REVIEW 3 major objections 6 minor 32 references
Quantum information loss
T0 review · 3 major / 6 minor · reviewed 2026-08-08 · deepseek-v4-flash
Pith's one-line read This paper proves that zero intrinsic information loss on a code space is exactly quantum error correctability, and applies the criterion to black-hole evaporation.
desk verdict A clean new measure of quantum information loss with a solid exact equivalence to quantum error correction, but the Hayden-Preskill application leans on an unproved continuity step. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the intrinsic information loss K(ρ,E), defined as the infimum, over pure-state decompositions ρ=Σ_i p_i |ψ_i⟩⟨ψ_i| and over POVMs {N_j} on the output, of the conditional entropy H(P)−H(Q) for the joint distribution P(i,j)=p_i Tr[E(|ψ_i⟩⟨ψ_i|)N_j]. The load-bearing equivalence is that K(ρ,E)=0 for all ρ on H_code iff the code satisfies the Knill-Laflamme conditions P E_α^† E_β P = c_αβ P; the proof uses the fact that positive semidefinite operators have orthogonal supports exactly when their trace product vanishes, together with contractivity of trace distance to force the spectral decomposition of a rank-two code state. In the Hayden-Preskill application, the machinery is the second Haar moment of the scrambling unitary, evaluated by the swap trick, which yields the bound on the Haar-averaged root fidelity between outputs for orthogonal inputs.
What would settle it
Compute, for a family of channels that nearly violate the Knill-Laflamme conditions on a small code space, the exact value of sup_ρ K(ρ,E) and the best achievable recovery fidelity; if arbitrarily small sup_ρ K(ρ,E) can coexist with recovery fidelity bounded away from 1 by a fixed constant, the asserted continuity from small information loss to near-perfect recovery fails.
Extended reading notes
Core claim
The paper's core claim is Theorem 3: for a code space H_code with projector P and a channel E with Kraus operators {E_α}, four statements are equivalent: E is universally pristine on H_code; K(ρ,E)=0 for every state ρ supported on H_code; there exists a positive-semidefinite matrix C=(c_αβ) such that P E_α^† E_β P = c_αβ P for all α,β; and there exists a recovery channel R with (R∘E)(ρ)=ρ on the code. The equivalence of the third and fourth statements is the Knill-Laflamme theorem, so the new content is that universal pristineness and vanishing intrinsic information loss each characterize quantum error correctability. The paper then proves Theorem 4: in the Hayden-Preskill model, the channel E_{U,s} from k injected qubits to the total radiation satisfies ε(s) ≤ $2^{{k/2−s}}$ in Haar average, where ε(s) is the maximal root fidelity between outputs of orthogonal pure inputs, so the intrinsic information loss goes to zero in Haar mean as the emitted radiation size s grows. The paper notes that the resulting s∼k/2 scale for this distinguishability bound is not by itself a sharper recovery threshold than the usual reference-assisted s∼k scale.
Load-bearing premise
The load-bearing premise is the unproved stability step after Theorem 3: if the maximum information loss on a code space is at most ε, then the channel's Kraus operators approximately satisfy the Knill-Laflamme conditions and a recovery channel with fidelity 1 − O(g(ε)) exists; without this step, small information loss alone does not guarantee near-perfect recovery.
Editorial extensions
If this is right
- Quantum error-correcting codes can be characterized without constructing a recovery map: a code is correctable for a channel exactly when the channel's intrinsic information loss vanishes on the code.
- An approximate version follows: if sup_ρ K(ρ,E) ≤ ε on the code, the channel's Kraus operators approximately satisfy the Knill-Laflamme conditions and a recovery map exists with worst-case root fidelity at least 1 − O(g(ε)), so near-zero information loss guarantees near-perfect recovery.
- In the Hayden-Preskill model, the evaporation channel becomes asymptotically universally pristine as the new radiation size s grows, so an infalling state can be recovered from Hawking radiation by a channel-theoretic criterion, without an external reference system.
- The information loss K obeys a data-processing inequality, vanishes for unitary channels, and reaches its upper bound S(ρ) for completely depolarizing channels, so it interpolates between reversible and completely irreversible dynamics.
Reading between the lines
- A testable extension the paper does not pursue is to make the cited continuity step explicit, yielding a quantitative, reference-free version of approximate quantum error correction in which the recovery fidelity is a single function of the intrinsic information loss.
- Because K(ρ,E) is defined through prepare-evolve-measure scenarios, it could be measured experimentally without auxiliary reference systems by preparing an ensemble, applying the channel, performing a POVM, and estimating the conditional entropy, offering a direct probe of recoverability in engineered open systems.
- The same channel-theoretic criterion could be evaluated in semiclassical models of evaporation beyond Hayden-Preskill, replacing the random-unitary channel with a channel derived from entanglement-island calculations, to see whether the operational retrodiction time coincides with Page time.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper introduces an operational measure of information loss K(ρ,E) for prepare-evolve-measure scenarios, defined as the minimum, over pure-state decompositions of ρ and output POVMs, of the classical conditional entropy of Alice's preparation given Bob's outcome. The authors prove that a channel is "pristine" with respect to an ensemble exactly when the output states have mutually orthogonal supports, and that vanishing K for all states on a code subspace is equivalent to "universal pristineness," to the Knill-Laflamme conditions, and to perfect recoverability of the code. They then apply the framework to the Hayden-Preskill model, showing Haar-averaged bounds on a measure of non-pristineness and on the intrinsic information loss, and conclude that the evaporation channel becomes asymptotically universally pristine and therefore that infalling quantum information admits near-perfect recovery from the radiation.
Significance. The exact equivalence in Theorem 3 is a clean and potentially useful information-theoretic reformulation of quantum error correction: it identifies zero intrinsic information loss on a code with the Knill-Laflamme conditions, without introducing an auxiliary reference system. The proofs in the Supplemental Material are detailed and appear correct; the measure is defined self-containedly, and no parameters are fitted to arrive at the main equivalence. The Hayden-Preskill application is suggestive and the Haar-averaged bounds in Theorem 4 are plausibly derived, but the advertised recovery conclusion is not rigorously supported by the stated results. If the missing approximate stability step can be supplied, the framework would be a valuable diagnostic for information retrieval in black hole evaporation; as it stands, the paper's central rigorous contribution is the exact equivalence, while the application overreaches.
major comments (3)
- [Section IV, paragraph beginning 'To rigorously establish the link to approximate quantum error correction'] The paper asserts that if sup_ρ K(ρ,E) ≤ ε for states supported on the code, then the Kraus operators approximately satisfy the Knill-Laflamme conditions, and that standard continuity bounds guarantee a recovery map with worst-case root fidelity at least 1 − O(g(ε)) for some monotonically increasing g vanishing at 0. No explicit g, no derivation, and no precise hypotheses are given, and the cited references are only invoked qualitatively. This step is load-bearing: without it, the paper does not establish that small intrinsic information loss implies near-perfect recovery. The statement should be formulated as a theorem with a concrete bound and a proof, or the application should be scaled back.
- [Section V, Theorem 4 and the paragraph following it] Theorem 4 supplies only Haar-averaged statements: E_U sup_{ψ⊥φ} Fr(E_{U,s}(ψ), E_{U,s}(φ)) ≤ 2^{k/2−s} and E_U K(ρ_A, E_{U,s}) ≤ f_m(Δ_HP(ρ_A,s)). The approximate error-correction step in Section IV requires a per-channel bound sup_ρ K(ρ,E) ≤ ε, not an average over U. Even with a valid stability theorem, the hypotheses of Theorem 4 do not match: smallness of the Haar mean does not imply smallness for a typical fixed channel unless a concentration argument is supplied. The sentence 'an unknown quantum state falling into a Hayden-Preskill black hole admits near perfect recovery' therefore does not follow from the proven results as they stand.
- [Section V, definition of ϵ(s) and Theorem 4(i)] The term 'asymptotically universally pristine' is used to describe a bound on the Haar average of the root fidelity between outputs of orthogonal inputs, whereas universal pristineness in Theorem 3 is an exact, per-channel condition that E(ψ)E(ϕ)=0 for all orthogonal code states. The paper does not define what an approximate or asymptotic version of universal pristineness means in a way that connects to the exact equivalence. A precise definition of approximate universal pristineness, and a proof that ϵ(s)→0 implies the relevant approximate recovery statement for a typical channel, are needed before the Page-time retrieval claim can be regarded as a consequence of the framework.
minor comments (6)
- [Section VI, Discussion] The word 'pristiness' appears in the first paragraph of the Discussion; it should be 'pristineness'.
- [Section II, paragraph following Theorem 1] The sentence contains 'the the intrinsic information loss'; the duplicated article should be removed.
- [Section III, first paragraph] The spelling 'probabilisitic' should be corrected to 'probabilistic'.
- [Section II, first paragraph] The phrase 'Bob's obtains' should read 'Bob obtains'.
- [Throughout] The notation for the code subspace is inconsistent: the main text uses both H_code and H code, and the Supplemental uses H_code. One notation should be used consistently.
- [Section IV, statement of Theorem 3] The sentence 'As the equivalence of conditions iii and iv have been established' should be 'has been established', and it would help to state explicitly that the cited Knill-Laflamme theorem is being invoked rather than proved.
Circularity Check
No material circularity: the central equivalences are proved from stated definitions and standard external theorems; the only self-citation is motivational and not load-bearing.
full rationale
The paper's central derivation is self-contained rather than circular. The intrinsic information loss K(rho, E) is defined directly as a minimization of conditional entropy over pure-state ensemble decompositions and output POVMs, and universal pristineness is an independently stated geometric condition on supports of channel outputs. Theorem 1 is proved from Shannon entropy properties; Theorem 2 from the structure of retrodiction maps; Theorem 3 is proved in the Supplemental Material from these definitions, trace-distance contractivity, and the externally established Knill-Laflamme theorem; and Theorem 4 follows from Haar-moment calculations and the Barnum-Knill bound. None of these steps fits a parameter to the target conclusion or defines one condition in terms of the other. The only self-citation is Fullwood and Parzygnat [12], used in the introduction as motivation for the axiomatic approach to information loss; no theorem or application reduces to it. The Hayden-Preskill application does contain an unproved continuity assertion: the passage beginning 'To rigorously establish the link to approximate quantum error correction' states that sup_rho K(rho,E) <= epsilon forces approximate Knill-Laflamme conditions and hence recovery fidelity 1 - O(g(epsilon)), without specifying g or providing a derivation. This is a genuine gap in the paper's advertised conclusion of near-perfect recovery, but it is a completeness or correctness gap, not circularity: the missing continuity statement is an external approximate-QEC result, not a restatement of the paper's own definitions or fitted quantities. The paper even honestly flags the finite-error scaling distinction in the paragraph beginning 'At first sight, the bound epsilon(s) <= 2^{k/2-s}'. Because no central claim reduces by construction to its inputs, the circularity score is low; the value 2 reflects only the presence of one minor, non-load-bearing self-citation.
Assumptions & free parameters
assumptions (6)
- domain assumption Quantum channels are completely positive trace-preserving maps between finite-dimensional Hilbert spaces.
- standard math Standard Shannon entropy properties: chain rule, Fano's inequality, and majorization of probability distributions by eigenvalues.
- standard math Knill-Laflamme theorem: condition P E_alpha^dagger E_beta P = c_alpha beta P is equivalent to the existence of a perfect recovery channel for the code.
- standard math Barnum-Knill bound on minimum-error state discrimination via the pretty-good measurement.
- domain assumption Approximate quantum error correction continuity: if P E_alpha^dagger E_beta P is close to c_alpha beta P, then a recovery channel with fidelity 1 - O(g(epsilon)) exists.
- domain assumption The Hayden-Preskill model assumes a Haar-random unitary models the scrambling of the black hole, and Theorem 4 averages over the Haar measure.
Cite this review
Pith. "Pith review of Quantum information loss." pith.science (2026). https://pith.science/paper/PM6QNYAS
@misc{pith2026260805535,
author = {Pith},
title = {Pith review of: Quantum information loss},
year = {2026},
howpublished = {\url{https://pith.science/paper/PM6QNYAS}},
note = {Machine review of arXiv:2608.05535}
}
abstract
We introduce a measure of information loss for any quantum process that may be modeled by a prepare-evolve-measure scenario: Alice prepares an ensemble of states that gets sent via a quantum channel to Bob, who then measures the output. As a quantum channel models open system dynamics, our measure of information loss quantifies Bob's inability to retrodict with certainty which state Alice sent through the channel. By minimizing this measure over all possible pure state ensemble decompositions of a fixed state $\rho$, and over all POVMs on the output of a channel $\mathcal{E}$, we arrive at an intrinsic notion of information loss for any state-channel pair $(\rho,\mathcal{E})$. We show that the vanishing of information loss with respect to all states supported on a fixed codespace $\mathcal{H}_{\text{code}}$ is equivalent to a condition we term \emph{universal pristineness}, which ensures that orthogonal pure states in $\mathcal{H}_{\text{code}}$ get sent via the channel $\mathcal{E}$ to possibly mixed states whose supports are orthogonal. Moreover, we prove universal pristineness is equivalent to the Knill-Laflamme conditions in quantum error correction, which are necessary and sufficient for the existence of a perfect recovery channel for all states supported on $\mathcal{H}_{\text{code}}$. As an application, we apply our framework to the Hayden-Preskill model of black hole evaporation, demonstrating that the evaporation channel becomes asymptotically universally pristine, thereby providing a purely channel-theoretic formulation of Page-time information retrieval.
Figures
Reference graph
Works this paper leans on
-
[1]
H. D. Zeh, On the interpretation of measurement in quan- tum theory, Found. Phys.1, 69 (1970)
1970
-
[2]
E. Joos and H. D. Zeh, The emergence of classical prop- erties through interaction with the environment, Z. Phys. B - Condensed Matter59, 223 (1985)
work page 1985
-
[3]
R. Landauer, Irreversibility and heat generation in the computing process, IBM Journal of Research and Devel- opment5, 183 (1961). 7
work page 1961
-
[4]
W. H. Zurek, Decoherence, einselection, and the quan- tum origins of the classical, Reviews of Modern Physics 75, 715 (2003)
2003
-
[5]
P. Busch, No Information Without Disturbance: Quan- tum Limitations of Measurement, inQuantum Reality, Relativistic Causality, and Closing the Epistemic Circle: Essays in Honour of Abner Shimony(Springer Nether- lands, Dordrecht, 2009) pp. 229–256
work page 2009
-
[6]
S. W. Hawking, Breakdown of predictability in gravita- tional collapse, Phys. Rev. D14, 2460 (1976)
1976
-
[7]
A. Akil and C. Bambi, eds.,The Black Hole Information Paradox: A Fifty-Year Journey, Springer Series in As- trophysics and Cosmology (Springer Nature Switzerland AG, 2025)
work page 2025
-
[8]
Landauer, The physical nature of information, Phys
R. Landauer, The physical nature of information, Phys. Lett. A217, 188 (1996)
work page 1996
Show all 32 references
-
[9]
Y. Li, S. Luo, Y. Sun, and S. Wang, Quantifying informa- tion loss of quantum channels, Phys. Rev. A112, 062440 (2025)
2025
-
[10]
N. J. Cerf, Entropic bounds on coding for noisy quantum channels, Phys. Rev. A57, 3330 (1998)
1998
-
[11]
J. C. Baez, T. Fritz, and T. Leinster, A characteriza- tion of entropy in terms of information loss, Entropy13, 1945–1957 (2011)
2011
-
[12]
Fullwood and A
J. Fullwood and A. J. Parzygnat, The information loss of a stochastic map, Entropy23, 10.3390/e23081021 (2021), arXiv:2107.01975 [cs.IT]
2021 arXiv
-
[13]
Knill, R
E. Knill, R. Laflamme, and L. Viola, Theory of quan- tum error correction for general noise, Physical Review Letters84, 2525–2528 (2000)
2000
-
[14]
Hayden and J
P. Hayden and J. Preskill, Black holes as mirrors: quan- tum information in random subsystems, Journal of High Energy Physics2007, 120–120 (2007)
2007
-
[15]
D. N. Page, Information in black hole radiation, Physical Review Letters71, 3743–3746 (1993)
1993
-
[16]
T. M. Cover and J. A. Thomas,Elements of Information Theory, 2nd ed. (Wiley-Interscience, Hoboken, NJ, 2006)
2006
-
[17]
Nielsen and G
M. Nielsen and G. Vidal, Majorization and the intercon- version of bipartite states, Quant. Inf. Comput.1, 76 (2001)
2001
-
[18]
Knill and R
E. Knill and R. Laflamme, Theory of quantum error- correcting codes, Phys. Rev. A55, 900 (1997)
1997
-
[19]
B´ eny and O
C. B´ eny and O. Oreshkov, General conditions for approx- imate quantum error correction and near-optimal recov- ery channels, Phys. Rev. Lett.104, 120501 (2010)
2010
-
[20]
H. K. Ng and P. Mandayam, Simple approach to approx- imate quantum error correction based on the transpose channel, Phys. Rev. A81, 042323 (2010)
2010
-
[21]
S. W. Hawking, Particle creation by black holes, Comm. Math. Phys.43, 199 (1975)
1975
-
[22]
Yoshida and A
B. Yoshida and A. Kitaev, Efficient decoding for the Hayden-Preskill protocol (2017), arXiv:1710.03363 [hep- th]
2017 arXiv
-
[23]
Penington, Entanglement wedge reconstruction and the information paradox, J
G. Penington, Entanglement wedge reconstruction and the information paradox, J. High Energy Phys.2020(2), arXiv:1905.08255 [hep-th]
2020 arXiv
-
[24]
Almheiri, N
A. Almheiri, N. Engelhardt, D. Marolf, and H. Maxfield, The entropy of bulk quantum fields and the entangle- ment wedge of an evaporating black hole, JHEP12, 063, arXiv:1905.08762 [hep-th]
1905 arXiv
-
[25]
Almheiri, R
A. Almheiri, R. Mahajan, J. Maldacena, and Y. Zhao, The page curve of hawking radiation from semiclassical geometry, JHEP03, 149, arXiv:1908.10996 [hep-th]
1908 arXiv
-
[26]
Penington, S
G. Penington, S. H. Shenker, D. Stanford, and Z. Yang, Replica wormholes and the black hole interior, JHEP03, 205, arXiv:1911.11977 [hep-th]
1911 arXiv
-
[27]
Almheiri, T
A. Almheiri, T. Hartman, J. Maldacena, E. Shaghou- lian, and A. Tajdini, Replica wormholes and the entropy of hawking radiation, JHEP05, 013, arXiv:1911.12333 [hep-th]
1911 arXiv
-
[28]
Reeb and M
D. Reeb and M. M. Wolf, An improved landauer principle with finite-size corrections, New Journal of Physics16, 103011 (2014)
2014
-
[29]
Barnum and E
H. Barnum and E. Knill, Reversing quantum dynam- ics with near-optimal quantum and classical fidelity, J. Math. Phys.43, 2097 (2002), arXiv:quant-ph/0004088 [quant-ph]. 8 Quantum information loss — Supplemental Material — In this Supplemental Material we provide proofs of Theo...
2002
-
[30]
There exists a positive semi-definite matrixC= (c αβ) such thatP E† αEβP=c αβPfor all Kraus indicesα, β
-
[31]
Proof.Throughout this proof, we writeψ:=|ψ⟩⟨ψ|for a pure state
There exists a quantum channelR:L(H B)→ L(Hcode) such that (R ◦ E)(ρ) =ρfor every stateρsupported onH code. Proof.Throughout this proof, we writeψ:=|ψ⟩⟨ψ|for a pure state. We shall repeatedly use that, for positive semidefinite operatorsX, Y≥0, Tr(XY) = 0⇐ ⇒XY= 0.(C1) Indeed,X...
-
[32]
3.E U K(ρ A,E U,s)≤f m ∆HP(ρA, s)
∆ HP(ρA, s)≤C p 2k/2−s. 3.E U K(ρ A,E U,s)≤f m ∆HP(ρA, s) . In particular,E U,s is asymptotically universally pristine onH A, and the intrinsic information lossK(ρ A,E U,s) ap- proaches 0 (in Haar mean) for larges. Proof.Let |Φ⟩HR = 1√dH dHX a=1 |a⟩H |a⟩R,Φ HR :=|Φ⟩⟨Φ| HR , an...
Reviewed August 8, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.