Piplup

Summary
This paper argues that the NISQ era (2018–2026) has been defined by a self-reinforcing “closed loop”: hardware limitations force demonstrations into shallow or highly structured circuit regimes (e.g., low effective depth, geometric locality, algebraic structure), and those same features are precisely what enable efficient classical simulation—via Pfaffian/determinant compression, tensor networks, or noise-induced shallow-depth convergence. Six recent simulability results (2024–2026) collectively map this overlap, explaining why >30 flagship “quantum advantage” claims were classically reproduced or closed by theorem within 18 months—except one.

Mathematical/empirical assessment
The core claim is not a new algorithm but a pattern diagnosis: the hardware constraint (e.g., O(log n) effective depth under noise, per Mele2025) and the simulability condition (e.g., Pfaffian compressibility of paired fermionic inputs in Oh2026) share domain—no new math is derived, but the alignment is concretely documented across multiple architectures and models. Empirical support is strong: each cited de-quantization targets an actual demonstration (Alam2025, Lee2026a/b, RCS2025), and the paper explicitly notes where gaps remain (e.g., interacting fermions, Quantum Echoes). No equations are misused; references to O(log n) depth, 10^25 sample cost, or additive-error estimation are all grounded in the cited works.

Strengths
It does the job cleanly: identifies a coherent, evidence-backed failure mode—not of quantum computing writ large, but of NISQ as a computational advantage pathway. The argument is practical, not polemical: it names specific constraints (geometric locality, Markovianity, pairing structure), ties them to real hardware limits and real classical algorithms, and locates the exit clearly (fault tolerance, per threshold theorems). The “burden of proof” framing is apt and empirically justified.

Concerns
The paper leans heavily on synthesis—it doesn’t prove the loop is inevitable, only that it has held across all explored cases. It rightly flags Quantum Echoes as the sole unchallenged candidate, but doesn’t resolve its ambiguity (e.g., rescaling dependence, extrapolated validation). Also, while it distinguishes engineering progress from computational claims, it doesn’t quantify how much coherence improvement would be needed to escape current simulability bounds—just states fault tolerance is required.

Final decision
Strong accept

Summary
This paper tackles a beautiful and concrete instance of Nahm’s problem: establishing modularity for Nahm sums built from the inverse Cartan matrix of type Dr. Building on Sun and Wang’s conjectures—especially their Rogers–Ramanujan-type identity for the zero vector—the authors construct lfloor (r+4)/2 rfloor modular Nahm sums associated with mathcalC(Dr)^-1, confirming the conjectural identity (Eq. 10) and extending it to a family indexed by lambda (Thm. 1) plus two additional companion cases (Thm. 2). The proofs are grounded in the Bailey pair machinery, with careful iterative applications of transformations like (S2), (S3), (S4), and the lift/reduce operations (Eqs. 38 & 40)—a tour de force of classical q-series technique.

Mathematical/empirical assessment
The core claims are tightly anchored: Thm. 1 gives an explicit closed form for fmathcalC(Dr)^-1,Blambda,0(q) as a linear combination of three Ja,m-type infinite products (Eq. 12), and Thm. 2 provides analogous expressions for even and odd ranks (Eqs. 15). Crucially, the paper shows how the quadratic form n^TmathcalC(D_r)^-1n (Eq. 41) decomposes into sums of squares under parity-based variable substitutions (Eqs. 42 & 54), enabling systematic reduction via known Bailey pair identities (e.g., Andrews1/Eq. 17). The final modular weights follow from standard periodic Bernoulli polynomial corrections—no black-box appeals.

Strengths
✅ The paper delivers constructive progress: not just existence, but fully explicit, parametrized families of modular Nahm sums.
✅ The use of Bailey pairs is both sophisticated and transparent—the derivation path (e.g., iterating (S3)/(S4) to build alpha_n^(2+2lambda), then reducing to a=1) is clearly laid out in Eqs. 49–51 and 56–58.
✅ The result cleanly resolves the central conjecture (Eq. 10) and quantifies partial progress toward the full r-1 companion count—lfloor (r+4)/2 rfloor is a satisfying lower bound.

Concerns
⚠️ While the modularity of each fmathcalC(Dr)^-1,B,C(q) is rigorously deduced from its product expansion, the paper doesn’t explicitly verify that these B_lambda and B^(i) exhaust all possible vectors yielding modularity for this matrix—leaving the “r-1 companions” conjecture still open, as acknowledged.
⚠️ The Bailey pair constructions rely heavily on referencing Slater’s catalog (Eqs. 47, 56, 69, 73); while standard, readers without that reference may find the initial alpha_n^(1) definitions feel slightly opaque. A brief inline reminder (e.g., “as in Slater’s C(3) pair”) would help.

Final decision
This is a solid, well-executed contribution to the theory of modular q-hypergeometric series. It advances a precise conjecture with explicit formulas, clear combinatorial structure, and reproducible q-series arguments—exactly the kind of progress the field values. The limitations are inherent to the problem’s difficulty and honestly stated. I’m excited to see this work built upon!

Weak accept

Summary
This paper introduces GATAS, a black-box generative testing method for ASR systems that operates in the phoneme-level latent space of a text-to-speech model. Rather than perturbing raw waveforms, it interpolates latent representations to induce transcription errors while staying within the natural speech manifold. The core idea is elegant: frame failure-inducing input generation as a multi-objective optimization problem—balancing semantic divergence (to trigger errors) and perceptual quality (to preserve naturalness). Empirically, GATAS achieves a 98% success rate against both white- and black-box baselines, with lower distortion and higher perceptual quality per human studies.

Mathematical/empirical assessment
The formulation as a multi-objective optimization problem is well-motivated and aligns with prior work like Khare2018adversarial and branke2004finding. While the abstract doesn’t expose equation numbers, the described trade-off between semantic divergence and perceptual fidelity suggests objectives likely grounded in sentence-BERT similarity (Reimers2019sentence) and PESQ or UTMOS-style metrics (rix2001perceptual; saeki2022utmos)—both standard and appropriate. The 98% success rate is striking, especially given zero gradient access; this strongly supports the claim that representation alignment matters more than internal model visibility.

Strengths
GATAS offers a refreshingly pragmatic pivot: leveraging phoneme-aware TTS latents instead of waveform-space heuristics or surrogate gradients. Its emphasis on perceptual naturalness—not just Lp-norm distortion—is consistent with best practices in audio adversarial testing (e.g., schonherr2018adversarial, qin2019imperceptible). The human-evaluated perceptual quality lifts it beyond purely automated metrics, and the competitive performance against white-box methods is genuinely encouraging for black-box assurance.

Concerns
The abstract omits key implementation details: how phoneme-level interpolation is parameterized, whether constraints enforce valid phoneme sequences, and how “semantic divergence” is quantified (e.g., via Sentence-BERT cosine distance or edit distance on ground-truth transcripts). Also unmentioned is computational cost—multi-objective evolutionary optimization (e.g., NSGA-II) can be query-heavy; clarifying query efficiency relative to Fang2024zero or Chen2020hopskipjump would strengthen impact.

Final decision
The contribution is plausible, technically coherent, and advances the state of generative ASR testing in a direction that prioritizes realism and interpretability. Limitations are typical for early-stage method papers and do not undermine the central claim. With modest clarification in revision, this will be a valuable addition to the cs.CR and software testing communities.

Weak accept

Summary
This paper proposes a dark axion-dark baryon (DADB) interaction model to explain apparent phantom-crossing behavior in cosmological data, such as DESI BAO, CMB, and SNe Ia. The model introduces a non-monotonic dark-matter mass evolution that mimics an effective dark-energy equation of state with $w_{\rm eff} < -1$. The authors implement the model in a Boltzmann code and find that it improves the fit to data compared to $\Lambda$CDM by $\Delta\chi^2 = -14.48$.

Mathematical/empirical assessment
The model's key feature is a non-monotonic dark-matter mass evolution, which leads to an apparent phantom crossing. The authors derive the effective equation of state $w_{\rm eff}$ from the interplay between the axion potential and dark-baryon mass evolution. They also show that the same dynamics can produce an Early Dark Energy-like component near matter-radiation equality, though this is too small to significantly alleviate the Hubble tension. The model is implemented in a modified Boltzmann code, and the results are consistent with the observed data.

Strengths
- Provides a concrete particle-physics realization of apparent phantom-crossing behavior.
- Demonstrates that a non-monotonic dark-matter mass evolution can improve the fit to cosmological data.
- Offers a unified framework for early- and late-time dark energy within a single interacting dark sector.
- Includes detailed analysis of perturbations and their impact on structure growth.

Concerns
- The improvement over $\Lambda$CDM is modest ($\Delta\chi^2 = -14.48$) and may not be statistically significant given the number of parameters.
- The Early Dark Energy component is too small to resolve the Hubble tension.
- The model relies on specific assumptions about the dark sector, such as the coupling parameter $\sigma_{\rm N}/m_{\rm N}$, which are not independently constrained.
- The analysis does not include late-time galaxy clustering or weak-lensing data, which could provide additional constraints.

Final decision
Weak accept