You are currently viewing a new version of our website. To view the old version click .
Entropy
  • Correction
  • Open Access

17 October 2018

Correction: Naudts, J. Quantum Statistical Manifolds. Entropy 2018, 20, 472

Departement Fysica, Universiteit Antwerpen, Universiteitsplein 1, 2610 Wilrijk Antwerpen, Belgium
This article belongs to the Special Issue Entropy: From Physics to Information Sciences and Geometry

Abstract

Section 4 of “Naudts J. Quantum Statistical Manifolds. Entropy 2018, 20, 472” contains errors. They have limited consequences for the remainder of the paper. A new version of this Section is found here. Some smaller shortcomings of the paper are taken care of as well. In particular, the proof of Theorem 3 was not complete, and is therefore amended. Also, a few missing references are added.
Theorem 1.
 
Theorem 2.
 

1. Corrections in Section 3

The display on top of page 5 should read
| | f ρ , K | | = sup A A f ρ , K ( A ) : | | A | | 1 = sup A A ( π ( A ) K Ω ρ , Ω ρ ) : | | A | | 1 = | | | K | 1 / 2 Ω ρ | | 2 | | | K | 1 / 2 | | 2 = | | K | | .
The operator K is replaced by | K | because K need not be positive.
The sentence ”This is a prerequisite for proving in the next Theorem that this map is the Fréchet derivative of the chart ξ ρ .” should read ”This is a prerequisite for proving in the next Theorem that this map is the Fréchet derivative of the inverse of the chart ξ ρ .”
The proof of the following Theorem is amended.
Theorem 3.
The inverse of the map ξ ρ : M B ρ , defined in Theorem 2, is Fréchet-differentiable at ω = ω ρ . The Fréchet derivative is denoted F ρ . It maps K to f ρ , K , where the latter is defined by (10).
Proof. 
Let K = ξ ρ ( ω σ ) . One calculates
| | ω σ ω ρ F ρ K | | = sup A A | ω σ ( A ) ω ρ ( A ) F ρ K ( A ) | : | | A | | 1 = sup A A | ( π ( A ) Ω ρ , [ e K α ( K ) I K ] Ω ρ ) | : | | A | | 1 | | e K α ( K ) I K | | | α ( K ) | + o ( | | K α ( K ) | | ) .
Note that
| α ( K ) | log | | e K | | | | K | |
and
| | K α ( K ) | | 2 | | K | | .
In addition, if | | K | | < 1 then one has
α ρ ( K ) log ( 1 + | | K Ω ρ | | 2 ) | | K Ω ρ | | 2 .
This holds because λ 1 implies exp ( λ ) 1 + λ + λ 2 . One concludes that (11) converges to 0 faster than linearly as | | K | | tends to 0. This proves that F ρ K is the Fréchet derivative of ξ ρ ( ω σ ) ω σ at σ = ρ . ☐

2. New Version of Section 4

Propositions 1 and 2 of [1] are not correct. This only has consequences for one sentence in the Introduction of [1] and for the results reported in Section 4 of [1]. The text in the Introduction “Next, an atlas is introduced which contains a multitude of charts, one for each element of the manifold. Theorem 4 proves that the manifold is a Banach manifold and that the cross-over maps are linear operators.” should be changed to “Next, an atlas is introduced which contains a multitude of charts, one for each element of the manifold. Theorem 4 proves that the manifold is a Banach manifold and that the cross-over maps are continuous.”
A new version of Section 4 follows below:

4. The Atlas

Following the approach of Pistone and collaborators [1,3,4,24], we build an atlas of charts ξ ρ , one for each strictly positive density matrix ρ . The compatibility of the different charts requires the study of the cross-over map ξ ρ 1 ( σ ) ξ ρ 2 ( σ ) , where ρ 1 , ρ 2 , σ are arbitrary strictly positive density matrices.
Simplify notations by writing ξ 1 and ξ 2 instead of ξ ρ 1 , respectively ξ ρ 2 . Similarly, write Ω 1 and Ω 2 instead of Ω ρ 1 , respectively Ω ρ 2 , and F 1 , F 2 instead of F ρ 1 , respectively F ρ 2 .
Proposition 1.
RETRACTED
Continuity of the cross-over map follows from the continuity of the exponential and logarithmic functions and from the following result.
Proposition 2.
Fix strictly positive density matrices ρ 1 and ρ 2 . There exists a linear operator Y such that for any strictly positive density matrix σ and corresponding positive operators X 1 , X 2 in the commutant A one has X 2 = Y X 1 Y * .
Proof. 
Using the notations of the Appendix of [1], one has
X i = J i ( ρ i 1 / 2 σ ρ i 1 / 2 I ) J i * , i = 1 , 2 .
Note that the isometry J depends on the reference state with density matrix ρ . Therefore, it carries an index i. The above expression for X i implies that
X 2 = Y X 1 Y * with Y = J 2 ( ρ 2 1 / 2 ρ 1 1 / 2 I ) J 1 * .
 ☐
Theorem 4.
The set M of faithful states on the algebra A of square matrices, together with the atlas of charts ξ ρ , where ξ ρ is defined by Theorem 1, is a Banach manifold. For any pair of strictly positive density matrices ρ 1 and ρ 2 , the cross-over map ξ 2 ξ 1 1 is continuous.
Proof. 
The continuity of the map X 1 X 2 follows from the previous Proposition. The continuity of the maps K 1 X 1 and X 2 K 2 follows from the continuity of the exponential and logarithmic functions and the continuity of the function α . ☐

3. Corrections in Section 9

In the proof of Proposition 4, the symbol Ω ρ is missing five times in obvious places. It has been added.

4. Added References

In the overview of papers devoted to the study of the quantum statistical manifold in the finite-dimensional case, the references [2,3] should be added. A quantum version of the work of Pistone and Sempi [4], alternative to [5], is found in [6]. Reference [7] to the work of Ciaglia et al. has been updated.

References

  1. Naudts, J. Quantum Statistical Manifolds. Entropy 2018, 20, 472. [Google Scholar] [CrossRef]
  2. Petz, D.; Sudar, C. Geometries of quantum states. J. Math. Phys. 1996, 37, 2662–2673. [Google Scholar] [CrossRef]
  3. Jenčová, A. Geometry of quantum states: Dual connections and divergence functions. Rep. Math. Phys. 2001, 47, 121–138. [Google Scholar] [CrossRef]
  4. Pistone, G.; Sempi, C. An infinite-dimensional structure on the space of all the probability measures equivalent to a given one. Ann. Stat. 1995, 23, 1543–1561. [Google Scholar] [CrossRef]
  5. Streater, R.F. Quantum Orlicz spaces in information geometry. Open Syst. Inf. Dyn. 2004, 11, 359–375. [Google Scholar] [CrossRef]
  6. Jenčová, A. A construction of a nonparametric quantum information manifold. J. Funct. Anal. 2006, 239, 1–20. [Google Scholar] [CrossRef]

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.