{"id":"2c820fbd-b7bd-42dd-ab80-74a460df47ad","arxiv_id":"2504.20623","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In a cell-free MRT network, slow fluid-antenna multiple access can outperform fast switching when the base station antenna count is large, and user-side fluid antennas greatly reduce the base-station antennas needed for a target outage.","lead":"This paper analyzes a wireless network where base stations use simple beamforming and each user has a reconfigurable fluid antenna that switches among many receiving positions to dodge interference. It derives outage probability formulas and finds that a simpler slow-switching strategy can beat the fast-switching one when base stations have many antennas.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 7's validity condition Ω>ω is violated by the paper's own simulations (Ω≈1.36<N), so the s-FAMA curves behind the claimed crossover are outside the proved regime and (28)'s sums are ill-defined for non-integer b.","rationale":"The reader's weakest assumption already flags Theorem 7's condition Ω>ω and the approximate Nakagami model. My stress-test sharpens this into a direct threat to the headline result: the parameters used to demonstrate the s-FAMA/f-FAMA crossover violate the theorem's own validity condition, and the non-integer value of b makes the displayed finite sums ill-defined. This is not merely an accuracy question; it means the curves in Fig. 3 may not be derivable from (28) as stated. The concern is distinct from the reader's broader approximation-accuracy point, hence partial agreement. I do not recommend changing the conditional verdict because the issue is fixable: the authors could extend Theorem 7 to the full parameter regime, clarify the non-integer summation, or supply Monte Carlo verification of the crossover. The paper's other components—f-FAMA analysis, the noise-limited special case, and the qualitative intuition about interference variance—are not undermined by this concern, so rejection is not warranted.","tokens_in":23168,"tokens_out":10735,"duration_ms":114567,"concrete_test":"Re-evaluate the s-FAMA outage probability for the exact simulated settings (N=2,4,5,8, K=2 or 10, r=[200,400,600,800], α=3, thresholds 14 and 18) by Monte Carlo simulation of the original selection rule in (13), and compare with (28) evaluated with b=N+Ω−1. If the Monte Carlo curves for N≥4 do not reproduce the crossover, or if (28) cannot be evaluated without ad hoc floor/analytic continuation for non-integer b, the central claim is unsupported. As a second check, re-derive the integral identity used in Appendix E from [56, Proposition 1] and state exactly which parameter inequalities are required for the finite-sum form; if Ω>ω is indeed necessary, the plotted curves are invalid.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing concern is that the slow-FAMA outage expression used to demonstrate the central claim is stated in Theorem 7 only under the condition Ω>ω, where ω=N is the desired-signal Nakagami shape and Ω is the interference shape in (25). For the paper's own simulation geometry r=[200,400,600,800] and α=3, Ω=(Σ r_i^{-3})^2/Σ r_i^{-6} ≈ 1.36. Fig. 2 uses N=2 and Fig. 3 uses N from 2 upward, so the condition Ω>ω is false for every nontrivial case in those plots, including the crossover region N=4–5 where s-FAMA is claimed to overtake f-FAMA. Moreover, b=ω+Ω−1=N+Ω−1 is then non-integer (e.g., 2.36 for N=2), so the finite sums over q=0,...,b−1 and p=0,...,b−q−1 in (28) are not literally defined as written. The paper acknowledges only a slight discrepancy and does not reconcile the violated condition. Since the headline s-FAMA-beats-f-FAMA result is read off Fig. 3, this is a load-bearing gap: the plotted s-FAMA curves are not backed by the theorem that produces them.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a downlink cell-free network in which each base station (BS) has N fixed antennas using maximum ratio transmission (MRT) precoding and each user has a single fluid antenna with K ports. Assuming interference-limited conditions, it derives integral-form outage probabilities for fast FAMA (Eq. (22)) and slow FAMA (Eq. (28)), based on joint distributions of the desired-signal magnitude and the interference magnitude. The numerical section reports two headline findings: with sufficiently large N, s-FAMA can outperform f-FAMA (Fig. 3), and user-side FAS can substantially reduce the number of BS antennas and the CSI burden on the BS (Figs. 4-5). Monte Carlo simulations are provided for the cases shown.","tokens_in":23434,"tokens_out":8545,"duration_ms":79103,"significance":"The contribution is potentially useful for the FAMA literature: it extends FAMA to a cell-free MRT downlink and provides integral outage expressions that are compared with Monte Carlo simulations for the f-FAMA case and one s-FAMA example. The claim that slow switching can beat fast switching is counterintuitive and, if correct, would be an interesting design insight. The paper is also careful to distinguish the exact f-FAMA derivations from the approximate s-FAMA treatment. However, the s-FAMA analysis is the weakest link: the theorem underlying the s-FAMA outage expression is stated under a condition that the paper's own simulation parameters violate, and the moment-matched Gamma approximation is not quantified outside one geometry. The central crossover claim currently rests on expressions that are not valid as written.","major_comments":[{"comment":"Theorem 7 states that the s-FAMA outage probability expression (27)/(28) holds when Ω>ω, where ω=N is the desired-signal Nakagami shape parameter. For the geometry r=[200,400,600,800] and α=3 used in Figs. 2 and 3, Eq. (25) gives Ω≈1.36, so the condition Ω>ω fails for every N≥2, including the crossover region N=4-5 in Fig. 3 where s-FAMA is claimed to outperform f-FAMA. The s-FAMA curves in these figures are therefore not covered by the theorem that produces them, and the paper's headline claim is not established as written.","section":"Section III-C, Theorem 7; Section IV, Figs. 2-3"},{"comment":"Independently of the condition Ω>ω, the finite sums in (27) and (28) run over q=0,...,b-1 and p=0,...,b-q-1 with b=ω+Ω-1. For the paper's own parameters, Ω≈1.36 and ω=N, so b is non-integer (e.g., b=2.36 for N=2); such upper summation limits are not defined literally. The authors do not provide a convention for non-integer b or a generalized expression, and the acknowledgment of a 'slight discrepancy' in Section IV does not address this definitional gap. Since the s-FAMA curves in Fig. 3 are evaluated using (28), this issue is load-bearing.","section":"Eqs. (27)-(28) and Theorem 7"},{"comment":"The s-FAMA analysis replaces the sum of scaled Gamma random variables by a moment-matched Nakagami distribution. The validation shown in Fig. 2 is for a single geometry and reports only a 'slight discrepancy'; no error metric or parameter sweep is provided, and Fig. 3 does not overlay Monte Carlo points in the crossover regime. Because the approximation error can depend on the number of interfering BSs and the heterogeneity of the distances r_i, the crossover location and even the existence of the s-FAMA/f-FAMA crossover could shift; a robustness check is needed before the central claim can be accepted.","section":"Section III-C, Theorem 5; Section IV, Fig. 2 vs. Fig. 3"}],"minor_comments":[{"comment":"The sentence 'Additionally, ι = Nσ^2 Σ_{i=1}^U r_i^{-α} > 0' is inconsistent with the definition ι = r_0^{-α} Nσ^2 in Eq. (16); presumably r_0^{-α} Nσ^2 is intended.","section":"Appendix A, proof of Theorem 1"},{"comment":"Remark 1 states that the correlation parameter is reduced to μ=1 for K=1, but Eq. (4) is undefined (division by zero) for K=1; please clarify the convention used for this special case.","section":"Remark 1 and Eq. (4)"},{"comment":"The sentence 'In the case with with 4 transmit BS antennas' contains a duplicated 'with'.","section":"Section IV, discussion of Fig. 4"},{"comment":"The notation E[σ_s |g_k^[s]|] and Var[σ_s |g_k^[s]|] is confusing because σ_s appears inside the expectation on the left but the subsequent variance expression is written as σ_s^2 times a bracket; please make the role of σ_s explicit and align the notation with Eqs. (11)-(12).","section":"Eq. (33)"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about Theorem 7's condition is confirmed by direct computation from Eq. (25) and the stated simulation parameters. The f-FAMA analysis and the Monte Carlo comparisons look genuine, and the paper's framework is worth pursuing, but the central crossover claim cannot be accepted until Eq. (28) is replaced with a valid expression for the parameters actually simulated, or until the numerical study is restricted to regimes where the theorem applies and is validated by Monte Carlo. This is repairable, but it requires more than a local edit."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a real extension, not a repackaging. The authors take f-FAMA and s-FAMA, previously analyzed with a single-antenna BS and no path loss, and work out outage probabilities for a cell-free downlink where each BS uses MRT and users have one fluid antenna. The f-FAMA side looks solid: Theorem 4 and Corollary 1 track the Monte Carlo curves, and the generalization follows the known correlated-Nakagami machinery. The paper also earns credit for reporting the counter-intuitive crossover where s-FAMA beats f-FAMA for large N, and for testing it against simulation rather than just asserting it.\n\nThe soft spot is not minor. Theorem 7, which produces the s-FAMA outage expression, is stated only under Ω>ω, where ω=N is the desired-signal Nakagami shape and Ω is the interference shape from the moment-matched Gamma sum. In the paper's own geometry, r=[200,400,600,800] and α=3, Ω≈1.36, so the condition fails for every N≥2 shown in Figs. 2 and 3, including the crossover region N=4–5 that drives the central claim. And because b=ω+Ω−1 is then not an integer, the finite sums over q and p in (28) are not even defined as written. The authors mention only a 'slight discrepancy' and do not reconcile the violated condition. This is load-bearing: the s-FAMA curves behind the headline result are not backed by the theorem that supposedly produces them.\n\nThere are smaller issues. The moment-matched Nakagami approximation for the interference is only checked on a few parameter sets, and the abstract's claim about reducing CSI estimation overhead goes beyond what the model actually captures. The heavy self-citation is fine here because the prior results being extended are the right prior results.\n\nThis paper is for FAS/FAMA researchers and cell-free MIMO analysts. The f-FAMA half is a sound incremental contribution, and the s-FAMA crossover is an interesting observation that may well survive, but as written the proof doesn't cover the regime where it is demonstrated. This deserves serious peer review with a demand for a fix: either supply a proof of (28) without the Ω>ω restriction, or restrict the claims and the plots to the proved regime and define the sums for non-integer b. If that is done, this becomes a solid subfield paper. I wouldn't cite the s-FAMA result in its current form.","headline":"Useful extension of FAMA to MRT-based cell-free networks, but the headline s-FAMA crossover rests on a theorem whose condition fails in the paper's own simulations.","tokens_in":23952,"tokens_out":2698,"would_cite":false,"duration_ms":28256,"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":"In a cell-free network with MRT precoding and fluid-antenna users, slow port switching can outperform fast port switching once the base station antenna array is large enough, and the paper derives integral outage expressions for both modes.","keywords":["fluid antenna system","fluid antenna multiple access","cell-free network","MRT precoding","outage probability","fast port switching","slow port switching","Nakagami fading"],"falsifier":"Re-run the Monte Carlo simulations of Section IV with unequal interference distances r = [200, 400, 600, 800] and $N=8$, $K=10$, $\\alpha=3$, comparing the empirical outage probability with expression (28); because the Gamma approximation in Theorem 5 is exact only when all interference distances are equal, a mismatch concentrated in the low-SIR tail would show that the moment-matched Nakagami model, not the port-selection mechanics, is responsible for the reported slow-FAMA crossover.","tokens_in":22945,"feed_emoji":"📡","tokens_out":10579,"duration_ms":96488,"temperature":0.7,"pith_summary":"This paper asks whether a reconfigurable fluid antenna on the user side can carry some of the interference-management load that today sits on base-station antenna arrays. It analyzes a cell-free downlink in which each multi-antenna base station uses only maximum-ratio transmission (MRT) precoding toward its own user, while each user selects among K positions (ports) of a fluid antenna to dodge inter-user interference, the fluid antenna multiple access (FAMA) idea. The paper derives outage probability expressions for the two standard port-switching strategies, fast FAMA (per-symbol selection) and slow FAMA (per-coherence selection), and reports a result that overturns the earlier expectation that fast FAMA should always win: with a sufficiently large base-station antenna array, slow FAMA can outperform fast FAMA. If correct, the practical stake is scalability: user-side fluid antennas could reduce both the number of base-station antennas and the channel-state-information overhead needed for dense cell-free networks.","feed_headline":"Slow FAMA can beat fast FAMA when BS antenna arrays grow large","feed_subtitle":"New outage analysis of cell-free fluid antenna networks shows user-side FAS can shrink base station antenna needs.","key_machinery":"The load-bearing object is the joint distribution of the K correlated fluid-antenna port magnitudes. For the desired signal this is a correlated Nakagami law with shape $\\omega = N$ and spread $\\iota = r_0^{-\\alpha} N \\sigma^2$; for the fast-FAMA interference each port's magnitude is Rayleigh with variance $\\sigma_I^2 = \\sum_i r_i^{-\\alpha} \\sigma^2 \\sigma_s^2$; and for slow-FAMA interference the port magnitude is approximated by a Nakagami law with shape $\\Omega = (\\sum_i r_i^{-\\alpha})^2 / \\sum_i r_i^{-2\\alpha}$ and spread $\\phi = \\sigma^2 \\sum_i r_i^{-\\alpha}$, obtained by matching the first two moments of the sum of Gamma-distributed interference powers. Ports are correlated through a parameter $\\mu^2$ determined by a Bessel-function model of the fluid antenna's spatial correlation. The port-selection event, in which the user keeps the port with the largest ratio of desired-signal magnitude to interference magnitude, is converted via the joint CDF and an integral identity for Marcum Q-functions and modified Bessel functions into the integral outage formulas (22) and (28).","core_discovery":"The central claim is that, in an interference-limited cell-free FAMA network with MRT precoding, the outage probability of a typical user is accurately captured by the double-integral expression (22) for fast FAMA and by (28) for slow FAMA, and that these expressions reveal a crossover: fast FAMA wins when each base station has few antennas (below $N=3$ in the simulated settings), while slow FAMA wins once $N$ is large, with the crossover moving from $N=4$ to $N=5$ as the SIR threshold rises from 14 to 18. The paper also claims that user-side fluid antennas substantially reduce the base-station burden: in the simulated interference-limited scenario, a user with a 12-port fluid antenna needs only 2 base-station antennas to reach an outage probability below $10^{-3}$, while a fixed-antenna user needs 15. The same benefit appears under noise-limited conditions, though with smaller savings. The analysis treats the fast-FAMA interference distribution exactly, and the slow-FAMA interference distribution approximately; the paper validates both against Monte Carlo results and attributes a slight visible discrepancy in the slow-FAMA case to the approximation.","pith_inferences":["An implicit consequence the paper does not draw out: the regime where slow FAMA wins is also the regime where CSI estimation is easiest, since slow switching only needs port selection once per coherence time, so the practical gain of the crossover is larger than the outage curves alone suggest.","The two-moment matching suggests an accuracy boundary: when one interfering base station dominates (large spread in distances), the Gamma sum is far from a single Gamma law, so the outage prediction should degrade; a testable extension is a mixture-Nakagami or log-moment-matched approximation that restores accuracy in that regime.","The antenna-tradeoff numbers imply a design curve between base-station array size $N$ and fluid-antenna port count $K$; if the paper is right, operators could trade BS antennas for user-side ports at a target outage, and the crossover's dependence on the SIR threshold shows the trade sharpens as the threshold rises.","The result motivates an architectural division of labor: base stations use only single-user MRT while users handle residual interference locally, removing the need for network-wide channel knowledge; this is a consequence the paper points to but does not itself simulate."],"forward_implications":["With a sufficient number of base-station antennas, slow FAMA outperforms fast FAMA in the interference-limited regime, so per-symbol fast switching is not always the better mode.","User-side fluid antennas can substitute for base-station antennas: in the paper's simulations a 12-port FAS user needs only 2 BS antennas for outage below $10^{-3}$, while a fixed-antenna user needs 15.","The outage expressions (22) and (28) apply for arbitrary base-station distances and path-loss exponents, extending single-cell FAMA analysis to cell-free deployments.","In noise-limited conditions, FAS still reduces the number of BS antennas required, but the savings are smaller than in interference-limited conditions.","Because each BS needs only MRT-precoding CSI for its own user, FAMA-equipped cell-free networks can avoid the network-wide CSI sharing that zero-forcing-style interference management requires."],"supporting_citations":[{"why":"Introduces fast fluid antenna multiple access and the per-symbol port-selection outage framework that the paper generalizes to cell-free MRT networks.","marker":"[39]"},{"why":"Introduces slow fluid antenna multiple access, supplies the s-FAMA analysis extended here, and states the expectation, overturned in this paper, that fast FAMA is superior.","marker":"[41]"},{"why":"Provides the fluid-antenna channel and correlation model, including the correlated Rayleigh joint PDF used for fast-FAMA interference.","marker":"[12]"},{"why":"Supplies the closed-form spatial correlation parameter $\\mu^2$ used to model port correlation in both signal and interference.","marker":"[13]"},{"why":"Supplies the joint PDF and CDF of correlated Nakagami-m magnitudes that underpin Theorems 2 and 3 for the desired signal across ports.","marker":"[53]"},{"why":"Provides the integral identity for combinations of Marcum Q-functions, modified Bessel functions, and power functions used to evaluate the outage integrals.","marker":"[56]"},{"why":"Gives the Gamma-sum approximation used in Theorem 5 to model slow-FAMA interference as Nakagami.","marker":"[60]"}],"fun_headline_variants":["Slow FAMA outshines fast when base stations have many antennas","Cell-free FAMA: slow port switching wins with large antenna arrays","User fluid antennas slash base station antenna count in cell-free networks","In interference, slow FAMA beats fast FAMA for large antenna arrays","Fluid antennas at users reduce BS antenna needs in cell-free FAMA"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The slow-FAMA outage analysis rests on approximating the combined interference from several base stations at different distances by a single Nakagami distribution matched only through its first two moments, then assuming the K ports share the same correlated structure as the desired signal; if that approximation is inaccurate for heterogeneous base-station distances, the outage curve and the slow-versus-fast crossover shift, and the paper itself notes a slight discrepancy in Fig. 2, while the condition $\\Omega > \\omega$ used in Theorem 7 is not met by some simulated parameter choices.","fun_headline_variants_meta":{"raw":{"variants":["Slow FAMA outshines fast when base stations have many antennas","Cell-free FAMA: slow port switching wins with large antenna arrays","User fluid antennas slash base station antenna count in cell-free networks","In interference, slow FAMA beats fast FAMA for large antenna arrays","Fluid antennas at users reduce BS antenna needs in cell-free FAMA"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0002,"raw_usage":{"total_tokens":1428,"prompt_tokens":1054,"completion_tokens":374,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":670,"completion_tokens_details":{"reasoning_tokens":282}},"tokens_in":670,"tokens_out":374,"duration_ms":4180,"temperature":1.0,"reasoning_tokens":282,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:25:41.290785+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the Monte Carlo simulations of Section IV with unequal interference distances r = [200, 400, 600, 800] and $N=8$, $K=10$, $\\alpha=3$, comparing the empirical outage probability with expression (28); because the Gamma approximation in Theorem 5 is exact only when all interference distances are equal, a mismatch concentrated in the low-SIR tail would show that the moment-matched Nakagami model, not the port-selection mechanics, is responsible for the reported slow-FAMA crossover.","supporting_citations":[{"cited_title":"Fluid antenna multiple access,","cited_arxiv_id":null,"evidence_quote":"Introduces fast fluid antenna multiple access and the per-symbol port-selection outage framework that the paper generalizes to cell-free MRT networks."},{"cited_title":"Closed-form expressions for spatial correlation parameters for performance analysis of fluid antenna systems,","cited_arxiv_id":null,"evidence_quote":"Supplies the closed-form spatial correlation parameter $\\mu^2$ used to model port correlation in both signal and interference."},{"cited_title":"Enhancing QoS through fluid antenna systems over correlated Nakagami-m fading channels,","cited_arxiv_id":null,"evidence_quote":"Supplies the joint PDF and CDF of correlated Nakagami-m magnitudes that underpin Theorems 2 and 3 for the desired signal across ports."},{"cited_title":"Laplace transform of product of gen- eralized Marcum Q, Bessel I, and power functions with applications,","cited_arxiv_id":null,"evidence_quote":"Provides the integral identity for combinations of Marcum Q-functions, modified Bessel functions, and power functions used to evaluate the outage integrals."},{"cited_title":"Multiuser MIMO in distributed antenna systems with out-of-cell interference,","cited_arxiv_id":null,"evidence_quote":"Gives the Gamma-sum approximation used in Theorem 5 to model slow-FAMA interference as Nakagami."}],"review_version":1}