{"id":"45d7add7-cf98-48f0-8192-b1578b00e0bb","arxiv_id":"2502.09612","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Random spin chains with XX+YY interactions show spin survival probability decaying slower than any power law, approximately as 1/log^2(t), captured by a new superspin RSRG-X formalism.","lead":"The paper develops an excited-state real-space renormalization group method for random spin chains, describing slow relaxation through collective 'superspin' flips. It matters because it provides a way to simulate late-time dynamics in disordered dipolar quantum simulators beyond the reach of exact diagonalization.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"For the long-range dipolar chain, the 1/log^2 asymptote is decided in a frequency window below the ED benchmark and depends on an unproven strong-randomness flow for the newly generated three-superspin and superspin-0 couplings; the paper's own text concedes the thermodynamic limit is inconclusive…","rationale":"The paper is doing something valuable: it constructs a closed RSRG-X for U(1)+Z2 random XX+YY models and benchmarks it against ED. The benchmark is real and fairly convincing in the window where both methods overlap. The load-bearing weak point is not an internal contradiction but the reach of the central late-time claim beyond that window. To get S_p(t) ~ B + A/log^2 t in the long-range chain one must believe two nested steps: (i) the superspin RSRG-X truncation is asymptotically exact in the low-frequency limit, and (ii) the N=40 simulation in Fig. 11b has already reached the universal regime. Concern (i) is the reader's weakest assumption; the paper's evidence is the Q<0 probability, which is necessary but not sufficient for control of the omitted couplings. Concern (ii) is acknowledged by the text: 'due to finite-size effects we are unable to confirm such scaling in the thermodynamic limit.' Either one failing would remove the long-range half of the central claim, while the nearest-neighbor and next-nearest-neighbor claims would survive. I therefore do not recommend rejecting or changing the conditional verdict; the verdict should remain conditional until the proposed diagnostics are run. The concrete test directly probes the suppression of three-superspin and superspin-0 couplings, and the threshold cross-check probes whether the numerical extrapolation is stable.","tokens_in":56208,"tokens_out":7497,"duration_ms":85130,"concrete_test":"During the Monte Carlo RSRG-X run used for Fig. 11b, record at every RG step with energy scale below 10^-4 the operator norm of the sum of all overlapping H1 terms that are three-superspin or superspin-0 couplings, normalized by ||H0||. Compute the median and 90th percentile of this ratio as a function of RG time for the long-range chain at N=16, 24, 32, and 40. If the ratio does not decrease with system size or RG time, or the 90th percentile exceeds roughly 0.1, the truncation is not controlled and the 1/log^2 fit cannot be trusted. As a cross-check, rerun the N=40 calculation with epsilon_thres reduced by a factor of 10^3; if omega Sp(omega) below 10^-5 shifts by more than the sample-to-sample error, the quoted threshold is not inert.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central prediction for interacting chains is built on Eq. (39): every RSRG-X state contributes independent coherent spin flips at frequencies 4Omega_n,j, with all other processes ignored. This picture is exact only if H0 dominates H1 at every step and if the terms outside the closed set listed in Eq. (35) - three-superspin couplings and X(0)X(0)/Y(0)Y(0) couplings - remain dynamically irrelevant. The validation in Appendix E reports only the probability Q<0, i.e., that some overlapping term is stronger than H0 (Table I). That does not establish that the ratio ||H1||/||H0||, or the accumulated effect of many weaker omitted couplings, flows to zero; no flow analysis is given, and the conclusion states that tracking three-site interactions remains open. The first RG step already generates three-superspin terms (Eq. 29). If these are weak but numerous, they can shift resonance frequencies or couple supposedly independent spin flips at late times. For the long-range chain, ED agreement (Fig. 10b) is only at N=16 and omega >= 1e-6, while the [log(omega0/omega)]^-3 fit encoding 1/log^2 t is extracted at N=40 below that window (Fig. 11b), where the text explicitly says finite-size effects prevent confirmation in the thermodynamic limit. Thus the abstract's no-deviation claim for the long-range model rests on exactly the least benchmarked regime of the least controlled approximation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops an excited-state real-space renormalization group (RSRG-X) formalism for random spin-1/2 systems with U(1) and Z2 symmetries, applied to XYZ? Actually XX+YY chains with nearest-neighbor, next-nearest-neighbor, and long-range dipolar interactions. The formalism represents the effective Hamiltonian in terms of two-level 'superspins' and predicts that the disorder-averaged spin survival probability decays at late times as S_p(t) ≈ B + A/log^2(t/t0), slower than any power law. For nearest-neighbor chains the prediction is derived analytically from the known infinite-randomness fixed point. For interacting chains the paper introduces a numerical RSRG-X algorithm and benchmarks it against exact diagonalization (ED) for N up to 16, finding quantitative agreement at low frequencies for NNN and LR chains. The paper also presents ED and RSRG-X results for two-dimensional power-law interacting models. The central claims are that 1D chains show the 1/log^2 asymptote, that the superspin formalism captures the dynamics, and that 2D systems display slow subdiffusive relaxation.","tokens_in":56585,"tokens_out":3287,"duration_ms":36948,"significance":"If the central claims hold, the paper makes a valuable contribution to the theory of slow dynamics in disordered spin systems: it extends RSRG-X to genuinely interacting long-range chains, gives a physical picture of many-body resonances, and provides a numerical method that reaches sizes beyond ED. The analytical derivation for the nearest-neighbor chain is clean and is supported by ED benchmarks at multiple system sizes. The quantitative ED agreement for the NNN and LR chains at accessible frequencies is a genuine strength, as is the explicit reporting of the RG-quality statistics in Appendix E. The main significance is therefore real, but it is currently weakened by an abstract-level overstatement for the long-range chain and by an unverified RG premise for the interacting case.","major_comments":[{"comment":"The abstract states that the results 'feature no significant deviation from the ~1/log^2(t) asymptote' for the long-range chain. This is stronger than what the paper establishes. The N=40 RSRG-X data in Fig. 11b show a fit to the universal form (7) only in the frequency window ω ≲ 10^-6, which is below the window 10^-6 ≲ ω ≲ 10^-2 where RSRG-X was benchmarked against ED (Fig. 10b). The text itself says that due to finite-size effects the scaling 'cannot be confirmed in the thermodynamic limit' (Sec. V B 3). Since the asymptotic statement is a central advertised result, the claim must either be supported by a convergence or finite-size scaling analysis, or explicitly downgraded to a tentative suggestion.","section":"Abstract and Sec. V B 3, Fig. 11b"},{"comment":"The self-contained nature of the superspin RSRG-X formalism rests on the assertion that interactions involving three or more superspins, and X(0)X(0) or Y(0)Y(0) terms, are negligible (Eq. (35) and surrounding text). The evidence provided in Appendix E and Table I is the probability that H0 overlaps with a stronger interaction (Q<0). This is not the same as demonstrating that the norm ratio ||H1||/||H0|| flows to zero, or that the accumulated effect of many weak omitted couplings does not shift resonance frequencies or couple supposedly independent spin flips at late times. The paper's own conclusion states that tracking three-site interactions remains an open problem. This gap directly affects the central prediction of Eq. (39) for the interacting chains, so the manuscript should either provide a flow analysis or clearly limit the claimed regime of validity.","section":"Sec. V A 3 and Appendix E"},{"comment":"For the two-dimensional model, Table I shows Q<0 probabilities up to 1.1% (N=16, α=3), and App. Fig. 13 shows that the RG quality distribution is substantially worse than in 1D. The paper acknowledges this and appropriately calls the 2D results preliminary. However, because the same superspin RSRG-X is applied in 2D without a separate validation of the strong-randomness premise, the 2D conclusions (power-law, subdiffusive decay with extracted exponents c) should be presented with a clearer caveat that the RG framework itself is less controlled in that setting. At present, the abstract's mention of 2D results is fair, but Sec. VI should more prominently state that the ED agreement is only qualitative for α ≤ 4 and that the numerical exponents should be regarded as effective rather than asymptotic.","section":"Sec. VI and Table I"}],"minor_comments":[{"comment":"In Eq. (D10), the expression for V- is identical to that for V+ (both written with a plus sign). This is presumably a typo; the V- state should use the minus sign, as correctly stated in Eq. (32) of the main text.","section":"Appendix D, Eq. (D10)"},{"comment":"The abstract says 'feature no significant deviation' for the long-range model, whereas Sec. V B 3 says the result is 'inconclusive'; please align the wording with the actual evidence.","section":"Sec. V B 3 and Abstract"},{"comment":"The term 'self-contained' is used to describe the RSRG-X formalism, but later the text admits that tracking three-site interactions remains open. Please clarify that 'self-contained' refers to the closed set of RG steps within the assumed dominance of one- and two-superspin couplings, not to the full validity of that assumption.","section":"Sec. V A 3"},{"comment":"The fits to the log^3 form and the power-law form are both shown on the same data; it would improve the presentation to state explicitly the fitted frequency ranges and the extracted parameters (A, ω0, c) for each case, so that the reader can assess the discriminative power of the fits.","section":"Sec. V B 3, Fig. 11b"},{"comment":"The manuscript has several sentences that appear incomplete or run-on (e.g., in the introduction of Sec. V A and in Appendix C 2 b). A careful proofreading pass would improve readability.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"This is a borderline case: the nearest-neighbor analytical derivation and the ED benchmarks for NNN chains are strong and could justify publication, but the abstract overstates the long-range result and the key RG premise for interacting chains is not fully established. The issues are fixable by changing the claims and adding a more careful analysis or an explicit caveat, so I do not recommend rejection. The paper also has a small but real typo in Eq. (D10) that should be corrected."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper before deciding anything. The superspin RSRG-X formalism is genuinely new: earlier RSRG-X treatments of long-range XX+YY chains explicitly dropped nontrivial superspins, and here the U(1) x Z2 symmetry structure keeps the effective Hamiltonian closed enough to track them. And the central one-dimensional claim, disorder-averaged spin survival decaying as 1/log^2(t), has real support: a clean analytic derivation for the nearest-neighbor model from the infinite-randomness fixed point, plus ED benchmarks at N=12-16 for NN, NNN, and long-range chains.\n\nCredit where due. The NN derivation in Sec. IV A is transparent: the pairing structure, the g(l) ~ 1/l^3 energy distribution, and the Fourier transform in Appendix A are all done carefully. The numerical RSRG-X reproduces ED quantitatively at low frequencies, including strong finite-size effects, which is the right validation target. The many-body spin-flip picture is clear, and the statistics in Fig. 9 showing abundant n>=2 collective flips makes the case that superspins matter. The RG premise is checked numerically (Appendix E, Table I) rather than assumed.\n\nSoft spots, in proportion. The abstract overstates the long-range chain. The body (Sec. V B 3) says plainly that for the LR model, finite-size effects prevent confirming the 1/log^2 asymptote in the thermodynamic limit. The N=40 fit to the universal form sits below the frequency window where ED validates the method, and the ED agreement itself degrades below omega ~ 10^-6 at N=16. Abstract says no significant deviation; body says inconclusive. Fix that mismatch. The 2D results are honestly labeled preliminary, and the Q<0 rate up to 1.1% in Table I is a real caveat, properly flagged. Minor items: no code or data released, figures lack error bars, and the conclusion cites Fig. 10 for the N=40 LR data when it is Fig. 11b. The three-superspin and X(0)X(0) couplings are argued negligible rather than proven, but the ED agreement is empirical evidence the approximation holds for these models in practice.\n\nWho it is for: people working on random spin chains, Rydberg and NV simulators, and RSRG methods generally. It deserves a serious referee. Send it to review; ask for the abstract-body alignment on the LR claim and for code and data release.","headline":"Genuinely new superspin RSRG-X formalism with a solid 1D derivation and honest ED benchmarks; the abstract overstates the long-range chain, which the body itself admits is unconfirmed in the thermodynamic limit.","tokens_in":57163,"tokens_out":5889,"would_cite":true,"duration_ms":53885,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"For random spin chains, spin survival decays as $B + A/\\log^2(t/t_0)$, slower than any power law, and a new superspin renormalization scheme reproduces exact diagonalization at low frequencies.","keywords":["RSRG-X","superspin","random spin chains","dipolar XX+YY model","spin survival probability","slower-than-power-law relaxation","infinite-randomness fixed point","many-body resonances"],"falsifier":"Compute $\\omega\\overline{S_p}(\\omega)$ for a long random nearest-neighbor $XX+YY$ chain with couplings drawn from the fixed-point distribution and compare with the predicted curved form $A[\\log(\\omega_0/\\omega)]^{-3}$; if the low-frequency data follow a straight power-law line instead, the infinite-randomness description of the decay is refuted.","tokens_in":55964,"feed_emoji":"🧲","tokens_out":7530,"duration_ms":78234,"temperature":0.7,"pith_summary":"This paper develops a renormalization-group method, superspin RSRG-X, for random spin-1/2 systems whose Hamiltonians conserve total $Z$ spin and are invariant under global spin flip. Using it, the authors argue that in one-dimensional random $XX+YY$ chains, including dipolar chains relevant to Rydberg and nitrogen-vacancy experiments, the disorder-averaged local spin survival probability decays at late times as $B + A/\\log^2(t/t_0)$, slower than any power law. The method resolves late-time dynamics into coherent flips of superspins, collective clusters of aligned spins, and provides a numerical algorithm that quantitatively matches exact diagonalization at low frequencies for nearest-neighbor, next-nearest-neighbor, and long-range dipolar chains. Applied to two-dimensional systems, it indicates subdiffusive power-law decay rather than universal logarithmic decay. If correct, this gives a controlled, experimentally relevant picture of slow relaxation of conserved densities in disordered spin systems.","feed_headline":"Random spin chains relax slower than any power law","feed_subtitle":"A superspin renormalization method matches exact numerics at low frequencies and reaches sizes exact methods cannot.","key_machinery":"The central machinery is the superspin RSRG-X update rule: at each step the algorithm identifies the strongest one- or two-superspin coupling, branches into an eigenspace of that coupling, and generates a new effective Hamiltonian on two-level superspins, where a superspin-$m$ is a collective degree of freedom whose two states differ in total magnetization by $2m$. The global U(1) and $\\mathbb{Z}_2$ symmetries restrict the allowed couplings to $XX+YY$, $ZZ$, and a dangling $X^{(0)}$ term, making the renormalization scheme self-contained. Each resonant branching creates a coherent flip between $|\\uparrow^{\\otimes I}\\downarrow^{\\otimes J}\\rangle$ and $|\\downarrow^{\\otimes I}\\uparrow^{\\otimes J}\\rangle$ at frequency $4\\Omega_{n,j}$, and the low-frequency spectral function is the sum over these resonances.","core_discovery":"The central discovery is that for one-dimensional random $XX+YY$ chains with U(1) and $\\mathbb{Z}_2$ symmetry, the disorder-averaged infinite-temperature spin survival probability decays as $\\overline{S_p}(t) \\simeq B + A/\\log^2(t/t_0)$ at late times, which is slower than any power law. The paper establishes this by constructing an RSRG-X formalism in which the effective Hamiltonian is expressed in terms of two-level superspins, so that conserved spin density relaxes through coherent collective spin flips. The numerical implementation matches exact diagonalization at low but nonzero frequencies for nearest-neighbor, next-nearest-neighbor, and long-range dipolar chains, and it extends the reachable system sizes to at least $N=40$ in one dimension. For two-dimensional power-law interacting models, the results indicate subdiffusive spin relaxation with a decay exponent that decreases as the interaction becomes more long-ranged.","pith_inferences":["A direct experimental test would be to prepare a randomly filled Rydberg chain with dipolar $XX+YY$ interactions and measure the conserved magnetization autocorrelation; if the disorder is strong, the predicted $1/\\log^2(t)$ tail should be visible over many orders of magnitude in time.","The two-dimensional results suggest a dimensional crossover: at small interaction exponent $\\alpha$ the late-time decay may be a genuine power law with continuously varying exponent rather than the logarithmic asymptote of one dimension, and a finite-size scaling study at fixed $\\alpha$ could separate the two possibilities.","Because the Monte Carlo branching samples RSRG-X states uniformly, observables dominated by rare strongly resonant eigenstates may require importance sampling; comparing the sampled resonance-energy distribution with exact spectra would test how far the method's quantitative reach extends.","The picture of independent coherent resonances implies that spin transport in these disordered chains is strongly suppressed and possibly frequency-dependent, which could connect the resonance distribution to rigorous bounds on sub-ballistic or subdiffusive transport in disordered spin systems."],"forward_implications":["For nearest-neighbor random $XX+YY$ chains, the disorder-averaged spin survival probability has the universal asymptote $\\overline{S_p}(t) = B + A/\\log^2(t/t_0)$ in the thermodynamic limit, with no power-law tail.","The same $1/\\log^2(t)$ form is consistent with next-nearest-neighbor and long-range dipolar chains at the largest sizes accessible to the method, indicating that logarithmic relaxation is not an artifact of the free-fermion mapping.","The numerical RSRG-X approach reaches system sizes beyond exact diagonalization while retaining quantitative agreement at low frequencies, allowing tests of late-time asymptotics that ED cannot access.","In two dimensions, the method predicts subdiffusive power-law decay of the spin survival probability, with the exponent decreasing as the interaction becomes more long-ranged.","Many-body spin flips involving $n \\ge 2$ aligned spins are abundant in interacting chains, so their dynamics differs qualitatively from the two-body spin-flip physics of nearest-neighbor chains.","The formalism also applies to Hamiltonians with $ZZ$ interactions, since those couplings are generated and renormalized within the same closed set of U(1)- and $\\mathbb{Z}_2$-symmetric terms."],"supporting_citations":[{"why":"Establishes the base RSRG-X scheme for random nearest-neighbor $XX+YY$ chains that this paper extends to interacting and long-range models.","marker":"[35]"},{"why":"Supplies the infinite-randomness fixed-point distribution of logarithmic interaction strengths used to derive the $1/\\log^2(t)$ decay form.","marker":"[21]"},{"why":"Earlier low-temperature dynamics result for the same fixed point; the present paper generalizes its $1/\\log^2$ behavior to infinite temperature and identifies the role of superspins.","marker":"[26]"},{"why":"Prior RSRG-X study of non-equilibrium dynamics in U(1)- and $\\mathbb{Z}_2$-symmetric chains whose initial states suppressed superspin contributions; the present formalism removes that restriction.","marker":"[32]"},{"why":"Provides the strong-randomness RSRG framework for two-dimensional systems and the background expectation that ground states there are not governed by infinite-randomness fixed points.","marker":"[19]"},{"why":"Exact numerical checks of the low-temperature $1/\\log^2(t)$ asymptote, which the paper's infinite-temperature RSRG-X prediction parallels.","marker":"[51]"}],"fun_headline_variants":["Superspin method unveils logarithmic relaxation in spin chains","Random spin chains defy power-law decay","New renormalization finds slower-than-power-law decay","Superspins explain slow relaxation in random magnets","Spin survival in random chains beats power laws"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument stands or falls on the strong-randomness premise that, at every renormalization step, the selected dominant one- or two-superspin coupling overwhelms every overlapping coupling and that three-superspin and superspin-0 pair terms remain negligible; the paper reports this premise is satisfied with high but not perfect probability in one dimension and significantly less often in two dimensions, so if those overlaps become strong the effective-Hamiltonian update breaks down.","fun_headline_variants_meta":{"raw":{"variants":["Superspin method unveils logarithmic relaxation in spin chains","Random spin chains defy power-law decay","New renormalization finds slower-than-power-law decay","Superspins explain slow relaxation in random magnets","Spin survival in random chains beats power laws"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000842,"raw_usage":{"total_tokens":3732,"prompt_tokens":1076,"completion_tokens":2656,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":692,"completion_tokens_details":{"reasoning_tokens":2584}},"tokens_in":692,"tokens_out":2656,"duration_ms":19972,"temperature":1.0,"reasoning_tokens":2584,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T20:51:12.790595+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $\\omega\\overline{S_p}(\\omega)$ for a long random nearest-neighbor $XX+YY$ chain with couplings drawn from the fixed-point distribution and compare with the predicted curved form $A[\\log(\\omega_0/\\omega)]^{-3}$; if the low-frequency data follow a straight power-law line instead, the infinite-randomness description of the decay is refuted.","supporting_citations":[{"cited_title":"Motrunich, K","cited_arxiv_id":null,"evidence_quote":"Establishes the base RSRG-X scheme for random nearest-neighbor $XX+YY$ chains that this paper extends to interacting and long-range models."},{"cited_title":"Barredo, H","cited_arxiv_id":null,"evidence_quote":"Supplies the infinite-randomness fixed-point distribution of logarithmic interaction strengths used to derive the $1/\\log^2(t)$ decay form."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the strong-randomness RSRG framework for two-dimensional systems and the background expectation that ground states there are not governed by infinite-randomness fixed points."},{"cited_title":"Braemer, T","cited_arxiv_id":null,"evidence_quote":"Exact numerical checks of the low-temperature $1/\\log^2(t)$ asymptote, which the paper's infinite-temperature RSRG-X prediction parallels."}],"review_version":1}