{"id":"aa4627aa-564b-4507-b1e3-78fb97800a00","arxiv_id":"2601.07361","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Interacting dark matter–dark energy models fit DESI BAO and CMB data as well as evolving-dark-energy (CPL) models, with a coupling preference that persists under DES-Dovekie supernova recalibration.","lead":"DESI's expansion-rate anomalies, often blamed on dark energy changing over time, can also be mimicked by dark matter and dark energy exchanging energy. The paper fits two such interacting models to CMB, BAO, and supernova data and finds a 3–5σ preference for interaction that survives a supernova recalibration aimed at weakening dynamical dark energy.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"CF model's 'robust' evidence depends on unvalidated ePPF perturbations in an extreme parameter region; a full-solver check or inclusion of WL/LSS data could collapse the both-scenarios claim.","rationale":"The paper is a serious Bayesian analysis with careful treatment of CMB nuisance, H0 sampling, and profile likelihoods. The CQ evidence (β>0 at ~5σ) is credible, and the comparison with CPL is informative. However, the headline claim treats CQ and CF symmetrically ('both IDE scenarios'), and the CF result is the load-bearing part: Table IV shows CF reaches Δχ²≈−15 (Dovekie) only at Ωm≈0.62 and σ8≈0.48. The ePPF formalism is an approximation; the original validation (Li et al. 2014) was for moderate couplings, and the authors provide no test in this extreme phantom regime. The supplementary note that 'a comprehensive assessment of the viability of this scenario in light of WL and LSS measurements is beyond the scope' is an explicit admission that the most discriminating datasets are absent. My concern is not that the CF model is 'outside consensus'—it is that the numerical perturbation prediction underlying the CF likelihood may be inaccurate, and that the resulting low growth prediction is inconsistent with existing data once included. Neither issue is resolved in the manuscript. The reader's CONDITIONAL verdict is appropriate; I would not change it.","tokens_in":24623,"tokens_out":8335,"duration_ms":89622,"concrete_test":"Validate the CF perturbation pipeline: using the CF best-fit parameters (e.g., CMB+DESI+Dovekie: w=-1.52, β=-2.59), compute CMB TT/TE/EE, lensing, and matter power spectra with an independent implementation of the same coupled-fluid model—e.g., a modified CLASS with explicit energy-momentum transfer terms from Eqs. (7)-(10)—and compare to IDECAMB's ePPF output. If the ΔCℓ/Cℓ exceed ~1% in the Planck/ACT/SPT multipole range (ℓ>30) or ΔP(k)/P(k) exceed ~5% for k<1 h/Mpc, the ePPF approximation is not reliable in the CF best-fit region, and the claimed 3–5σ interaction preference is not robust. Additionally, run a further MCMC including DES Y3 cosmic shear and galaxy clustering; if the CF posterior shifts or is excluded at >3σ, the 'both scenarios' claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract's central claim—'both IDE scenarios show robust evidence for non-vanishing interactions at the 3–5σ level'—depends critically on the CF model, whose good fit is obtained only in an extreme region of parameter space (w≈−1.5, Ωm≈0.6, σ8≈0.48; Table IV). This region is exactly where the ePPF perturbation framework (refs [115,118]) is least tested: the model has a constant phantom EoS w<−1 and a strong coupling β≈−2.5, far from the ΛCDM-like regimes where ePPF is calibrated. The paper asserts ePPF 'effectively eliminates the instability while preserving accurate small-scale dynamics' (Supplementary Sec. I) but provides no convergence or cross-code validation for these parameters. If ePPF overestimates the suppression of structure (σ8≈0.48, S8≈0.68), the CF model's CMB fit—and hence its 3–5σ interaction preference—could be an artifact. Moreover, the same low S8 value is in direct tension with weak-lensing and galaxy-clustering measurements that are explicitly excluded (Supplementary Sec. III). This is not a mere refinement: the entire 'both scenarios' conclusion collapses if CF's perturbation treatment or its growth predictions do not survive contact with the omitted data.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses CMB (Planck+ACT+SPT), DESI DR2 BAO, and three SN compilations (PantheonPlus, DESY5, DES-Dovekie) to constrain two interacting dark-energy models: coupled quintessence (CQ) and a phenomenological coupled fluid (CF). The authors report a robust preference for non-vanishing dark-sector interactions at the 3–5σ level in both models, with fit quality comparable to or better than the CPL dynamical-dark-energy parametrization for the same number of free parameters. They use profile-likelihood checks to argue that the interaction signal is likelihood-driven, and they emphasize that the preference survives the DES-Dovekie recalibration that weakens the CPL evidence. The CF model achieves its best fits only in an extreme parameter region (Ωm≈0.6, σ8≈0.48, w≈−1.5), a point the authors acknowledge.","tokens_in":25003,"tokens_out":4674,"duration_ms":51008,"significance":"The CQ analysis is a potentially significant contribution: it shows that a Lagrangian-based interacting dark-energy model with the same number of parameters as CPL can reproduce the DESI-driven preference for deviations from ΛCDM, with a profile-likelihood minimum away from β=0 and consistent results across several SN compilations. The paper also includes careful methodological choices, such as sampling H0 and computing θ* numerically rather than using the Hu–Sugiyama approximation, and it makes model-comparison tables transparent. However, the headline claim that 'both IDE scenarios' show robust evidence depends critically on the CF model, whose extreme parameter region is supported by an ePPF perturbation treatment that is not cross-validated and whose low S8 prediction is in tension with existing weak-lensing and galaxy-clustering data. As presented, the CF result is not yet on the same evidential footing as the CQ result.","major_comments":[{"comment":"The ePPF framework is asserted to 'effectively eliminate the instability while preserving accurate small-scale dynamics' (Supplementary Sec. I), but no validation is shown for the CF best-fit region: Table IV gives w≈−1.4 to −1.6, β≈−2.3 to −3.1, Ωm≈0.6, σ8≈0.48. This is far from the ΛCDM-like regime in which such approximate perturbation schemes are normally calibrated. Because CF provides the largest Δχ² and ΔDIC improvements in Table II, the central claim that 'both' IDE models robustly beat ΛCDM rests on the reliability of ePPF in this untested phantom/strong-coupling regime. A cross-check against a full perturbation solver, or at least a demonstration that the conclusions are insensitive to the ePPF closure, is needed.","section":"Supplementary Sec. I; Table IV"},{"comment":"The paper explicitly states that a comprehensive assessment against weak-lensing and LSS measurements is beyond its scope, but the CF model predicts S8≈0.68 (σ8≈0.48) with Ωm≈0.6 (Table IV), values that are in strong tension with current galaxy-lensing and clustering data. Since CF is the model that gives the best Δχ² and ΔDIC in Table II, the broad claim that interacting dark energy is statistically competitive with CPL and “robustly” preferred over ΛCDM is conditional on a scenario whose growth predictions appear to conflict with existing probes. Please include representative WL/LSS likelihoods (e.g., DES Y3, KiDS, or an equivalent) or explicitly restrict the robustness claim to the CQ model.","section":"Supplementary Sec. III; Table II"},{"comment":"The lower panel of Fig. 2 plots Δχ² only over β∈[−4,−1.5]. This range excludes β=0, the non-interacting ΛCDM limit. The accompanying text states that the CF profile likelihood is 'significantly displaced from the non-interacting limit,' but the plotted curve does not show Δχ² at β=0 and therefore does not directly support a significance claim for CF. Please extend the x-axis to β=0 or explicitly report Δχ²(β=0) for the CF model. Without this, the profile-likelihood argument against prior-volume effects is incomplete for the CF scenario.","section":"Fig. 2 (lower panel); text near profile-likelihood discussion"}],"minor_comments":[{"comment":"The conclusion says the marginalized constraints and profile likelihood reach 'the 5σ level.' Table I shows that 5σ is reached for CQ, but for CF the significance varies by dataset (e.g., β=−2.37±0.82 for CMB+DESI is roughly 3σ). Please be precise about which model and dataset support which significance.","section":"Conclusion"},{"comment":"The caption states the CQ panel is on the right and the CF panel on the left, but the text in Supplementary Sec. III refers to 'the left panel of Fig. 7' in the CQ discussion. Please correct this cross-reference or the panel ordering.","section":"Fig. 7 caption and Supplementary Sec. III text"},{"comment":"The binned distance-modulus residuals would be easier to interpret if the error bars on the bins were shown explicitly, or if the binning prescription were stated in the caption; currently the visual weight per point is unclear.","section":"Fig. 5"},{"comment":"The phrase 'robust evidence for non-vanishing interactions at the 3–5σ level' is used before the CF caveats are introduced. Consider stating in the abstract that the CF scenario lives in an extreme corner of parameter space and that its viability depends on perturbation physics and structure-growth data.","section":"Abstract and Sec. 1"}],"recommendation":"major_revision","confidential_remarks":"The paper relies heavily on the authors' own IDECAMB/ePPF codes, which are established in the literature but are not independently validated for the extreme CF region. The CQ result is likely the more durable contribution; the CF result currently carries the 'both scenarios' headline and should be either substantiated with additional data/validation or separated from the main claim. The editor may wish to encourage the authors to make the CF dependence on ePPF and the excluded LSS/WL data a central rather than peripheral issue in the revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know. First, the CQ model—Lagrangian coupled quintessence with an inverse power-law potential—is a genuinely competitive alternative to CPL: with the same number of free parameters it matches CPL's fit to CMB+DESI+SN and it holds up under the DES-Dovekie recalibration that weakens CPL. That is new and worth engaging with. Second, the CF model is the soft half of the paper. Its best fit sits at Ωm≈0.6, σ8≈0.48, S8≈0.68, a corner that existing WL/LSS data would likely disfavor, and the ePPF perturbation treatment is least validated in exactly that corner. The paper itself admits that a comprehensive WL/LSS assessment is out of scope. So the abstract's 'both IDE scenarios show robust evidence' overstates what is actually shown.\n\nThe analysis is competent. They sample H0 directly instead of using the Hu-Sugiyama θ* fit, which is the right call for IDE. They check profile likelihoods to guard against prior-volume effects. They use multiple SN compilations, and the CQ preference is stable across them. The unified perturbation mapping is a useful organizational device, and IDECAMB/ePPF are their own established tools rather than black boxes.\n\nSoft spots, in proportion. The CQ prior β∈[0,0.15] enforces the sign of the coupling; the claim of a positive coupling is conditional on not allowing negative couplings. For CF, the conclusion's 'reaching the 5σ level' is not supported by Table I: β≈−2.4 with errors around 0.8 is more like 3σ. The profile-likelihood plot for CF shows only Δχ² near the minimum; there is no Δχ² at β=0, so the 5σ claim is not demonstrated. The CF perturbation predictions need a cross-code check or inclusion of WL/LSS data before the both-scenarios claim can be taken at face value.\n\nBottom line: CQ is a credible result that deserves referee time; CF needs either validation or a downgraded summary. Send it to peer review, but ask for the CF perturbation validation and a more honest abstract. For a reading group, I'd bring it if you want to debate whether the CF corner is an artifact.","headline":"CQ is a credible CPL rival; the CF half of the 'robust both-scenarios' claim sits in a corner that structure-growth data would likely reject.","tokens_in":25480,"tokens_out":2928,"would_cite":true,"duration_ms":32051,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["95.35.+d","95.36.+x","98.80.-k"],"model":"deepseek-v4-flash","headline":"DESI's deviations from standard cosmology can be explained by dark matter–dark energy interactions at 3–5 sigma, matching or beating evolving-dark-energy fits without requiring a time-varying equation of state.","keywords":["interacting dark energy","dark matter-dark energy coupling","DESI BAO","dynamical dark energy","CPL parametrization","coupled quintessence","coupled fluid","cosmological perturbations"],"falsifier":"Include weak-lensing and galaxy-clustering measurements in the same MCMC analysis. The coupled-fluid model predicts S8≈0.68, while current lensing data cluster near S8≈0.76–0.82; if adding those data pulls the coupled-fluid posterior back toward ΛCDM, or worsens its Δχ² enough to remove the preference, then the paper's claim that both interacting models robustly beat ΛCDM at 3–5 sigma is falsified.","tokens_in":24500,"feed_emoji":"🌌","tokens_out":4681,"duration_ms":47489,"temperature":0.7,"pith_summary":"DESI baryon acoustic oscillation measurements deviate from the standard ΛCDM cosmology, and the usual interpretation is that dark energy's equation of state changes over time. This paper tries to establish a different physical explanation: dark matter and dark energy exchange energy and momentum, and that interaction reshapes the expansion history in a way that mimics dynamical dark energy. Using a unified treatment of perturbations for two interacting models (coupled quintessence and coupled fluid), the authors find a 3–5 sigma preference for a nonzero coupling across all data combinations. They claim these models fit the CMB, BAO, and supernova data as well as, or better than, the standard CPL parametrization with the same number of free parameters, even after a supernova recalibration that weakens the evidence for dynamical dark energy. If correct, the DESI results do not require dark energy to evolve; a dark-sector interaction is a statistically competitive physical alternative.","feed_headline":"Interacting dark sector rivals evolving dark energy at 3–5 sigma","feed_subtitle":"Two interacting models match or beat the standard CPL parametrization with the same number of free parameters, even after SN recalibration.","key_machinery":"The load-bearing tool is a unified parameterization of IDE perturbations, in which the energy and momentum transfer rates are written as linear combinations of the density and velocity perturbations of the two dark components, with model-specific mapping coefficients. This lets both the coupled-quintessence and coupled-fluid scenarios be evolved with the same treatment. For the coupled fluid, the extended parameterized post-Friedmann (ePPF) framework replaces the ill-defined large-scale pressure condition with a parametrized momentum relation, preventing the instabilities that normally plague constant-equation-of-state interacting fluids. The same number of free parameters as the CPL paramet","core_discovery":"On the paper's own terms, the central discovery is that the apparent DESI preference for dynamical dark energy is not unique. Two interacting dark-energy (IDE) realizations—coupled quintessence, defined at the Lagrangian level, and a phenomenological coupled fluid with a constant equation of state—both produce a background expansion that departs from ΛCDM in the redshift range probed by BAO and supernovae while fitting the data at least as well as CPL. The authors report that the coupling parameter is constrained away from zero at roughly 3–5 sigma, that the coupled fluid model gives the largest improvement over ΛCDM in both Δχ² and DIC, and that the preference survives the DES-Dovekie recal","pith_inferences":["I infer that the paper's two-model claim is not symmetric: the coupled-fluid model lives in a corner of parameter space (w≈-1.5, Ωm≈0.6, S8≈0.68) that existing weak-lensing measurements would likely penalize, so its competitive fit may degrade once those data are included.","I infer the strongest near-term test is not better BAO or supernova distances but structure growth: measuring S8 and the growth rate at z<1 should separate the coupled-fluid model from ΛCDM and from coupled quintessence.","I draw a non-obvious consequence: if future data favor an interacting scenario, coupled quintessence offers a less radical cosmology (H0 and Ωm close to ΛCDM) than the coupled fluid, so the two IDE realizations have very different implications for the Hubble tension and matter clustering.","I infer from the parameter correlations that the model rankings are partly driven by the lowest-redshift supernova bin; a systematic shift in that anchor would likely reorder CPL, CQ, and CF."],"forward_implications":["If the paper is right, the DESI BAO deviations can be explained without introducing a time-dependent dark-energy equation of state; an interaction between dark matter and dark energy is a viable alternative.","The coupled-quintessence model achieves its fit while remaining quintessence-like at all redshifts, so a phantom-crossing equation of state is not required by the data.","The coupled-fluid model predicts a distinctive structure-formation signature: low σ8 and S8 (about 0.48 and 0.68) with high matter density, meaning future weak-lensing and galaxy-clustering data can directly discriminate it from ΛCDM and CPL.","Even after the supernova recalibration that weakens dynamical-dark-energy evidence, the interaction preference persists, so the result is not an artifact of a single supernova compilation.","Because the background distances are highly degenerate between IDE and CPL, distinguishing the paradigms requires perturbation-level observations rather than more distance measurements."],"fun_headline_variants":["Dark sector interactions rival evolving DE at 3–5σ","Coupled dark matter and dark energy fit DESI at 3–5σ","Interacting dark sector beats ΛCDM and matches CPL","Dark coupling explains DESI's dark energy hint at 3–5σ"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The claim depends on the perturbation framework being reliable for the coupled-fluid model in the exact region the data prefer (w≈-1.5, Ωm≈0.6, σ8≈0.48), while weak-lensing and galaxy-clustering data that would directly test that region are omitted from the analysis.","fun_headline_variants_meta":{"raw":{"variants":["Dark sector interactions rival evolving DE at 3–5σ","Coupled dark matter and dark energy fit DESI at 3–5σ","Interacting dark sector beats ΛCDM and matches CPL","Dark coupling explains DESI's dark energy hint at 3–5σ"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000975,"raw_usage":{"total_tokens":3986,"prompt_tokens":756,"completion_tokens":3230,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":500,"completion_tokens_details":{"reasoning_tokens":3152}},"tokens_in":500,"tokens_out":3230,"duration_ms":27107,"temperature":1.0,"reasoning_tokens":3152,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T11:06:28.188566+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Include weak-lensing and galaxy-clustering measurements in the same MCMC analysis. The coupled-fluid model predicts S8≈0.68, while current lensing data cluster near S8≈0.76–0.82; if adding those data pulls the coupled-fluid posterior back toward ΛCDM, or worsens its Δχ² enough to remove the preference, then the paper's claim that both interacting models robustly beat ΛCDM at 3–5 sigma is falsified.","supporting_citations":[],"review_version":1}