{"id":"142b8b6a-510a-492a-86d9-e6623b4220f8","arxiv_id":"2603.08472","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Multi-mode pinching-antenna systems using mode combining or mode selection can outperform single-mode PASS and hybrid beamforming in sum rate.","lead":"Two protocols for multi-mode pinching-antenna wireless systems are compared: locking each antenna to one guided mode, or letting it mix several modes. Simulations suggest multi-mode systems beat single-mode pinching antennas and hybrid beamforming, but the printed channel model omits user locations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (6) defines identical distances for all users (no user index), so the printed model cannot support the multi-user multiplexing that the headline gains depend on.","rationale":"The reader's weakest_assumption is precisely the missing user index in Eq. (6). I read the full manuscript and the strongest claim: the gains rely on mode-domain multiplexing of two users. The printed channel model makes all user channels identical, which contradicts the simulation setup. This is the most load-bearing concern because it strikes at the core of the paper's contribution. The proposed remedy—checking the released code and confirming a user-dependent channel—would settle whether this is a typographical omission or a fundamental error. I see no reason to change the CONDITIONAL verdict; the paper needs a corrected equation and a reproducibility check before its claims can be accepted.","tokens_in":9040,"tokens_out":3029,"duration_ms":27177,"concrete_test":"Download the released code at https://github.com/xiaoxiaxusummer/multi_mode_pinching_antenna and extract the exact user-coordinate setup used to generate Figs. 2-3. Verify whether the channel construction uses a user-dependent distance in R_{n,k} (e.g., distinct user x/y positions). If it does, Eq. (6) is a typo and the central multiplexing claim may be valid after correction; if it does not, the reported multi-user rates are impossible under the stated model and the central claim fails. An additional analytic check: compute the rank of H_eff for the intended user positions; rank ≥ 2 is necessary for two-user multiplexing.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Eq. (6), the LoS distance is defined as R_{n,k} = sqrt(x_n^2 + h_PA^2), with no dependence on user k. Thus every user sees the same channel vector h_k(x) = h(x), so the effective channel H_eff in Eq. (7) has identical columns. The sum-rate objective (9)-(10) then collapses to a single-user rate (or at most rank-1 transmission). The two-user rate gains in Figs. 2-3—and the central claim of mode-domain multiplexing—are not derivable from the system model as printed. Either the authors intended user-dependent coordinates (e.g., R_{n,k} = sqrt((x_n - u_k)^2 + h_PA^2)) but omitted them, or the simulation results are inconsistent with the equations. This is load-bearing because the entire letter is about multi-user multiplexing; if all user channels are identical, the qualitative comparison against single-mode TDMA and hybrid beamforming reduces to a single-SNR comparison with no multiplexing benefit.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The letter proposes two protocols for multi-mode pinching-antenna systems (PASS): mode selection, in which each PA's propagation constant is matched to one guided mode, and mode combining, in which the propagation constant is continuously tuned to couple to multiple modes. A sum-rate maximization problem is formulated for a downlink multi-user system with jointly optimized digital beamforming, PA positions, and PA propagation constants, and a PSO-based algorithm with a KKT-parameterized beamforming structure (PSO-KPBF) is proposed. Simulation results are presented to show that multi-mode PASS, including a simple uniform mode-combining design, outperforms single-mode PASS and conventional hybrid MISO beamforming, and that mode combining achieves higher spectral efficiency than mode selection at the cost of higher hardware complexity. The paper also releases code on GitHub.","tokens_in":9297,"tokens_out":3954,"duration_ms":38598,"significance":"If the central claims are correct, the paper offers a useful architectural comparison for a recently proposed technology: it gives two concrete operating protocols, a tractable heuristic optimization framework, and a reproducible implementation. The idea of exploiting multiple guided modes in PASS to support multi-user multiplexing from a single waveguide is timely and could be of interest to the antenna-array and MIMO communities. The release of code is a positive feature. However, the current manuscript contains a model inconsistency that directly undermines the multi-user multiplexing claims: the channel vector in Eq. (6) has no user index, so all users see the same channel and the effective channel matrix is at most rank one. In addition, the precoder expression in Eq. (11) has a dimension inconsistency that makes the central algorithmic step undefined as printed. These issues must be resolved before the reported gains can be assessed.","major_comments":[{"comment":"The LoS channel from PA n to user k is defined as R_{n,k} = sqrt(x_n^2 + h_PA^2), with no dependence on k. Consequently h_k(x) is the same vector for every user, so the effective channels h_eff,k = G^H h_k are identical for all k, and H_eff has rank one. The sum-rate objective (9)-(10) then degenerates to a single-user rate, and the two-user multiplexing gains shown in Figs. 2 and 3 cannot be produced by the printed model. The simulations must therefore rely on an unstated user-position model (e.g., R_{n,k} = sqrt((x_n - u_k)^2 + h_PA^2)). This is load-bearing: the entire letter is about mode-domain multiplexing, and the qualitative comparison against TDMA/hybrid beamforming is meaningless if all user channels coincide. Please state the user-dependent distance explicitly and update Eq. (6) accordingly.","section":"Eq. (6), Section II-B"},{"comment":"There is a dimensional inconsistency in the definition of the KPBF precoder. In Eq. (7) H_eff is defined as a M x K matrix (columns are heff,k in C^{M x 1}), but Eq. (11) says 'Given H_eff in C^{K x M}' and then defines fW = P_rel^{1/2} (I_M + (1/σ^2) H_eff^H Λ H_eff)^{-1} H_eff^H. With H_eff K x M, the matrix H_eff^H Λ H_eff is M x M, so the inverse and H_eff^H are conformable, but left-multiplying by P_rel^{1/2} (which is K x K if p_rel is length K) is not conformable. If H_eff is instead M x K, then H_eff^H Λ H_eff is undefined (K x M times K x K). The equations as printed cannot define W. Moreover, the claim that this parameterization spans KKT-stationary solutions of the WMMSE sum-rate problem needs a clear statement of the exact transformation and a derivation or a precise reference. This is load-bearing because Algorithm 1 reconstructs the precoder using Eq. (11).","section":"Eq. (11), Section III-A"},{"comment":"The numerical results are presented as single curves with no error bars, no number of random trials, no convergence trajectories for the PSO, and no explicit user positions or channel parameters. Since PSO is a stochastic heuristic for a highly nonconvex problem, the claimed relative ordering of mode selection, mode combining, and uniform mode combining could be an artifact of a particular initialization. The GitHub link is a positive step, but the letter should at least report means over independent runs and specify the random seed or the variance across runs. In addition, the simulation setup should state the user coordinates, the PA length, the waveguide length, and other parameters that are currently deferred to the repository; this is particularly important because Eq. (6) omits the user-position dependence.","section":"Section IV"}],"minor_comments":[{"comment":"The formula for the uniform mode-combining propagation constant is ambiguous: it can be read as β1 + β2 + ... + βM/2 rather than (β1+β2+...+βM)/2. Please add parentheses or clarify in words.","section":"Eq. (5), Remark 2"},{"comment":"Typos: 'addictive white Gaussian noise' should be 'additive white Gaussian noise'; 'hardwre complexity' should be 'hardware complexity'.","section":"Abstract/Conclusion"},{"comment":"The complexity analysis states O(IKM) for forming H_eff, but the problem dimensions suggest O(NKM) since G is N x M and h_k is N x 1. Please correct the notation.","section":"Algorithm 1, Section III-B"},{"comment":"The caption uses 'PA=(1+2)/2' and the figure text writes 'PA=( 1+ 2)/2'; this should be 'β_PA = (β1+β2)/2' or similar.","section":"Fig. 3 caption"},{"comment":"Eq. (7) defines H_eff in C^{M x K}, but Eq. (11) uses H_eff in C^{K x M}. Even if the dimension issue in the major comment is resolved, please keep the notation consistent throughout.","section":"Section II-B"}],"recommendation":"major_revision","confidential_remarks":"The most serious issue is the missing user index in Eq. (6), which appears to be a typographical omission rather than a deliberate modeling choice, since the simulation figures clearly show two users. Nevertheless, it is load-bearing and must be fixed. The dimension inconsistency in Eq. (11) is also likely fixable, but it currently makes the central algorithm undefined. The paper is a reasonable letter-sized contribution if these points are corrected and the simulation claims are properly qualified; the heuristic PSO results would benefit from statistical reporting."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nIf you're thinking about reading this one, know this first: the headline claim of multi-user multiplexing gains doesn't follow from the printed model. Eq. (6) defines R_{n,k} = sqrt(x_n^2 + h_PA^2) with no user index, so every user sees the same LoS channel vector. That makes the effective channel rank one, and the sum-rate problem collapses to a single-user rate. The two-user rate curves in Figs. 2 and 3 cannot be produced by the equations as written. The stress-test note is right, and I haven't found anything in the paper that fixes it.\n\nThat's a shame, because the underlying idea is worth a look. The paper compares two operating protocols for multi-mode pinching-antenna systems: mode selection, where each PA phase-matches to a single guided mode, and mode combining, where the propagation constant is tuned continuously to radiate multiple modes. The formulation is sensible, and the authors propose a PSO-KPBF algorithm that embeds a KKT-parameterized precoder into a particle swarm over PA positions and propagation constants. They also release code, which is a real plus for a paper like this.\n\nThe soft spots beyond the Eq. (6) issue are milder. The simulation section is thin: one curve per scheme, no error bars, no sensitivity analysis for the PSO. The KKT parameterization is asserted via citation to their earlier work rather than derived here; that's acceptable in a letter, but a referee might ask for a reminder of the stationarity conditions. The self-citation pattern is heavy but fair given the line of research.\n\nMy take: the concept—mode-domain multiplexing on a single waveguide—is plausible and relevant to 6G antenna design. The paper deserves a serious referee, but only with the channel model corrected. If the authors intended R_{n,k} = sqrt((x_n - u_k)^2 + h_PA^2) and simply omitted the user positions, that's a fixable typo, but it's load-bearing. If that's not the intended model, the simulations are unexplained. Either way, this needs a major revision before acceptance.\n\nI'd bring the revised version to a reading group; the current one is useful mainly as a cautionary example of an index error.","headline":"The mode-combining idea is interesting, but Eq. (6) as printed gives all users the same channel, so the multiplexing gains in the simulations aren't reproducible from the stated model.","tokens_in":9779,"tokens_out":3000,"would_cite":false,"duration_ms":27561,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A single dielectric waveguide can carry several orthogonal guided modes, and pinching antennas along it can multiplex multiple users in the mode domain, provided each user has a distinct line-of-sight channel.","keywords":["multi-mode pinching antenna systems","mode-domain multiplexing","mode selection","mode combining","particle swarm optimization","KKT-parameterized beamforming","sum-rate maximization","dielectric waveguide"],"falsifier":"Run the released simulation code with the channel distance formula exactly as printed in Eq. (6): if R_{n,k} has no user-dependent term, the sum-rate curves in Figs. 2-3 cannot be reproduced for two distinct users; alternatively, check whether user coordinates enter the distance computation in the code.","tokens_in":8921,"feed_emoji":"📡","tokens_out":3786,"duration_ms":34199,"temperature":0.7,"pith_summary":"This paper claims that a single dielectric waveguide can carry several orthogonal guided modes, and that pinching antennas placed along it can multiplex multiple users in the mode domain. Two operating protocols are proposed: mode selection, in which each antenna phase-matches one mode, and mode combining, in which each antenna continuously tunes to radiate several modes. The authors formulate a sum-rate maximization problem and solve it with a particle-swarm algorithm whose beamforming step is parameterized by KKT optimality conditions. Simulation results indicate that multi-mode operation beats single-mode pinching antennas and conventional hybrid beamforming, and that even a fixed, uniform mode-combining design does so. If this holds, mode-domain multiplexing becomes a practical single-waveguide path to multiuser millimeter-wave links.","feed_headline":"One waveguide carrying two modes serves two users at once","feed_subtitle":"A single dielectric waveguide with two orthogonal modes can multiplex two users; even fixed combining outperforms hybrid beamforming.","key_machinery":"The central object is the mode-domain effective channel h_eff,k = G^H(x,beta) h_k(x), where G contains the in-waveguide coupling coefficients between each pinching antenna and each guided mode, and h_k is the line-of-sight wireless channel from antennas to user k. The coupling coefficient eta_n,m is governed by coupled-mode theory and depends on the phase mismatch between the antenna's tunable propagation constant beta_n^PA and the mode's propagation constant beta_m. The paper's optimizer is PSO-KPBF: particle swarm optimization over antenna positions x, propagation constants beta^PA, and low-dimensional beamforming parameters (lambda, p_rel), with the precoder reconstructed from the KKT con","core_discovery":"The central claim is that the extra degrees of freedom from multiple guided modes in a pinching-antenna waveguide are exploitable: by tuning each pinching antenna's propagation constant, the system shapes how much power leaks from each mode, creating an effective multi-antenna channel from one physical waveguide. Mode selection restricts each antenna to phase-match one mode, suppressing the others; mode combining leaves the propagation constant continuous so several modes radiate at once. The paper's proposed PSO-KPBF algorithm jointly optimizes antenna positions, propagation constants, and a KKT-parameterized precoder, and its simulations show mode combining reaches the highest spectral eff","pith_inferences":["The same KKT-parameterized swarm approach could apply to other position-dependent antenna architectures, such as fluid or movable antennas, where the channel depends on continuous placement variables, though the paper only demonstrates it for pinching antennas.","The channel-model ambiguity around user positions should be resolved before comparing protocols: if users have identical line-of-sight vectors, the reported multiuser gains would not follow from the printed equations.","A natural extension is to more than two modes and more than two users; the paper's M=2, K=2 simulations do not show whether the mode-combining advantage scales.","Hardware prototypes of spoof-surface-plasmon or liquid-crystal-tuned waveguides could validate the assumed sinusoidal coupling model and measure real insertion losses."],"forward_implications":["Mode-domain multiplexing can serve multiple users from a single dielectric waveguide, removing the need for multiple waveguides or large antenna arrays.","Mode combining gives the highest spectral efficiency across power and antenna-count regimes, while mode selection approximates it with lower hardware complexity.","Even a preconfigured uniform mode-combining design, with no per-antenna tuning, outperforms single-mode PASS and hybrid beamforming in the reported settings.","The KKT parameterization shrinks the search space for beamforming, making joint position and propagation-constant optimization feasible with a modest number of particles and iterations.","Multi-mode PASS gains grow with transmit power as the system becomes interference-limited, because the mode-domain precoder can manage interference across modes."],"fun_headline_variants":["One waveguide, two modes, multiple users","Pinching antennas get a second mode for free","Mode combining wins over hybrid beamforming in pinching-antenna test","Two modes on one waveguide: a new multiplexing trick"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The multiuser benefit rests on users sitting at different locations so their line-of-sight channels differ, but the printed channel model omits user coordinates, so as written every user sees the same channel vector.","fun_headline_variants_meta":{"raw":{"variants":["One waveguide, two modes, multiple users","Pinching antennas get a second mode for free","Mode combining wins over hybrid beamforming in pinching-antenna test","Two modes on one waveguide: a new multiplexing trick"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000497,"raw_usage":{"total_tokens":2293,"prompt_tokens":786,"completion_tokens":1507,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":530,"completion_tokens_details":{"reasoning_tokens":1442}},"tokens_in":530,"tokens_out":1507,"duration_ms":10055,"temperature":1.0,"reasoning_tokens":1442,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T18:31:58.344562+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the released simulation code with the channel distance formula exactly as printed in Eq. (6): if R_{n,k} has no user-dependent term, the sum-rate curves in Figs. 2-3 cannot be reproduced for two distinct users; alternatively, check whether user coordinates enter the distance computation in the code.","supporting_citations":[],"review_version":1}