{"id":"0bb491c9-2089-494d-9ac0-03ce6dcc5ee1","arxiv_id":"2506.00811","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"A true/fake frequency multiplexing scheme with joint power allocation and a fitted multiplexing factor aims to hide real data inside decoy signals in wireless adversarial channels.","lead":"This paper proposes a wireless transmission scheme that sends fake decoy signals to deceive an eavesdropper while hiding the real signal in overlapping frequency bands. It might interest engineers working on physical-layer security because it combines decoy signaling with optimized power allocation to improve both secrecy and deception.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"P1→P2 drops |h_e,i|^2 in deception constraint (11c), so P2-optimal powers can violate the decoy-dominance requirement; Section V results do not establish the claimed CTSF guarantee.","rationale":"The reader's weakest_assumption is exactly the load-bearing point. The claim that CTSF+BADO outperforms benchmarks in both secrecy and deception depends on Algorithm 1 solving P1. Since P2 is the problem actually solved, and P2 constraint (15e) is not equivalent to P1 constraint (11c), the feasible region is wrong and the output is not guaranteed to satisfy the deception condition. This is an internal inconsistency, not a disagreement with prevailing consensus: even accepting Remark 1, the variable substitution drops the Eve channel gains. There is no code or machine-checked proof that could indirectly validate the reformulation, and Section V does not report whether (11c) holds at the returned power allocations. The residual-interference admission in Theorem 2 is secondary; it is a modeling idealization, whereas the P2 reformulation error changes the optimization problem itself. I therefore see no basis for keeping the paper's central claim without a corrected reformulation and rerun simulations, so the reader's REJECT verdict remains appropriate.","tokens_in":16372,"tokens_out":9740,"duration_ms":87915,"concrete_test":"Analytical check: choose one true band k and one fake band n with c_n=c_k=1, |h_e,n|^2=0.1, |h_e,k|^2=1, ξ_n=1, ξ_k=0.5. Then (15e) holds (1≥0.5) while (11c) fails (0.1<0.5), proving the reformulation changes the feasible set. Computational follow-up: run Algorithm 1 on random Rician realizations, recover p_i=ξ_i/c_i using the Newton-fitted α, and evaluate (11c) directly; if any returned point violates (11c), P1 and P2 are not equivalent. Replacing (15e) by the correct constraint ξ_n|h_e,n|^2≥ξ_k|h_e,k|^2 and re-running BADO will show a different secrecy/deception trade-off than Figs. 3–8.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Even granting the paper's Remark 1, the central derivation fails at the P1-to-P2 reformulation. P1 constraint (11c) is p_n|h_e,n|^2 ≥ p_k|h_e,k|^2, the necessary condition for the decoy signal to dominate at Eve in (10). With the announced substitution ξ_i = p_i c_i, and with self-correlation c_n=c_k=1, this constraint is exactly ξ_n|h_e,n|^2 ≥ ξ_k|h_e,k|^2. P2 instead imposes (15e), ξ_n ≥ ξ_k. The two coincide only when |h_e,n|^2 = |h_e,k|^2, which is not guaranteed and is generally false under the Rician fading model with user-dependent channel means and variances. Consequently, P2's feasible set is not equivalent to P1's: a solution of P2 can satisfy all P2 constraints while violating (11c), so the power allocation returned by Algorithm 1 is not certified to make the decoy signals dominate at Eve. The simulation comparisons in Section V evaluate this weakened problem, so the claimed anti-interception and deception advantage over OFDM and equal-power benchmarks is not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a 'conceal truth while show fake' (CTSF) transmission scheme in which multiple sources transmit true and decoy signals over non-orthogonal frequency bands. It defines interception and deception probabilities, formulates an optimization problem P1 that maximizes the sum secrecy rate of the true signals subject to a decoy-dominance constraint at the eavesdropper, and proposes a bi-stage alternating dual-domain optimization (BADO) algorithm together with a Newton's method for fitting the T/F frequency-multiplexing factor alpha. Numerical simulations compare the proposed scheme with OFDM and equal-power allocation and report higher secrecy rate and better deception performance.","tokens_in":16672,"tokens_out":8173,"duration_ms":83539,"significance":"The basic idea of simultaneously hiding true information in overlapping spectrum and actively deceiving an eavesdropper is timely and practically relevant for adversarial wireless scenarios. The paper provides an explicit multi-source signal model, probability-based performance metrics, a complexity analysis of the proposed algorithms, and a systematic set of simulation comparisons. However, the central optimization reformulation is not equivalent to the original problem, and the proposed Newton fit for alpha is tautological. Because the numerical results are generated by solving the weakened reformulated problem, the claimed advantages over the benchmarks are not established. As presented, the contribution is therefore not ready for publication.","major_comments":[{"comment":"The reformulation of P1 into P2 is not equivalent. Constraint (11c) is p_n |h_e,n|^2 >= p_k |h_e,k|^2 for all n in \\tilde{K} and k in K; substituting xi_i = p_i c_i gives (xi_n/c_n)|h_e,n|^2 >= (xi_k/c_k)|h_e,k|^2, and with c_n=c_k=1 for the reference bands, xi_n |h_e,n|^2 >= xi_k |h_e,k|^2. The P2 constraint (15e) is simply xi_n >= xi_k, which drops the Eve channel gains |h_e,n|^2 and |h_e,k|^2. Under the Rician fading model these gains are generally unequal, so the P2 feasible set is strictly larger than that of P1; a P2-optimal solution can violate the decoy-dominance condition (11c), and the Section V results therefore do not certify the claimed deception guarantee.","section":"Section IV, Eqs. (11c) and (15e)"},{"comment":"The Newton fit is circular and does not implement joint optimization over alpha. The quantity c_i in (18) is the same quantity defined in (4) as the squared sinc ratio, so minimizing the squared error between c_i and that expression simply recovers the alpha that produced c_i; it is a definitional identity, not an independent optimization. Moreover, after the substitution xi_i = p_i c_i, P2 is solved with c_i absorbed into xi_i, so alpha is no longer a decision variable; fitting alpha after recovering p_i and c_i cannot restore the joint optimization over {p_i} and alpha stated in P1. The paper therefore does not support the claim that power allocation and the T/F multiplexing factor are jointly optimized.","section":"Section IV, Eq. (18) and Algorithm 2"},{"comment":"The theorem is not correct as stated. The derivative in (25) is dR_{s,k}/dp_k = a_k/(C_k ln2(1 + a_k p_k/C_k)) - a_{k,e}/(C_{k,e} ln2(1 + a_{k,e} p_k/C_{k,e})), whose sign depends on the relative magnitudes of the Bob and Eve channel gains and the interference denominators. The proof's assertion that this derivative is positive is unjustified, and the subsequent interpretations of Figs. 3 and 5 rely on this monotonicity.","section":"Section IV, Theorem 1"}],"minor_comments":[{"comment":"The displayed T1 objective appears malformed; it should presumably be a sum over k of log(b_k mu_k), not the logarithm of a sum of logarithms, and b depends on k although the notation in (16a) suggests a single b.","section":"Section IV, Eq. (16a)"},{"comment":"The symbols alpha_k and alpha_{k,e} are introduced without definition and appear to conflict with the frequency-multiplexing factor alpha; they should be renamed to the channel gains used in (25).","section":"Section IV, Eq. (24)"},{"comment":"The simulation section reports no error bars, confidence intervals, or number of channel realizations, so it is difficult to assess the statistical significance of the observed advantages over the benchmarks.","section":"Section V"},{"comment":"The equivalence leading to (8) should state its implicit assumptions, namely c_n=c_k=1 and equal noise powers at Eve, so that the subsequent constraint derivation can be checked.","section":"Remark 1, Eq. (8)"}],"recommendation":"reject","confidential_remarks":"The two load-bearing issues are independent: the reformulation error invalidates the claimed feasibility guarantee, and the alpha-fitting procedure invalidates the joint-optimization claim. Because the simulations solve the reformulated problem, they do not provide evidence for the original P1. A repair would require re-deriving the equivalence with channel-dependent constraints and a genuine joint treatment of alpha, which goes beyond a routine revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague: The paper's core idea is genuinely different—use non-orthogonal frequency multiplexing so decoy signals dominate at Eve while true signals are hidden, and optimize power/correlation for sum secrecy. I haven't seen this exact T/F multiplexing model with a variable correlation factor before, so there is something new here. The definitions of interception and deception probabilities are reasonable, and the simulations show the scheme beating OFDM and equal-power baselines in secrecy and deception rates.\n\nThat said, the math has a load-bearing gap. The jump from P1 to P2 is not equivalent. Constraint (11c) is p_n |h_e,n|^2 ≥ p_k |h_e,k|^2; with ξ_i = p_i c_i and self-correlation c=1, that becomes ξ_n |h_e,n|^2 ≥ ξ_k |h_e,k|^2. P2 instead imposes ξ_n ≥ ξ_k. These coincide only when the Eve channel gains are equal, which is not true under Rician fading with user-dependent means. So a P2-optimal power allocation can violate the decoy-dominance condition on actual channels, and the simulation results in Section V evaluate this weakened problem. The authors seem not to have noticed.\n\nThe α-fitting step is also circular. The c_i used as data in Algorithm 2 are defined in Eq. (4) as functions of α; after the ξ optimization you get c_i = ξ_i / p_i, and then you fit α to make the sinc expression match. That's just inverting the defining equation, not optimizing α over the secrecy rate. The paper even concedes in the Theorem 2 discussion that residual decoy interference at Bob is ignored in the main model, which further weakens the numerical claims.\n\nMinor issues: inconsistent notation, no error bars, no code released. But those are secondary.\n\nBottom line: this deserves a serious referee, not a desk reject. The idea is interesting enough that an editor should send it out, but the current form doesn't substantiate the central claim. I'd tell the authors to fix the constraint transformation—either carry the channel gains through or redefine ξ_i to include |h_e,i|^2—and replace the circular Newton fit with a real optimization over α, if that's what they intend. Then the secrecy/deception trade-off might actually hold.","headline":"Interesting deception-plus-secrecy scheme, but the P1→P2 reformulation drops Eve channel gains and the α-fit is circular, so the performance claims don't yet hold up.","tokens_in":17141,"tokens_out":3438,"would_cite":false,"duration_ms":32893,"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":"The paper proposes a true/false frequency multiplexing scheme that hides confidential signals inside decoy signals, so an eavesdropper locks onto the fake while the true bits stay secure.","keywords":["anti-interception transmission","conceal truth while show fake","T/F frequency multiplexing","secrecy rate","deception probability","power allocation","physical layer security","non-orthogonal frequency division multiplexing"],"falsifier":"Take a Rician-faded channel realization with $|h_{e,n}|^2 < |h_{e,k}|^2$ on the eavesdropper link, run the proposed optimization, and check whether the output satisfies $p_n|h_{e,n}|^2 \\geq p_k|h_{e,k}|^2$ whenever $\\xi_n \\geq \\xi_k$; a single realization where the optimized powers satisfy $\\xi_n \\geq \\xi_k$ but violate the original dominance inequality would show the reformulation is not equivalent and the deception guarantee does not hold.","tokens_in":16176,"feed_emoji":"📡","tokens_out":7621,"duration_ms":67815,"temperature":0.7,"pith_summary":"This paper argues that a transmitter can do two things at once against a passive eavesdropper: keep confidential bits secret and actively feed the eavesdropper a fake signal. It proposes a true/false (T/F) frequency multiplexing scheme in which multiple sources transmit true and decoy signals on intentionally overlapping non-orthogonal frequency bands, so the decoy signal is what the eavesdropper locks onto while the true signal is buried in the overlap. The central optimization maximizes the sum secrecy rate of the true signals subject to a deception constraint that the decoy signal's SINR at the eavesdropper stays above a threshold and dominates the true signal's SINR. A bi-stage alternating dual-domain optimization approach jointly allocates power and correlation coefficients, and a Newton method fits the frequency multiplexing factor. Simulations with four sources under Rician fading report higher secrecy rates and lower interception probabilities than orthogonal frequency-division multiplexing and equal-power baselines.","feed_headline":"Fake signals hide true data while fooling eavesdroppers","feed_subtitle":"A frequency-multiplexing scheme optimizes power so decoys dominate at the eavesdropper while true-signal secrecy stays high","key_machinery":"The load-bearing objects are the T/F frequency multiplexing factor $\\alpha$ and the induced correlation coefficient $c_i = \\left| \\frac{\\operatorname{sinc}(\\alpha(i-k))}{\\operatorname{sinc}(\\alpha(i-k)/K)} \\right|^2$, which measures how much two frequency bands overlap and therefore how much interference a signal in band $i$ creates in band $k$. The optimization substitutes $\\xi_i = p_i c_i$, converting the decoy-dominance constraint $p_n|h_{e,n}|^2 \\geq p_k|h_{e,k}|^2$ into $\\xi_n \\geq \\xi_k$; the bi-stage alternating dual-domain optimization then alternates between solving for auxiliary ratio variables and for $\\xi_i$, each subproblem being convex. A Newton iteration later fits $\\alpha$ from the optimized correlation coefficients.","core_discovery":"The authors claim that the 'conceal truth while show fake' goal is achievable as a single optimization problem: maximize the sum secrecy rate of true signals while forcing the decoy signal received by the eavesdropper to be stronger than any true signal it is trying to intercept. The mechanism is non-orthogonal frequency multiplexing: true and decoy signals share overlapping bands, so at the legitimate receiver the known decoys are cancelled and only true-signal interference remains, while at the eavesdropper all overlapping signals interfere and can be steered so the decoy dominates. The paper formalizes interception and deception probabilities and shows numerically that, for a certain deception-threshold range, the deceptive probability stays at one while the interception probability falls to zero.","pith_inferences":["The paper assumes Bob knows the decoy signals exactly and cancels them; with imperfect cancellation, residual decoy interference would reduce true-signal SINR, so the secrecy-rate gains are an upper bound until cancellation errors are modeled.","A natural extension the paper does not pursue is to optimize the frequency multiplexing factor jointly inside the secrecy-rate problem instead of fitting it after power allocation, which could reveal whether the overlap factor itself is the main deception lever.","The equivalence gap between $\\xi_n \\geq \\xi_k$ and decoy dominance suggests a reformulation that keeps the eavesdropper's channel gains in the constraint, or optimizes over channel uncertainty, would be needed before deployment in fading channels.","If Eve is modeled as adaptive and able to estimate the correlation coefficients, the deception guarantee weakens; testing the scheme against such an informed eavesdropper is left for future work."],"forward_implications":["Secrecy and deception can be pursued in one physical-layer design: the same overlapping-band transmission that hides the true signal makes the decoy the strongest signal at the eavesdropper.","With the deception threshold in the reported range, the simulated system achieves perfect deception probability along with zero interception probability.","The proposed optimization outperforms both OFDM and equal-power allocation in sum secrecy rate and in decoy dominance at the eavesdropper.","Increasing the required decoy quality first helps secrecy by degrading the eavesdropper's channel through overlap interference, then hurts it as decoy power crowds out true-signal power.","The optimized correlation coefficients can be realized by fitting the single frequency multiplexing factor with Newton's method, so only one physical parameter needs to be set."],"supporting_citations":[{"why":"Supplies the secrecy-rate metric used in the objective and the multi-source secure-transmission setting.","marker":"[12]"},{"why":"Provides the physical-layer security formulation for vehicle-satellite links that motivates the secrecy-rate objective.","marker":"[14]"},{"why":"Gives the covert-communication baseline where a signal is hidden inside NOMA interference.","marker":"[16]"},{"why":"Defines covert air-ground communication, the anti-detection baseline that CTSF extends toward deception.","marker":"[17]"},{"why":"Combines covertness and secrecy in NOMA with an internal eavesdropper, the closest prior joint goal.","marker":"[19]"},{"why":"Demonstrates an offensive full-duplex jammer that prevents eavesdropping, a precedent for active deception.","marker":"[28]"},{"why":"Frames physical-layer covert communication and its detection limits, motivating the hiding of true signals.","marker":"[30]"},{"why":"Models simultaneous secure and covert transmission against two adversarial goals, a template for the joint secrecy-deception constraint.","marker":"[38]"}],"fun_headline_variants":["Decoys mask secrets: optimal power splits deceive and hide","T/F frequency trick: decoys dominate eavesdropper, true signals stay safe","Maximizing secrecy with decoys: non-orthogonal bands outwit interceptor","Decoy-steered secrecy: power optimization foils interception","Hide truth in decoy: frequency multiplexing anti-intercept scheme"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reformulated constraint $\\xi_n \\geq \\xi_k$ is treated as equivalent to the decoy-dominance condition $p_n|h_{e,n}|^2 \\geq p_k|h_{e,k}|^2$, which silently assumes the eavesdropper's channel gains on true and fake frequencies are comparable or cancel out; under Rician fading they generally do not, and if they differ enough the optimized power allocation can fail to make the decoy dominant.","fun_headline_variants_meta":{"raw":{"variants":["Decoys mask secrets: optimal power splits deceive and hide","T/F frequency trick: decoys dominate eavesdropper, true signals stay safe","Maximizing secrecy with decoys: non-orthogonal bands outwit interceptor","Decoy-steered secrecy: power optimization foils interception","Hide truth in decoy: frequency multiplexing anti-intercept scheme"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000842,"raw_usage":{"total_tokens":3653,"prompt_tokens":914,"completion_tokens":2739,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":530,"completion_tokens_details":{"reasoning_tokens":2646}},"tokens_in":530,"tokens_out":2739,"duration_ms":16840,"temperature":1.0,"reasoning_tokens":2646,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:57:35.258415+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a Rician-faded channel realization with $|h_{e,n}|^2 < |h_{e,k}|^2$ on the eavesdropper link, run the proposed optimization, and check whether the output satisfies $p_n|h_{e,n}|^2 \\geq p_k|h_{e,k}|^2$ whenever $\\xi_n \\geq \\xi_k$; a single realization where the optimized powers satisfy $\\xi_n \\geq \\xi_k$ but violate the original dominance inequality would show the reformulation is not equivalent and the deception guarantee does not hold.","supporting_citations":[{"cited_title":"Dt-ass isted multi-point symbiotic security in space-air-ground integ rated networks,","cited_arxiv_id":null,"evidence_quote":"Supplies the secrecy-rate metric used in the objective and the multi-source secure-transmission setting."},{"cited_title":"Uav- assisted physical layer security in multi-beam satellite- enabled vehicle communications,","cited_arxiv_id":null,"evidence_quote":"Provides the physical-layer security formulation for vehicle-satellite links that motivates the secrecy-rate objective."},{"cited_title":"Covert communication of star-ris aided noma net works,","cited_arxiv_id":null,"evidence_quote":"Gives the covert-communication baseline where a signal is hidden inside NOMA interference."},{"cited_title":"Covert communication in uav-assisted air-ground networks,","cited_arxiv_id":null,"evidence_quote":"Defines covert air-ground communication, the anti-detection baseline that CTSF extends toward deception."},{"cited_title":"Covert and secure co mmuni- cations in noma networks with internal eavesdropping,","cited_arxiv_id":null,"evidence_quote":"Combines covertness and secrecy in NOMA with an internal eavesdropper, the closest prior joint goal."},{"cited_title":"Full-duplex c onstant-envelope jamceiver and self-interference suppression by highpass fi lter: Exper- imental validation for wi-fi security,","cited_arxiv_id":null,"evidence_quote":"Demonstrates an offensive full-duplex jammer that prevents eavesdropping, a precedent for active deception."},{"cited_title":"Physical lay er covert communication in b5g wireless networks—its research, appl ications, and challenges,","cited_arxiv_id":null,"evidence_quote":"Frames physical-layer covert communication and its detection limits, motivating the hiding of true signals."},{"cited_title":"Simultaneous secure and covert transmissions a gainst two attacks under practical assumptions,","cited_arxiv_id":null,"evidence_quote":"Models simultaneous secure and covert transmission against two adversarial goals, a template for the joint secrecy-deception constraint."}],"review_version":1}