{"id":"5093f27a-d856-4ec6-bbb1-1e2736072aff","arxiv_id":"2605.24857","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":4.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":5,"one_line_summary":"Bayesian optimization of magnetization data for α-RuCl₃ yields a parameter set with large positive Γ that agrees with other observables, though the Kitaev coupling K scale remains undetermined by the magnetization curves alone.","lead":"This paper applies Bayesian optimization to fit five spin Hamiltonian parameters for the Kitaev candidate α-RuCl₃ directly from experimental magnetization curves along the b- and c-axes. A smart generalist might read it to see a data-driven alternative when ab initio calculations give uncertain or method-dependent results for quantum magnets.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"Cost function insensitivity to |K| (and weak dependence on Γ', J) prevents unique determination of the minimizing parameter set","rationale":"The reader's weakest assumption (completeness of the 5-term model and solver accuracy) is relevant but secondary; the abstract itself flags the non-uniqueness with respect to K as a direct limitation on what the magnetization fit can determine. This is an internal consistency issue for the central claim rather than an external-consensus disagreement.","tokens_in":1888,"tokens_out":349,"duration_ms":25281,"concrete_test":"Re-optimize the four parameters (Γ, Γ', J, g_c) for several fixed K values in [-3, -12] meV; if multiple K values achieve cost within 5% of the reported minimum but produce specific-heat peak positions differing by >20%, the uniqueness of the large-Γ scenario is not established by the magnetization data alone.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The abstract explicitly states that the cost function is insensitive to the absolute value of K, so magnetization data alone do not determine its energy scale, and depends only weakly on Γ' and J. The reported set fixes K = -6.0 meV arbitrarily while favoring large positive Γ. When the same set is used to compute specific heat, susceptibility, and structure factor (quantities sensitive to overall energy scale and relative couplings), any other K value that keeps the magnetization cost comparable could produce different thermodynamic signatures. This makes it unclear whether the reported set is the only one that 'yields good agreement' or whether the large-Γ preference is robust once the energy scale is allowed to vary.","agreement_with_reader":"disagree"},"referee_report":{"model":"grok-4.3","summary":"The paper proposes using Bayesian optimization to fit five parameters of the spin Hamiltonian for α-RuCl₃ (K, Γ, Γ', J, g_c) directly to experimental magnetization curves along the b- and c-axes. The minimizing set is reported as (K, Γ, Γ', J, g_c) = (-6.0, 7.5, -0.3, -1.75, 2.3) meV; the cost function is insensitive to |K| and depends only weakly on Γ' and J while favoring large positive Γ. The authors then evaluate the static spin structure factor, magnetic susceptibility, and specific heat for this set, claiming it yields good agreement with experiment and supports the large-Γ scenario over the small-g_c alternative.","tokens_in":2051,"tokens_out":483,"duration_ms":27245,"significance":"A reliable data-driven route to spin-Hamiltonian parameters would usefully complement ab-initio methods whose results can vary with methodological details. The explicit use of Bayesian optimization together with low-energy solvers, the transparent reporting of cost-function insensitivity, and the cross-validation on three additional observables constitute concrete strengths that, if the numerical procedures are shown to be robust, would make the methodology transferable to other quantum magnets.","major_comments":[{"comment":"Abstract: the cost function is stated to be insensitive to the absolute value of K, yet the reported minimizing set fixes K = -6.0 meV. Because specific heat, susceptibility and the structure factor depend on the overall energy scale, it is necessary to show that the reported agreement with experiment and the preference for large Γ remain stable when K is varied over the range that keeps the magnetization cost function comparably low.","section":"Abstract"},{"comment":"Abstract (and the description of the optimization procedure): the central claim that the fitted parameters yield good agreement with experiment rests on the premise that the low-energy numerical solvers inside the cost function accurately reproduce the b- and c-axis magnetization for arbitrary parameter values. No convergence tests, error estimates, or cross-checks against exact diagonalization on small clusters are referenced, leaving the reliability of the entire optimization chain unquantified.","section":"Abstract"}],"minor_comments":[],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the careful reading of the manuscript and the recommendation for minor revision. The two major comments identify important aspects of robustness that we address point by point below. Both can be resolved with targeted additions to the manuscript.","responses":[{"response":"We agree that, although the manuscript already states the insensitivity of the magnetization cost function to |K|, it is necessary to demonstrate that the conclusions drawn from the other observables remain stable. In the revised manuscript we will add calculations for several values of K lying within the low-cost region of parameter space and show that both the preference for large positive Γ and the level of agreement with the static structure factor, susceptibility, and specific heat are preserved. This will make explicit that the support for the large-Γ scenario is independent of the precise overall energy scale chosen for K.","revision_made":"yes","referee_comment":"[Abstract] Abstract: the cost function is stated to be insensitive to the absolute value of K, yet the reported minimizing set fixes K = -6.0 meV. Because specific heat, susceptibility and the structure factor depend on the overall energy scale, it is necessary to show that the reported agreement with experiment and the preference for large Γ remain stable when K is varied over the range that keeps the magnetization cost function comparably low."},{"response":"We acknowledge that explicit convergence tests, error estimates, and cross-checks against exact diagonalization were not included in the original manuscript. In the revised version we will add a dedicated subsection (or appendix) that reports convergence tests of the low-energy solvers with respect to bond dimension, truncation error, and cluster size, together with direct comparisons to exact diagonalization on small clusters for representative parameter points. Quantitative error estimates on the magnetization curves will also be provided to quantify the reliability of the cost function.","revision_made":"yes","referee_comment":"[Abstract] Abstract (and the description of the optimization procedure): the central claim that the fitted parameters yield good agreement with experiment rests on the premise that the low-energy numerical solvers inside the cost function accurately reproduce the b- and c-axis magnetization for arbitrary parameter values. No convergence tests, error estimates, or cross-checks against exact diagonalization on small clusters are referenced, leaving the reliability of the entire optimization chain unquantified."}],"tokens_in":1616,"tokens_out":499,"duration_ms":26044,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main thing to know is that the cost function built from b- and c-axis magnetization is insensitive to the absolute value of K, so the authors simply set K = -6.0 meV and report the other four parameters that minimize it. They also note weak dependence on Γ' and J. That limitation is stated plainly.\n\nWhat the work actually does is apply Bayesian optimization to five parameters against experimental magnetization curves, then use the resulting set to compute the static spin structure factor, susceptibility, and specific heat. They argue these favor the large-Γ choice over a small-gc alternative and that the set matches experiment reasonably. The transparency about what the magnetization data can and cannot constrain is the useful part.\n\nThe soft spot is the one flagged in the stress test. Because the magnetization cost does not pin K, other values of K that keep the cost comparable could shift the energy scale and change the thermodynamic signatures. The paper claims the chosen set works well on those quantities, but without a check of how sensitive they are to K it is not clear whether the large-Γ preference is stable. Everything also depends on the low-energy solvers inside the optimization loop being reliable across the sampled parameter space.\n\nThis is a specialized methods paper for people working on Kitaev materials and parameter extraction. A reader who needs to know the current experimental constraints on α-RuCl3 would get value from the reported set and the explicit warning about K. It is worth sending to referees because the limitation is documented and the approach is practical, even if the Hamiltonian is not fully determined.","headline":"Magnetization data leaves K undetermined so the authors fix it by hand at -6 meV while the fit favors large positive Γ; the paper is honest about this limit but the thermodynamic checks may not be robust to it.","tokens_in":2533,"tokens_out":408,"would_cite":false,"duration_ms":23543,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Bayesian optimization of magnetization curves selects a spin Hamiltonian for α-RuCl₃ that favors large positive Γ over the small-g_c alternative.","keywords":["α-RuCl3","Kitaev material","spin Hamiltonian","Bayesian optimization","magnetization curves","Γ interaction","quantum magnets"],"falsifier":"A direct measurement of specific heat or magnetic susceptibility at temperatures and fields where the optimized parameters predict a clear mismatch with existing data.","tokens_in":2796,"feed_emoji":"📊","tokens_out":733,"duration_ms":18025,"temperature":0.7,"pith_summary":"The paper fits five parameters of the spin Hamiltonian for the Kitaev candidate α-RuCl₃ to experimental magnetization data along the b and c axes using Bayesian optimization. The cost function is minimized at (K, Γ, Γ', J, g_c) = (-6.0, 7.5, -0.3, -1.75, 2.3) meV, yet remains insensitive to the absolute scale of K and only weakly dependent on Γ' and J. Computations of the static spin structure factor, magnetic susceptibility, and specific heat with this parameter set agree with experiment and prefer the large-Γ regime over a competing small-g_c regime. The method combines data-driven fitting with low-energy numerical solvers to extract Hamiltonian parameters directly from measurements.","feed_headline":"Bayesian fit to magnetization selects large-Γ Hamiltonian for α-RuCl3","feed_subtitle":"Optimization of b- and c-axis data yields parameters whose computed susceptibility and specific heat match experiment better than the small-","key_machinery":"Bayesian optimization of a cost function built from experimental magnetization curves, evaluated with low-energy numerical solvers for each candidate parameter set.","core_discovery":"The parameter set that minimizes the cost function defined on b- and c-axis magnetization curves is (K,Γ,Γ',J,g_c)=(-6.0, 7.5, -0.3, -1.75, 2.3) meV. Magnetization data alone do not fix the energy scale of the Kitaev interaction K. The optimization favors a large positive Γ. When the static spin structure factor, magnetic susceptibility, and specific heat are computed from this set, they match experiment and support the large-Γ scenario over the small-g_c scenario.","pith_inferences":["Additional observables such as neutron scattering intensities would be needed to constrain the Kitaev coupling scale that magnetization leaves free.","The same optimization workflow could be applied to other quantum magnets once magnetization curves along principal axes are available.","If the five-term form is incomplete, the apparent preference for large Γ could shift when further interaction terms are included."],"forward_implications":["The optimized parameters reproduce the measured b- and c-axis magnetization curves.","Static spin structure factor, susceptibility, and specific heat computed from the same set agree with experiment.","These thermodynamic and structural quantities distinguish the large-Γ regime from the small-g_c regime.","Magnetization data by themselves leave the absolute Kitaev scale undetermined.","The combination of Bayesian optimization and accurate solvers offers a systematic route to Hamiltonian parameters from experimental data."],"fun_headline_variants":["Bayesian magnetization fit favors large Γ in α-RuCl3","Large Γ selected via Bayesian optimization of α-RuCl3 curves","Bayesian optimization of magnetization selects large Γ for α-RuCl3","Large positive Γ favored by Bayesian magnetization optimization in RuCl3","α-RuCl3 b and c axis data Bayesian optimized to large Γ"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The spin Hamiltonian is completely captured by the five fitted terms and the numerical solvers accurately reproduce the magnetization curves for any values of those terms.","fun_headline_variants_meta":{"raw":{"variants":["Bayesian magnetization fit favors large Γ in α-RuCl3","Large Γ selected via Bayesian optimization of α-RuCl3 curves","Bayesian optimization of magnetization selects large Γ for α-RuCl3","Large positive Γ favored by Bayesian magnetization optimization in RuCl3","α-RuCl3 b and c axis data Bayesian optimized to large Γ"]},"model":"grok-4.3","cost_usd":0.012696,"raw_usage":{"total_tokens":5610,"prompt_tokens":848,"num_sources_used":0,"completion_tokens":91,"cost_in_usd_ticks":126962000,"prompt_tokens_details":{"text_tokens":848,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":4671,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":848,"tokens_out":91,"duration_ms":46545,"temperature":1.0,"reasoning_tokens":4671,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-30T00:09:39.554001+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A direct measurement of specific heat or magnetic susceptibility at temperatures and fields where the optimized parameters predict a clear mismatch with existing data.","supporting_citations":[],"review_version":1}