the Kullback–Leibler divergence measures in momentum space representation.
Thus, we will write the Kullback–Leibler divergence in the momentum space as
Z ∞ !
2 2 2 |ψ̃(p)|2
DKL (|ψ̃(p)| ||ψ(p)| ) = |ψ̃(p)| log dp (11)
−∞ |ψ(p)|2

We claim that this Kullback–Leibler divergence depends on the system used to test a specific
modification of quantum mechanics. We test our claim by using two simple systems of quantum
mechanics modified by a deformation of the Heisenberg algebra, i.e., we will analyze a simple har-
monic oscillator and a particle in a box. Even for such simple systems, with a simple modification
of quantum mechanics (by spatial non-locality produced by a deformation of the Heisenberg alge-
bra) we observe that the Kullback–Leibler divergence is different. Hence, we conclude that one of
them is better suited to detect such modifications of quantum mechanics than others.

III. HARMONIC OSCILLATOR

In the previous section, we developed a formalism to classify the deviations of quantum me-
chanics. In this section, we will apply this formalism to a concrete example. Now we analyze a
quantum harmonic oscillator and obtain its Kullback–Leibler divergence. The quantum gravity
corrections to the wave function for harmonic oscillator in momentum space are given by [35] .
s √
λ n!(n + λ) β 1+λ λ
ψ̃n (p) = 2 Γ(λ) c Cn (s) (12)
2πΓ(n + 2λ)
p √
where λ = (1/2) + (1/4) + (1/k 4 ), k = βr−1 , c = (1 + βp2 )−1/2 , s = βp(1 + βp2 )−1/2 ,
p

1
r= mωℏ and Cn (s) is the Gegnbauer polynomial. Furthermore, we assume that the ground state
 7

probability density is concentrated at the origin, which means the particle spends most of its time
at the bottom of the harmonic potential well, as one would expect for a state with little energy.
Thus for ground state, n = 0, we can write (ψ̃0 (p) = ψ˜p )
s √
λ β 1
ψ˜p = 2λ Γ(λ) (13)
2πΓ(2λ) (1 + βp2 ) 1+λ
2

To find the probability density, we need to take the square modulus of the given wave function
 √ 2λ−1 2 
λ β2 Γ (λ) 1
|ψ˜p |2 = (14)
πΓ(2λ) (1 + βp2 )1+λ

The probability density for the usual Harmonic Oscillator, can be written as [37]
r
2 r
|ψp | = exp (−rp2 ) (15)
π

We note that one can arrive at the unmodified wave function Eq.(15) from the corrected wave
1 2
function Eq.(14) by using the standard definition of limits lim 2 1+λ
= e−rp . Now we can
β→0 (1 + βp )
write the Kullback–Leibler divergence between the original and modified probability densities as
Z ∞
2 2 1 h
2 1+λ 2
i
DKL (ψ̃p |||ψp ) = −B 2 1+λ
log A(1 + βp ) exp (−rp ) dp (16)
−∞ (1 + βp )

where A and B are given by

√ √
21−2λ rπΓ(2λ) λ β22λ−1 Γ2 (λ)
A= √ , B= (17)
βλΓ2 (λ) πΓ(2λ)
√
Using the Gamma function–Legendre formula 22z−1 Γ(z)Γ(z + 1/2) = πΓ(2z), we can further
simplify A and B as

r Γ(λ + 12 )
r r
β Γ(λ)
A= , B=λ (18)
β λΓ(λ) π Γ(λ + 12 )

Finally, we make use of Sterling asymptotic approximation for large z, Γ(z + 1/2) = Γ(z)z 1/2 , and
further simplify A and B. Here, the use of Sterling approximation is justified because for small β,
λ takes large values
r r
r βλ
A= , B= (19)
βλ π

Now binomial expanding it in terms of β, λ ≈ 1/2 + r/β + β/8r. Note that λ strongly depends
upon the second term only as it diverges quickly for small values β; therefore for practical purposes
we have for small β, we have λ = r/β. Using these approximations, we can write A and B as
r
r
A ≈ 1, B ≈ (20)
π
 8

Now Eq.(16) can be written as

r Z ∞
r −r
h r
i
DKL (ψ̃p2 |||ψp2 ) =− (1 + βp2 ) β log (1 + βp2 ) β exp (−rp2 ) dp (21)
π −∞
Finally, we can make use of these two Taylor’s expansions, which are valid for small values of β
−r 2 1  2

(1 + βp2 ) β = e−p r + β p4 re−p r + O β 2


(22)
2
r 2 1  2

(1 + βp2 ) β = ep r − β p4 rep r + O β 2


(23)
2
We can thus obtain the Kullback–Leibler divergence for a quantum harmonic oscillator for the
probability density obtained from HUP (Heisenberg’s Uncertainty Principle) and GUP (Generalised
Uncertainty Principle) up to the strength of O(β 1 ) as
r Z ∞
r 2
2 2
DKL (ψ̃p |||ψp ) ≈ β p4 re−p r dp (24)
π 0
This integral is the standard Gaussian integral multiplied by the polynomial in the p space, so we
obtain an analytical solution for DKL of the harmonic oscillator as
3
DKL (ψ̃p2 |||ψp2 ) ≈ β (25)
8r
We plot this analytical solution along with the Kullback–Leibler divergence for a particle in a box.

IV. PARTICLE IN A BOX

In this section, we will apply the formalism developed to a different physical system. Thus, we
will be able to show that the difficulty to detect a specific modification of quantum mechanics will
depend on the system chosen to detect it. This is an important observation, as it can act as a
guiding principle for tabletop experiments, as it can directly give us important about the ability of
such experiments to detect a specific deviation. We will analyze the Kullback–Leibler divergence
of a particle in a box, which can be represented by an infinite well potential. The wave function for
√
the nth excited state of this system can be written as ψ(x) = 2a−1 sin nπxa−1 . It may be noted
that up to the first order in β, in the framework of spatial non-local deformation, there is no change
in position space eigenstates of the quantum particle in box. However, as we have a shift of energy
levels [34], one needs to obtain wave functions in the momentum space. The momentum-space
wave function can be obtained by the Fourier transform of the already available position-space
wave function. For the particle trapped in a one-dimensional box, the Fourier transform is given
by the following expression:
Z a
r
1 2 nπx
ψ(p) = √ exp (−ιxp) sin dx (26)
2π 0 a a
 9

For ground state, with n = 1 and setting a = 1, we obtain

√ 1 + exp (−ιp)
 
ψ(p) = π (27)
π 2 − p2

Now we obtain the corrected wave function in the momentum space representation. Due to the
modification of the commutation algebra, we need to perform the modified Fourier Transform of
our deformed wave function, which can be obtained as [32, 33]
Z ar
1 − 3βp2
exp −ipx 1 − 3βp2 sin(πx) dx



ψ̃(p) = (28)
0 π

By solving the above equation, we arrive at the deformed wave function for the infinite potential
well in the momentum space representation as
 2

π − 3πβp2 1 + e−ip(1−3βp )
p
(29)
π 2 − p2 (1 − 3βp2 )2

To find the probability density, we take the square modulus of this wave function

2π 1 − 3βp2 1 + cos p 1 − 3βp2

2
|ψ̃(p)| =  2 (30)
π 2 − p2 (1 − 3βp2 )2

To find the Kullback–Leibler divergence, we need both the modified and the unmodified proba-
bility density distributions. The unmodified probability density function for infinite potential well
is given by [34]

1 + cos p
|ψ(p)|2 = 2π (31)
(π 2 − p2 )2

simplifying it further, we obtain the Kullback–Leibler divergence as

2 2
  !
π 2 −p2 (1−3βp2 ) (cos(p)+1)
3βp2 − 1 cos p − 3βp3 + 1 log




2π (p2 −π 2 )2 (1−3βp2 )(cos(p−3βp3 )+1)
Z ∞
DKL (ψ̃p2 |||ψp2 ) = dp  2
−∞ π 2 − p2 (1 − 3βp2 )2
(32)
In fact, we can further simplify by expanding the log term in the integral up to the first order of
the β
∞ 6πβ p5 (− sin(p)) − 3p4 − 3p4 cos(p) + π 2 p3 sin(p) − π 2 k 2 − π 2 p2 cos(p)
Z

DKL (ψ̃p2 |||ψp2 )p ≈ dp
−∞ (π 2 − p2 )3
(33)
As it is arduous to solve this expression analytically, we use numerical methods to plot DKL for
infinite well potential.
 10

(a) log DKL vs β for large values of β. (b) log DKL vs β for small values of β.

FIG. 1: log DKL vs β for the quantum harmonic oscillator and particle in the box

To test a modification of quantum mechanics, for a system to be able to detect small changes
in the parameter β, the system should not be robust with respect to β. This can be obtained
if the system has higher Kullback–Leibler divergence [38, 39]. Therefore, we have analyzed the
Kullback–Leibler divergence for different values of β. We have taken the length of the well equal
to a = 1 and r = 1 for the harmonic oscillator. As the value of Kullback–Leibler divergence is
small, we plot the value of log DKL vs β. Figures 1a and 1b show our results for large and small
values of β, respectively. In these figures, the orange curve describes the value of log DKL for
the particle in the box while the blue curve describes the same for a harmonic oscillator. As the
value of the log DKL is larger for a harmonic oscillator than the particle in a box, it would be
easier to experimentally test such a deviation from standard quantum behavior using a harmonic
oscillator than a particle in a box. This is because the original and deformed probability densities
are increasingly different from each other for a harmonic oscillator than a particle in a box. This
behavior also holds for both small and large values of β. Thus, for any experiment used to test
non-locality, the deviation from standard quantum behavior will be larger for a harmonic oscillator
than for a particle in a box. Hence, any experimental setup which uses the non-local modification
of a harmonic oscillator will be better suited to detect non-locality than any experimental setup
which uses a particle in a box. Now this behavior might be different for different modifications,
and different systems, but what we have demonstrated is that the DKL (and hence the difficulty to
experimentally detect) can be different for various physical systems, even for the same modification.
 11

V. CONCLUSION

The results obtained in this letter can have various important applications. They can classify
systems that are better suited to detect a specific modification of quantum mechanics, and hence
be of use to experiments in deciding which experiments would be better suited to detect a specific
modification of quantum mechanics. Even though we found that for a given modification a specific
system is clearly better than another to detect the deviation from ordinary quantum behavior.
This may not be the case with other modifications of quantum mechanics. Hence, a thorough
study of various systems and different modifications to quantum mechanics has been made to
understand how different experiments can be used to probe different modifications to quantum
mechanics. Here, the important result is that Kullback–Leibler divergence can be used to classify
various systems that are being used to detect a specific deviation from standard quantum behavior.
The Kullback–Leibler divergence has been used to test the robustness of the system under the
change of parameter, and this is done in the sensitivity analysis of that parameter [40, 41]. It would
be interesting to analyze the robustness of β, which can be used to fix a bound on how accurately
can a given system measure β. Thus, not only can the Kullback–Leibler divergence be used to
classify experiments that can test a particular modification, it can provide important information
on the accuracy of such tests.
 12

[1] T. J. Hollowood, Contemp. Phys. 57, no.3, 289-308 (2016)

[2] R. Penrose, Gen. Relativ. Gravit. 28, 581 (1996); L. Diosi, Phys. Lett. A 105, 199 (1984)
[3] L. Mertens, M. Wesseling and J. van Wezel, SciPost Phys. 14, 114 (2023)
[4] M. Carlesso, A. Vinante, and A. Bassi, Phys. Rev. A 98, 022122 (2018)
[5] B. Schrinski, Y. Yang, U. von Lüpke, M. Bild, Y. Chu, K. Hornberger, S. Nimmrichter, and M. Fadel,
Phys. Rev. Lett. 130, 133604 (2023)
[6] J. Christian, Phys. Rev. Lett. 95, 160403 (2005)
[7] M. Bahrami, A. Smirne, and A. Bassi, Phys. Rev. A 90, 062105 (2014)
[8] M. Carlesso, A. Bassi, P. Falferi, and A. Vinante, Phys. Rev. D 94, 124036 (2016)
[9] L. Petruzziello and F. Illuminati, Nature Commun. 12, no.1, 4449 (2021)
[10] I. Pikovski, M. R. Vanner, M. Aspelmeyer, M. Kim and C. Brukner, Nature Phys. 8, 393 (2012)
[11] A. Kempf, G. Mangano and R. B. Mann, Phys. Rev. D 52, 1108-1118 (1995)
[12] M. Faizal, A. F. Ali and A. Nassar, Int. J. Mod. Phys. A 30, no.30, 1550183 (2015)
[13] M. Faizal, A. F. Ali and A. Nassar, Phys. Lett. B 765, 238-243 (2017)
[14] N. A. Shah, A. Contreras-Astorga, F. Fillion-Gourdeau, M. A. H. Ahsan, S. MacLean and M. Faizal,
Phys. Rev. B 105, no.16, L161401 (2022)
[15] N. Seiberg and E. Witten, JHEP 09, 032 (1999)
[16] A. Laddha and M. Varadarajan, Class. Quant. Grav. 27, 175010 (2010)
[17] A. Belenchia, D. M. T. Benincasa, S. Liberati, F. Marin, F. Marino and A. Ortolan, Phys. Rev. Lett.
116, no.16, 161303 (2016)
[18] S. Das and E. C. Vagenas, Phys. Rev. Lett. 101, 221301 (2008)
[19] P. Pedram, K. Nozari and S. H. Taheri, JHEP 03, 093 (2011)
[20] F. Marin, F. Marino, M. Bonaldi, M. Cerdonio, L. Conti, P. Falferi, R. Mezzena, A. Ortolan, G. A. Prodi
and L. Taffarello, et al. Nature Phys. 9, 71-73 (2013)
[21] C. Quesne and V. M. Tkachuk, Phys. Rev. A 81, 012106 (2010)
[22] H. Erbin, A. H. Fırat and B. Zwiebach, JHEP 01, 167 (2022)
[23] A. S. Koshelev, K. Sravan Kumar and P. Vargas Moniz, Phys. Rev. D 96, no.10, 103503 (2017)
[24] R. Trinchero, Phys. Rev. D 98, no.5, 056023 (2018)
[25] A. S. Koshelev, K. Sravan Kumar and P. Vargas Moniz, Phys. Rev. D 96, no.10, 103503 (2017)
[26] V. K. Mishra, G. Fai, P. C. Tandy and M. R. Frank, Phys. Rev. C 46, 1143-1146 (1992)
[27] R. E. Wagner, M. R. Ware, E. V. Stefanovich, Q. Su and R. Grobe, Phys. Rev. A 85, 022121 (2012)
[28] A. Belenchia, D. M. T. Benincasa and S. Liberati, JHEP 03, 036 (2015)
[29] A. Belenchia, D. M. T. Benincasa, S. Liberati, F. Marin, F. Marino and A. Ortolan, Phys. Rev. D 95,
no.2, 026012 (2017)
[30] A. Belenchia, D. M. T. Benincasa, F. Marin, F. Marino, A. Ortolan, M. Paternostro and S. Liberati,
 13

Class. Quant. Grav. 36, 155006 (2019)

[31] S. B. Giddings, Phys. Rev. D 74, 106005 (2006)
[32] A. N. Tawfik and A. M. Diab, Rept. Prog. Phys. 78, 126001 (2015)
[33] A. N. Tawfik and A. M. Diab, Int. J. Mod. Phys. D 23, 1430025 (2014)
[34] K. Nozari and T. Azizi, Gen. Rel. Grav. 38, 735-742 (2006)
[35] L. N. Chang, D. Minic, N. Okamura and T. Takeuchi, Phys. Rev. D 65, 125027 (2002)
[36] S. Das, M. P. G. Robbins and M. A. Walton, Can. J. Phys. 94, no.1, 139-146 (2016)
[37] S. S. Coelho, L. Queiroz and D. T. Alves, Entropy 24, no.12, 1851 (2022)
[38] I. Csiszar, Ann. Probab. 3, 146-158 (1975).
[39] S. Kullback, Ame. Statistician 41, 340–341 (1987).
[40] R. Teixeira, A. O’Connor and Maria Nogal, Structural Safety 81, 101860 (2019)
[41] M. Zhu, S, Liu and J. Jiang, Int. J. Appro. Reaso. 90, 37 (2017)