{"id":"2a4911f7-7895-43e3-b4d7-a3795c7e188e","arxiv_id":"2508.15375","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Simulation-based study claims 15 dB channel gain from RIS-aided high-speed train MISO optimization, but the optimality and outage-elimination claims are analytically unsupported.","lead":"A proposed block coordinate descent algorithm tunes a reconfigurable intelligent surface's phases and a base station's beamforming to improve channel gain for high-speed train links, reporting a 15 dB gain over a no-RIS baseline. The paper's claims of optimal phase design and complete elimination of outage probability are not supported by the derivation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Outage-elimination claim is impossible under the paper's own non-central chi-square model: any finite variance yields positive outage below a positive threshold, so Fig. 3(b)'s zero is inconsistent with Eq. (24).","rationale":"The reader's REJECT verdict is correct. My load-bearing concern is the outage-elimination claim: the paper's own Eq. (24) is a non-central chi-square CDF, which is strictly less than 1 for every finite threshold b > 0, so Fig. 3(b)'s zero outage is mathematically impossible under the manuscript's model. This is decisive independently of any external consensus. The reader's weakest_assumption emphasized the phase update in Section III-A and the Gaussian conflation in Eqs. (22)–(23); I agree with the Gaussian-conflation concern and the overall rationale, but I would foreground the impossibility of zero outage because it directly falsifies a headline contribution. The phase-update suboptimality is also real: with κ = 3 the NLOS power is 0.25, so aligning only the LOS component and then applying a scalar rotation cannot be the global maximizer of P1 for fixed w; hence the repeated use of 'optimal' is unsupported. A corrected paper could rename the BCD update a heuristic, provide a rigorous finite-outage analysis, and benchmark against existing RIS-HST methods. As submitted, rejection is warranted, so the reader's verdict stands unchanged in substance.","tokens_in":8191,"tokens_out":11306,"duration_ms":128394,"concrete_test":"Re-run the outage evaluation used for Fig. 3(b) with a Monte Carlo of 10^7 independent channel realizations drawn from Eqs. (1) and (6) with Table I parameters (κ = 3, γth = 10 dB, NI = 1600), applying the proposed phase update and MRT. Count the fraction of realizations with |ĥ_k|² < σ²γth. Any positive count refutes 'eliminates outage'. Alternatively, plug the simulation's ζ_k and γ0,k into Eq. (24): with σ²_h,k > 0 and γ0,k > 0, the expression 1 − Q1(√ζ_k, √γ0,k) is strictly positive. If the plotted zeros are merely numerical underflow, the authors must replace the claim with a finite outage probability and revise the abstract/conclusion.","verdict_should_be":"REJECT","load_bearing_attack":"Section IV-C derives Pout,k = 1 − Q1(√ζ_k, √γ0,k) (Eq. 24) under the assumption ĥ_k ~ CN(μ_h,k, σ²_h,k). For any finite μ and σ² > 0, |ĥ_k|² has a non-central chi-square distribution with 2 degrees of freedom, whose CDF is strictly positive at any finite positive threshold. Q1(a,b) = 1 only when b = 0; here γ0,k = σ²γth / σ²_h,k > 0. Eq. (23) gives σ²_h,k = ̺²|h_d,k^H w_k|² + ̺⁴|h_k,(1)e^{jǫ*}|², which is positive for the stated κ = 3 (̺² = 1/4). Therefore the abstract, Fig. 3(b), and conclusion claim that 'the introduction of RIS eliminates outage probability' cannot be true under the paper's own distributional model. Moreover, the Gaussian model is itself unjustified: the rotation e^{jǫ*} in Eq. (17) depends on h_k,(1) and h_d,k^H w_k, so multiplying the NLOS Gaussian term by e^{jǫ*} destroys circular Gaussianity, and Eqs. (22)–(23) conflate deterministic and random parts by using |h_d,k^H w_k|² as a variance though h_d,k^H w_k contains a deterministic LOS component. This is not a minor caveat: the central performance claim is internally inconsistent.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a RIS-assisted high-mobility HST MISO downlink. It formulates a channel-gain maximization problem P1 and proposes a BCD algorithm that alternates between a RIS phase update (LOS Doppler cancellation plus a scalar phase rotation) and MRT transmit beamforming. It then presents analytical expressions for achievable rate, channel capacity, outage probability, and BER, and reports simulations claiming an average 15 dB channel gain improvement over baselines and that the RIS 'eliminates' outage probability.","tokens_in":8532,"tokens_out":5872,"duration_ms":63749,"significance":"The problem is timely and the system model is representative of current RIS-railway studies. If the optimization were actually optimal and the outage model were correct, the claimed 15 dB gain and outage elimination would be noteworthy. A positive feature is that the simulation setup is self-contained and uses explicit parameters; no machine-checked proofs or reproducible code are provided. However, the central technical claims are not supported: the proposed phase update is not the maximizer of the phase subproblem, and the outage analysis is internally inconsistent with the paper's own distributional model. The headline results therefore cannot be accepted as stated.","major_comments":[{"comment":"The proposed phase update does not solve the phase subproblem of P1. Equation (13) aligns only the LOS component of the cascaded channel; the NLOS term in Eq. (14), sqrt(1/(1+kappa)) h_NLOS,k Phi_k G w_k, is left uncontrolled and depends on the RIS phases through h_NLOS,k^H Phi_k G w_k. With kappa=3, the NLOS weight 1/(1+kappa)=0.25 is not negligible. The subsequent scalar rotation e^{j epsilon*} in Eq. (17) cannot compensate for per-element phase mismatches in the NLOS component. Thus Phi_k^* is not optimal, Algorithm 1's monotonicity/convergence is not established, and the labels 'optimal phase'/'optimal beamforming and phase' in Figs. 2-3 are unsupported. The 15 dB gain claim may be an artifact of this incomplete optimization rather than a true optimum.","section":"Section III-A, Eqs. (13)-(17)"},{"comment":"The outage-elimination claim is mathematically impossible under the paper's own model. With hat_h_k ~ CN(mu_h,k, sigma^2_h,k) and sigma^2_h,k > 0 (as given by Eq. (23) for kappa=3), |hat_h_k|^2 has a non-central chi-square distribution with 2 degrees of freedom, whose CDF is strictly positive at any finite positive threshold. Since gamma_{0,k} = sigma^2 gamma_th / sigma^2_h,k > 0 for gamma_th=10 dB, Eq. (24) gives P_out,k > 0. Figure 3(b) shows P_out=0 for all time slots, contradicting Eq. (24). The Abstract and Section V claim that the RIS 'eliminates outage probability' is therefore false under the paper's own distributional assumptions.","section":"Section IV-C, Eqs. (21)-(24)"},{"comment":"The Gaussian model for hat_h_k is not derived and is generally incorrect. The phase e^{j epsilon*} in Eq. (17) depends on the random realizations of h_{k,(1)} and h_d,k^H w_k; multiplying the Gaussian NLOS component by this data-dependent phase breaks circular symmetry, so hat_h_k is not CN(mu_h,k, sigma^2_h,k). Moreover, Eq. (22) treats h_d,k^H w_k as a mean, while Eq. (23) uses |h_d,k^H w_k|^2 as a variance, but h_d,k^H w_k contains a deterministic LOS part. The quantity in Eq. (23) is the squared magnitude of a deterministic-plus-random quantity, not the variance of the composite channel. Thus Eq. (24) does not follow, and the simulated outage curve in Fig. 3(b) is not a consequence of the stated formulas.","section":"Section IV-C, Eqs. (22)-(23)"}],"minor_comments":[{"comment":"Typographical errors should be corrected: 'Achivable Rate' in Section IV-A and Fig. 2(b); 'Outrage probability' in Fig. 3(b); 'chanel gain' in Eq. (9); 'dictance' in Section V; 'coverage probability' should read 'outage probability' in the paragraph after Eq. (23).","section":"Throughout"},{"comment":"The notation is inconsistent: v_k is introduced as the vector of RIS phases, but Eq. (13) suddenly uses v_{k,(1)} without formally defining it. Also, the relationship between diag(h_LOS^H) G w_k in Eq. (12) and diag(a_y(theta_2, phi_2) circle a_z(phi_2)) G w_k in Eq. (13) should be stated explicitly.","section":"Section III-A, Eqs. (11)-(13)"},{"comment":"The symbols rho and varrho are visually similar and used with different powers; using the explicit factors kappa/(1+kappa) and 1/(1+kappa) would improve clarity and avoid confusion in the variance formula.","section":"Section IV-C, Eqs. (22)-(23)"},{"comment":"The simulation section does not report Monte-Carlo repetitions, confidence intervals, or a comparison with an exhaustive/brute-force phase search for small N. Given that optimality is at issue, such a validation would be necessary to support the 15 dB gain claim.","section":"Section V"}],"recommendation":"reject","confidential_remarks":"The manuscript's central claims are internally inconsistent: the phase update is not the true maximizer of the phase subproblem, and the outage-elimination statement contradicts the paper's own non-central chi-square model. These are not presentation issues but load-bearing flaws. Even with extensive revision, the current contribution would need to be substantially reworked; I recommend rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Honestly, this one has a good setting but the main results don't survive contact with its own equations. The paper is a straightforward RIS phase-alignment plus MRT for high-speed train MISO, and the simulations show a 15 dB gain over their baselines, which is plausible. But the two headline claims—'optimal' phase design and 'eliminates outage'—are both overreach. The phase update in Section III-A aligns only the LOS component and then applies a common rotation; with κ=3 the NLOS component is not negligible, so this is a heuristic, not the maximizer of P1. The outage analysis is worse: Eq. (22)–(23) assert the composite channel after phase adjustment is Gaussian, but the phase in Eq. (17) is a function of the random channel, so the distribution isn't circular Gaussian. Even granting the Gaussian model, Eq. (24) with a finite non-centrality parameter gives positive outage for any positive threshold; Figure 3(b) shows exactly zero, which is internally inconsistent. That's not a minor caveat—'outage eliminated' is in the abstract and conclusion. The paper also doesn't benchmark against prior RIS-HST algorithms, only random phase and no RIS, so the novelty is incremental. On the plus side, the system model is clearly laid out, the BCD algorithm is easy to follow, and the simulation setup is transparent enough to reproduce. This could be a useful reference for RIS-HST modeling if the claims were trimmed. As it stands, the paper needs major revision: retract or soften the outage claim, rename the phase design a heuristic, and add comparisons to existing methods. I'd send it to peer review—the topic is active and the errors are instructive—but I'd expect rejection or major overhaul. Not something I'd cite in its current form.","headline":"The outage-elimination claim is contradicted by the paper's own non-central chi-square model, and the 'optimal' phase design is a heuristic; otherwise a routine RIS-HST extension.","tokens_in":9030,"tokens_out":3182,"would_cite":false,"duration_ms":34601,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Jointly optimizing RIS phases and transmit beamforming gives a high-speed train downlink 15 dB more channel gain and zero outage probability.","keywords":["reconfigurable intelligent surface","high-speed train communications","MISO beamforming","block coordinate descent","channel gain","outage probability","Doppler shift","Rician fading"],"falsifier":"Simulate the same high-speed MISO link at Rician factor 3 with 1600 elements, and for a fixed beamformer compare the paper's line-of-sight-only phase update against a phase vector found by coordinate ascent or a direct search over phases; a measurable gain increase would falsify the claimed optimality. Separately, compute the outage count from the empirical distribution of |h_hat_k|^2 instead of the Gaussian approximation in Eqs. (22)-(23) and check whether outage remains exactly zero across all time slots.","tokens_in":8052,"feed_emoji":"🚄","tokens_out":7037,"duration_ms":80123,"temperature":0.7,"pith_summary":"This paper tries to show that adding a reconfigurable intelligent surface to a high-speed train downlink, and tuning it together with the base station's beamformer, turns a channel that suffers Doppler shift and fast fading into one with consistently high gain. It proposes a block coordinate descent algorithm that alternates between optimizing the RIS phase shifts and the transmit beamforming vector, and reports an average channel gain improvement of 15 dB over fixed-phase or no-RIS baselines. In the simulated setting, this gain is enough to drive outage probability to zero and to improve rate, capacity, and bit error rate. If the result holds, it would make a passive surface a practical way to solve the central wireless problem of high mobility.","feed_headline":"RIS-aided train links gain 15 dB and lose their outage probability","feed_subtitle":"A joint phase-and-beamforming optimization keeps the train channel strong enough that outages vanish.","key_machinery":"A block coordinate descent loop that alternates between two blocks: the RIS phase-shift matrix and the transmit beamforming vector. The phase block is solved by using the structure of the line-of-sight path: the phase vector is chosen to cancel the Doppler phase e^{j2πk f_d T_c} and align with G w_k, and then a single phase rotation e^{jε_k} aligns the composite reflected path with the direct path. The beamforming block is the standard maximum-ratio transmission along the equivalent channel h_{r,k}^H Φ_k G + h_{d,k}^H. Together the two steps target the channel-gain objective F(w_k, Φ_k) in problem P1.","core_discovery":"The central claim is that in a Rician high-speed train MISO downlink, the joint optimization of RIS phases and an MRT beamformer—not the beamformer alone—unlocks most of the performance. The proposed BCD algorithm averages 15 dB higher channel gain than MRT with random or zero RIS phases, and with 1600 RIS elements the optimized RIS eliminates outage probability at a 10 dB SNR threshold. The mechanism is a two-step phase update: first the RIS phases cancel the Doppler phase on the line-of-sight component of the reflected channel and beamform that component toward the receiver, then a common scalar rotation aligns the reflected signal with the direct signal; the beamformer is then updated as","pith_inferences":["In my reading, the claimed optimality of the phase update stops after aligning the line-of-sight component; with Rician factor 3, the non-line-of-sight component still carries noticeable variance, so a phase update that also matches those parts should yield additional gain beyond the reported 15 dB.","The zero-outage result is tied to the chosen operating point: smaller RIS arrays, lower Rician factors, or thresholds above 10 dB should bring outage back, and the paper's own Gaussian approximation is the first place to test that.","A natural extension the paper does not pursue is applying the same BCD idea with discrete per-element phase constraints and imperfect channel estimates; the 15 dB and outage-elimination numbers would need recalibration under those constraints."],"forward_implications":["At the simulated parameters (1600 RIS elements, Rician factor 3, 360 km/h), the optimized RIS removes outage entirely rather than merely reducing it against a 10 dB SNR threshold.","Achievable rate grows with the number of RIS elements for the RIS-aided link, while the no-RIS baseline stays flat in the number of elements.","Channel capacity improves at every time slot, by up to about 1.5 kbps, suggesting the RIS suppresses the time-varying loss of the direct path.","Bit error rate first worsens and then improves as the train moves from the base station toward the RIS, meaning the reflected path substitutes for the direct path.","The 15 dB average gain is measured against schemes that use MRT without optimizing the RIS phase, so phase tuning itself is what carries most of the reported improvement."],"supporting_citations":[],"fun_headline_variants":["Train RIS: joint phase and beamforming yields 15 dB gain, no outages","Optimized RIS phases kill outage probability in train links, +15 dB","Joint RIS-beamer update gives 15 dB boost, zero outage for train links","Train MISO with optimized RIS: 15 dB stronger, outage-free"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that ignoring the non-line-of-sight part when setting the RIS phases, and then treating the whole optimized channel as one complex Gaussian, does not change the outcome; if either simplification is wrong, the 15 dB and zero-outage claims lose their support.","fun_headline_variants_meta":{"raw":{"variants":["Train RIS: joint phase and beamforming yields 15 dB gain, no outages","Optimized RIS phases kill outage probability in train links, +15 dB","Joint RIS-beamer update gives 15 dB boost, zero outage for train links","Train MISO with optimized RIS: 15 dB stronger, outage-free"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000565,"raw_usage":{"total_tokens":2482,"prompt_tokens":677,"completion_tokens":1805,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":421,"completion_tokens_details":{"reasoning_tokens":1722}},"tokens_in":421,"tokens_out":1805,"duration_ms":13294,"temperature":1.0,"reasoning_tokens":1722,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T17:56:41.888717+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the same high-speed MISO link at Rician factor 3 with 1600 elements, and for a fixed beamformer compare the paper's line-of-sight-only phase update against a phase vector found by coordinate ascent or a direct search over phases; a measurable gain increase would falsify the claimed optimality. Separately, compute the outage count from the empirical distribution of |h_hat_k|^2 instead of the Gaussian approximation in Eqs. (22)-(23) and check whether outage remains exactly zero across all time slots.","supporting_citations":[],"review_version":1}