REVIEW
Robust exponential lower bounds for fermionic and bosonic Gaussian ranks
Quantum Physics · Submitted 1 Oct 2026 · 49 references
Pith's verdict: Resolves the Gaussian analogue of the stabilizer-rank exponential lower bound problem with clean, checkable proofs; the bosonic hinge is real but holds up on inspection.
T0 · AI review
Why Pith says this
Highlight any text to comment on it.
Referee report
The paper
The paper claims
This paper proves that for every fixed pure non-Gaussian state on finitely many fermionic or bosonic modes, the approximate border Gaussian rank of its tensor powers grows exponentially in the number of copies, at any fixed norm error…
The reading
What carries the argument
Two independent mechanisms carry the proof. For fermions, the target is reduced to four modes by Gaussian postselection, producing a normal form $|M_t\rangle = (|0000\rangle + t|1111\rangle)/\sqrt{1+t^2}$ with success probability $p_\psi > 0$, and a lower bound is obtained by comparing ranks of the matrix $F_p(|\phi\rangle)$ whose columns are degree-$p$ Majorana monomials applied to $|\phi\rangle$: on $k$ copies of $|M\rangle$ this matrix has an orthonormal block of $D_{k,p}$ columns, while on any Gaussian state it has rank at most $C_{4k,p}$, a fact resting on the osculating-space structure of spinor varieties; the Eckart-Young theorem then converts the matrix-rank gap into a bound on the state-approximation error. For bosons, the target is postselected near a Fock number state, coupled to a vacuum environment by a beam-splitter and two-mode squeezer isometry, and a photon cutoff bounds each Gaussian term's Schmidt rank by $\binom{n+p}{p}$, while a uniform Renyi-entropy gap $c_{\mathcal{K}} = \inf_{|\phi\rangle \in \mathcal{K}_\psi} H_s(\rho_\phi) - d_n(\mu) > 0$ of the environment above the thermal-entropy benchmark $d_n(\mu)$ forces exponentially many Schmidt coefficients to be needed; the entropy rank bound then sets the exponential base, and a typicality argument over the postselection record transfers the bound back to the original tensor powers.
Extended reading notes
Core claim
The central claim, Theorem 1, is that for any fixed normalized pure non-Gaussian state $|\psi\rangle$ on finitely many modes there is a positive rate $c_\psi$ such that the approximate border Gaussian rank $\chi_{\mathrm{G}}^{(\delta)}(|\psi\rangle^{\otimes k})$ is at least $e^{c_\psi k}$ for all sufficiently large $k$, uniformly for every fixed error $0 \le \delta < 1$, with definite parity assumed in the fermionic case. Because the approximate border rank dominates the exact, border, and approximate notions through the chain $\chi_{\mathrm{G}}^{(\delta)} \le \bar{\chi}_{\mathrm{G}}^{(\delta)} \le \overline{\chi}_{\mathrm{G}} \le \chi_{\mathrm{G}}$, one bound controls all four rank notions simultaneously. The rate $c_\psi$ is independent of the error tolerance, though the threshold in $k$ may depend on it. For the two model states the paper derives explicit exponential bases: $343/243$ for the four-mode fermionic state $|M\rangle = (|0000\rangle + |1111\rangle)/\sqrt{2}$, and $11/10$ (with $(5/4)^k$ at zero error) for the bosonic single-photon state $|1\rangle$.
Load-bearing premise
For the bosonic half of the theorem, the exponential rate rests on the assertion that every state produced by the successful postselection channel keeps its environment Renyi entropy strictly above the thermal-entropy benchmark $d_n(\mu)$ by a uniform positive amount, a gap established by a specific beam-splitter/squeezer isometry, a small-coupling limit, and continuity in the weighted norm $\|\cdot\|_D$ — if that gap collapsed under perturbation of the outputs near the number state, the bosonic exponential bound, and with it the no-energy-assumption scope, would fail.
Editorial extensions
If this is right
- Decomposition-based classical simulation of any fixed non-Gaussian state, fermionic or bosonic, provably costs exponentially many Gaussian terms in the number of copies at any fixed norm error below one, so non-Gaussianity is a hard resource on its own.
- The single-photon state, the natural continuous-variable resource, has approximate border Gaussian rank at least $(11/10)^k$ and exact border rank above $(5/4)^k$, placing limits on Fock-state-based bosonic simulation.
- Gaussian rank and border Gaussian rank genuinely differ in both settings: the paper exhibits a six-mode fermionic state with border rank $2$ and exact rank $3$, and confirms the single-photon state has border rank $2$ but infinite exact rank.
- The dichotomy for pure states is sharp: Gaussian states have rank one in every tensor power, while every non-Gaussian target has exponentially growing tensor-power rank, which the paper conjectures reflects a structural property of the free family rather than details of the specific states.
Reading between the lines
- The template may transfer to the stabilizer-rank problem: if the exponential-growth certificate here lies in a structural property of free families (Majorana matrices for fermions, environment spectra for bosons), then identifying an analogous certificate for stabilizer states would go a long way toward the open exponential stabilizer-rank conjecture, which the paper explicitly leaves as a directi
- The bosonic construction, which needs no energy assumption, transfers hardness to nearly-Gaussian low-photon states; a testable consequence is that even weak non-Gaussianity, such as a single photon among many vacuum modes or a small squeezing tail, should force exponential Gaussian decomposition complexity, because the entropy gap persists in a neighborhood of the Fock state in the weighted norm
- The postselection transfer works by typicality of the success count, so the lower bound should hold not only for identical copies but for any tensor product drawn from the successful family; this suggests the exponential bound is stable under mixing the target with vacuum or weak noise, a robustness the authors do not state explicitly.
- Because the rate $c_\psi$ is uniform in the error $\delta$, the lower bound reflects structure of the state itself rather than the chosen approximation error; matching constructions against the finite-copy values in Table I would settle whether the optimal rate is $\delta$-independent, the question flagged in the outlook.
Original abstract
The power and limitations of classical simulation are central to understanding quantum computational advantages. A leading simulation paradigm is based on coherent decomposition into classically tractable free states where the decomposition rank determines the simulation complexity. Proving strong lower bounds on this number is a notoriously difficult and mathematically rich problem, as exemplified by the qubit stabilizer rank problem. Here, we study the Gaussian version of this problem in both bosonic and fermionic systems and establish robust exponential lower bounds on Gaussian rank. In particular, we prove that for every pure non-Gaussian state on finitely many modes, with definite parity in the fermionic case, the approximate border Gaussian rank of its tensor powers grows at least exponentially at any fixed norm error below one. Our proofs combine reduction to four modes with Majorana spectral bounds for fermions, and Gaussian postselection with an entropy-based rank bound for bosons. As concrete examples, we derive explicit exponential lower bounds for the four-mode fermionic GHZ state and the bosonic single-photon state. The bosonic results require no assumption on the energy of the target state. Our results show that non-Gaussianity universally entails exponential Gaussian decomposition complexity, setting fundamental limitations on decomposition-based classical simulation of bosonic and fermionic systems.
The compiled record
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
The abstract, sentence by sentence
The power and limitations of classical simulation are central to understanding quantum computational advantages. A leading simulation paradigm is based on coherent decomposition into classically tractable free states where the decomposition rank determines the simulation complexity. Proving strong lower bounds on this number is a notoriously difficult and mathematically rich problem, as exemplified by the qubit stabilizer rank problem. Here, we study the Gaussian version of this problem in both bosonic and fermionic systems and establish robust exponential lower bounds on Gaussian rank. In particular, we prove that for every pure non-Gaussian state on finitely many modes, with definite parity in the fermionic case, the approximate border Gaussian rank of its tensor powers grows at least exponentially at any fixed norm error below one. Our proofs combine reduction to four modes with Majorana spectral bounds for fermions, and Gaussian postselection with an entropy-based rank bound for bosons. As concrete examples, we derive explicit exponential lower bounds for the four-mode fermionic GHZ state and the bosonic single-photon state. The bosonic results require no assumption on the energy of the target state. Our results show that non-Gaussianity universally entails exponential Gaussian decomposition complexity, setting fundamental limitations on decomposition-based classical simulation of bosonic and fermionic systems.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | The power and limitations of classical simulation are central to understanding quantum computational advantages. | background 1.00 | 0.04 | 0.02 | asserted 1.00 |
| s1 | A leading simulation paradigm is based on coherent decomposition into classically tractable free states where the decomposition rank determines the simulation complexity. | background 1.00 | 0.04 | 0.02 | asserted 1.00 |
| s2 | Proving strong lower bounds on this number is a notoriously difficult and mathematically rich problem, as exemplified by the qubit stabilizer rank problem. | background 0.94 | 0.04 | 0.02 | asserted 1.00 |
| s3 | Here, we study the Gaussian version of this problem in both bosonic and fermionic systems and establish robust exponential lower bounds on Gaussian rank. | contribution 0.97 | 0.95 | 0.66 | asserted 1.00 |
| s4 | In particular, we prove that for every pure non-Gaussian state on finitely many modes, with definite parity in the fermionic case, the approximate border Gaussian rank of its tensor powers grows at least exponentially at any fixed norm error below one. | result 1.00 | 0.95 | 0.68 | asserted 1.00 |
| s5 | Our proofs combine reduction to four modes with Majorana spectral bounds for fermions, and Gaussian postselection with an entropy-based rank bound for bosons. | method 1.00 | 0.93 | 0.04 | asserted 0.99 |
| s6 | As concrete examples, we derive explicit exponential lower bounds for the four-mode fermionic GHZ state and the bosonic single-photon state. | result 0.97 | 0.96 | 0.06 | asserted 1.00 |
| s7 | The bosonic results require no assumption on the energy of the target state. | result 0.79 | 0.88 | 0.04 | asserted 1.00 |
| s8 | Our results show that non-Gaussianity universally entails exponential Gaussian decomposition complexity, setting fundamental limitations on decomposition-based classical simulation of bosonic and fermionic systems. | interpretation 1.00 | 0.92 | 0.92 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
References
Discussion
Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.