{"id":"e33bcdd3-9836-45fb-a8ee-2e81984485e2","arxiv_id":"2608.06970","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A single Rydberg-excited control qubit is detected non-destructively with 99.81% assignment fidelity by mapping its state onto a collective phase shift of eight Rydberg-dressed target atoms.","lead":"This experiment shows that a single atom's quantum state can be read out through the collective response of eight neighboring atoms, using Rydberg interactions to imprint a phase shift. The scheme detects a single Rydberg excitation without destroying it, with an assigned fidelity of 99.81 percent, which matters for mid-circuit measurement in neutral-atom quantum computers.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reported 99.81% is a positive predictive value, not a state-assignment fidelity: by the paper's own P(≥3|S)=61.5%, a below-threshold outcome would be misassigned, so balanced assignment accuracy is ~81%, not 99.81%.","rationale":"The reader's weakest assumption is that target spins act as independent Bernoulli trials, so the 99.81% fidelity rests on a binomial model. That is a real concern, but I find a more direct and internal problem: the headline fidelity is not a two-outcome state-assignment fidelity as the abstract states. The SI itself defines the second-protocol fidelity as P(S_c|≥3 cts), a positive predictive value, and reports P(≥3 cts|S_c)≈61.5%. Consequently, if the protocol is used to assign the control qubit's state, roughly 38.5% of Rydberg-prepared shots fall below threshold and would be assigned to the ground state. The numbers in the paper therefore imply an average assignment fidelity of about 81% for equal priors, not 99.81%. This is not a speculative external failure mode; it follows from the paper's own quoted probabilities. It is also more load-bearing than the independence issue because it affects the central claim even if all target atoms are perfectly independent. The concrete test is straightforward: compute the full confusion matrix from the SI values, or re-analyze the 26 shots at the shot level to see how the confidence intervals change. Both would settle whether the headline number denotes what the abstract claims. Because the underlying mechanism and experimental demonstration remain sound, the appropriate verdict is still conditional—the paper should be accepted only after the fidelity claim is reframed or corrected. This does not change the reader's verdict, hence UNCHANGED, but it sharpens the condition.","tokens_in":14892,"tokens_out":11566,"duration_ms":118485,"concrete_test":"Reconstruct the full 2×2 confusion matrix for the θ_R=3π/2 detection from the SI numbers with equal priors: TPR=P(≥3|S)=0.615, FPR=P(≥3|g)=0.0012, so average assignment fidelity=(0.9988+0.615)/2≈0.807 and Youden's J≈0.614. If these values are not what the authors mean by F, restate the abstract as '99.81% precision for a positive detection event, with 61.5% detection efficiency.' Separately, re-analyze the 26 shots at the shot level rather than as 208 atom-level Bernoulli trials, computing a Wilson interval for P(≥3|g); with 0 observed events in 26 shots, the 95% upper bound is ~11%, which would lower P(S_c|≥3 cts) to ~85% even with TPR=61.5%, showing the atom-level binomial extrapolation is load-bearing.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is not the quantity the text calls 'state-assignment fidelity.' In the SI, 'Deriving and estimating the detection fidelities F', the θ_R=3π/2 protocol defines F via Bayes as P(S_c | ≥3 cts), using P(≥3 cts | g_c)≈0.12% and P(≥3 cts | S_c)≈61.5%. If this protocol is used to assign the control-qubit state, the complementary outcome (<3 counts) is naturally assigned to the ground state. The full confusion matrix then gives P(g_c|g_c)≈99.88% and P(S_c|S_c)≈61.5%, so for equal priors the average state-assignment fidelity is ≈80.7%, not 99.81%. Using the paper's own Youden's J definition from Eq. S3 gives J≈61.4%, not 99.81%. The quoted 99.81% is the positive predictive value of a threshold crossing; it does not quantify the 38.5% of Rydberg-prepared shots that fall below threshold and would be misassigned as ground. For mid-circuit readout, detection efficiency is a central performance metric, not a minor statistical detail. This internal inconsistency directly affects the abstract's 'state-assignment fidelity' claim and should be resolved before the headline number is used.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an experimental scheme in which a single control qubit, encoded in a ground-to-Rydberg transition, imprints a collective phase onto a Rydberg-dressed ensemble of target atoms in a 3x3 optical-lattice array. The central mechanism is a state-dependent AC-Stark shift: when the control is in the Rydberg state, strong control-target interactions suppress the target dressing shift, changing the Ramsey fringe contrast. The authors measure the distance dependence of this shift, show that the per-target phase is nearly independent of target number by comparing N=8 and N=4 ensembles, and demonstrate coherent mapping of control-spin Rabi oscillations onto the ensemble. They then use the collective response as a non-destructive Rydberg-state detector, reporting a state-assignment fidelity of F = 99.81^{+0.17}_{-1.47}% based on a threshold of at least three target spins in the θ_R=3π/2 protocol. The supplementary material provides derivations of the interaction model and the statistical estimate.","tokens_in":15140,"tokens_out":4114,"duration_ms":46252,"significance":"If the claims hold, the work introduces a useful loss-resilient interface between a single Rydberg qubit and a mesoscopic spin ensemble, with potential applications in mid-circuit readout and quantum simulation. The experimental evidence for the mechanism is substantial: the distance-dependent phase shift follows the calculated pair potentials, the Rabi oscillation frequency seen on the target ensemble (2π × 4.48(5) MHz) matches the control-spin Rabi frequency (2π × 4.45(6) MHz), and the N=8 versus N=4 comparison supports the claimed insensitivity to target-atom number. The paper also provides a detailed SI with the interaction model, loss measurements, and statistical analysis. However, the headline fidelity number is mislabeled: the reported 99.81% is a positive predictive value conditioned on a threshold crossing, not a state-assignment fidelity in the usual balanced sense, and the paper's own numbers imply a substantially lower balanced assignment accuracy. This issue is load-bearing because the abstract and main text advertise the result as a state-assignment fidelity for non-destructive detection.","major_comments":[{"comment":"The quantity F = 99.81^{+0.17}_{-1.47}% is not a state-assignment fidelity. It is computed as P(S_c | ≥3 counts), the positive predictive value of crossing the threshold. By the paper's own numbers, P(≥3 | S_c) ≈ 61.5%, so 38.5% of Rydberg-prepared shots fall below threshold. If the complementary outcome (<3 counts) is assigned to the ground state, the full confusion matrix gives P(g_c | g_c) ≈ 99.88% and P(S_c | S_c) ≈ 61.5%, yielding a balanced assignment accuracy of roughly 80.7% for equal priors. The Youden's J statistic defined in Eq. S3 would be approximately 61.4%, not 99.81%. The manuscript should either report the 99.81% as the positive predictive value / conditional detection probability, or present the full confusion matrix and report the balanced state-assignment fidelity. The abstract's 'state-assignment fidelity' claim should be corrected accordingly.","section":"Main text, Fig. 4; SI, 'Deriving and estimating the detection fidelities F'"},{"comment":"The low sensitivity P(≥3 | S_c) ≈ 61.5% is a central performance limitation for a readout protocol and cannot be hidden by quoting only the positive predictive value. For mid-circuit readout, missed detections (38.5% of true Rydberg events) are as important as false positives. The manuscript should explicitly discuss this detection efficiency, including how it would affect repeated or cascaded readout, and should report the full receiver-operating-characteristic information, not only the threshold-crossing conditioned probability.","section":"SI, Eq. S5 and surrounding text"},{"comment":"The statistical model underlying the 99.81% claim treats each target spin as an independent Bernoulli trial, and the Wilson interval in Eq. S4 sums over 'up to 208 independent events.' However, the SI also reports that loss in the Rydberg-control case is enhanced (0.07(1) atoms/µs versus 0.03(2) atoms/µs), and state-preparation errors of the control are global rather than per-target. These correlated error sources could violate the independence assumption in a way not captured by the binomial model. The manuscript should justify or test the per-target independence more directly, or quantify how correlated loss and control preparation errors affect the confidence interval and the fidelity number.","section":"SI, 'Deriving and estimating the detection fidelities F'; 'Atom loss from off-resonant coupling'"},{"comment":"The measured AC-Stark shift, δ_g = 2π × 578(4) kHz, differs from the theoretical estimate δ_g^theory = 2π × 670 kHz by about 14%. This is described as 'reasonable agreement,' but the discrepancy is comparable to the systematic uncertainties in the dressing parameters. Since the quantitative phase-imprinting model is the basis of the protocol, the manuscript should provide a more detailed account of this mismatch, for example by including calibration of the dressing Rabi frequencies and detunings, or by identifying the missing correction terms.","section":"Main text, Fig. 2c"}],"minor_comments":[{"comment":"The caption states that θ_R = 3π/2 'inverts the detection outcome compared to θ_R = π/2'; the relationship between the two histograms, especially why the θ_R = 3π/2 panel creates a background-free region, could be stated more explicitly.","section":"Fig. 4 caption"},{"comment":"The Wilson score interval is quoted as P(≥3 cts | g_c) ≈ 0.12^{+0.91}_{-0.11}% without giving the observed counts and total n; stating n = 208 and the number of observed threshold-crossing events would make the interval transparent.","section":"SI, Eq. S4"},{"comment":"The entry P(S_c | S_c) ≲ 99% for the θ_R = π/2 scheme is described as 'conservatively estimated' without a derivation; because the control Rydberg preparation cannot be verified on a single shot, the basis for this estimate should be specified.","section":"SI, 'Deriving and estimating the detection fidelities F'"},{"comment":"The loss rates 0.03(2) and 0.07(1) atoms/µs are reported without stating the pulse duration or the number of repetitions used for the loss measurement; adding these details would improve reproducibility.","section":"Fig. S3"}],"recommendation":"major_revision","confidential_remarks":"The experimental work appears technically sound and the mechanism is well supported by distance-dependence and atom-number-independence data. The main obstacle is the mislabeled fidelity metric: the headline 99.81% is a conditional positive predictive value, and the paper's own SI numbers imply a balanced assignment fidelity near 81% (and a Youden's J near 61%). This is fixable without new experiments by re-reporting the metric and being explicit about detection efficiency. If the authors are unwilling to correct the abstract and headline claim, the paper would not be acceptable; with that correction, it is a strong candidate for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the real content. The scheme couples a single Rydberg control to a dressed target ensemble and imprints a phase per target that is independent of target number. That is a genuinely new combination, and the data back it: the distance dependence tracks the calculated pair potentials, the N=8 vs N=4 comparison shows the phase per atom is unchanged, and the target ensemble follows the control Rabi oscillation at the expected frequency. The theoretical model in the SI is a standard AC-Stark calculation with no circular fitting to the headline number. This is a well-executed experiment and a useful building block.\n\nNow the problem. The 99.81% fidelity claim does not mean what it says. The SI defines it as P(S_c | ≥3 counts), the probability that the control was Rydberg given that a threshold was crossed. The complement—the probability that a Rydberg-prepared control produces a below-threshold count and gets assigned to ground—is P(<3 | S_c) ≈ 38.5%. So a single-shot assignment of the control state, with equal priors, is correct only about 81% of the time: ~99.9% for ground, ~61.5% for Rydberg. That is the number that matters for mid-circuit readout. The abstract calls it 'state-assignment fidelity,' which is wrong. Youden's J, the paper's own measure for the θ_R=π/2 scheme, gives about 61% for the θ_R=3π/2 scheme. The high 99.81% is the purity of the click, not the accuracy of the assignment. The information is in the SI, so it is a framing error, not a hidden one—but it is a significant one, and the summary of the paper should not be a number that overstates recall by ~38 points.\n\nOther issues are minor by comparison. The binomial estimate comes from 26 shots, and the lower bound on the per-target fidelity is around 94%. The measured AC-Stark shift is 578 kHz against an ab initio 670 kHz, a 14% gap that is unexplained. Target loss is higher when the control is in the Rydberg state, which could introduce correlations not captured by the independent-Bernoulli model. None of these undercut the mechanism, but they should be addressed.\n\nMy recommendation: send it for peer review. The mechanism is novel, the experiment is careful, and the critique is fixable. Ask the authors to recompute and relabel the fidelity—report the confusion matrix and the recall, or at least state that 99.81% is the conditional detection purity, not the per-shot assignment accuracy. Until then, don't quote the 99.81% number without the caveat.","headline":"A genuinely new collective phase-imprinting readout for Rydberg states with solid experimental backing, but the 99.81% 'state-assignment fidelity' is really a positive predictive value; per-shot assignment accuracy is about 81%.","tokens_in":15750,"tokens_out":5263,"would_cite":true,"duration_ms":50555,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper shows that a single Rydberg-encoded qubit can be detected without destroying it by imprinting a collective phase on a Rydberg-dressed neighboring ensemble, with per-atom phase independent of ensemble size and a demonstrated…","keywords":["Rydberg dressing","collective phase","single-qubit detection","non-destructive readout","optical lattice","Ramsey interferometry","atom-number robustness","Rydberg qubit"],"falsifier":"Measure the per-target Ramsey oscillation frequency for $N = 8$ and for larger ensembles at fixed nearest-neighbor spacing and identical dressing parameters. The paper reports essentially equal rates for $N=8$ and $N=4$; a statistically significant drift of $\\delta_g$ with $N$, or a histogram of detected $|0\\rangle$ atoms showing multi-atom correlations that violate independent Bernoulli trials, would falsify the robustness claim.","tokens_in":14650,"feed_emoji":"⚛️","tokens_out":9314,"duration_ms":87080,"temperature":0.7,"pith_summary":"The paper establishes a method to read out the state of a single qubit by mapping it onto a small ensemble of neighboring atoms in an optical lattice. A control atom in a Rydberg state suppresses an AC-Stark shift that its ground-state counterpart would otherwise impose on each target atom, so a Ramsey measurement on the targets reveals which control state was present. Because the target atoms are Rydberg-dressed, the phase imprinted per atom is insensitive to how many target atoms are there, making the scheme robust to atom loss and number fluctuations. The experiment demonstrates detection of a single Rydberg excitation with state-assignment fidelity $\\mathcal{F} = 99.81^{+0.17}_{-1.47}\\%$, which would offer non-destructive mid-circuit readout for Rydberg-atom quantum processors.","feed_headline":"Single Rydberg qubit read out with 99.81% fidelity","feed_subtitle":"A control atom's state is mapped onto an eight-atom ensemble whose response is robust to atom loss.","key_machinery":"The load-bearing mechanism is a qubit-controlled collective phase: each of $N$ Rydberg-dressed target atoms acquires the same phase $\\phi = \\delta_g t_T$ when the control qubit is in its ground state and essentially no phase when the control is in its Rydberg state. The per-atom dressing shift $\\delta_g = \\Omega_P^2/4\\Delta$ is produced by off-resonant coupling of the target to a Rydberg $|P\\rangle$ state, with admixture $\\beta = \\Omega_P/2\\Delta$; fourth-order interactions among targets scale as $\\beta^4\\Delta$ and are suppressed, which is what makes the imprinted phase per target independent of $N$. A Ramsey interferometer converts the collective phase into a measurable population difference, and a final $3\\pi/2$ pulse makes the detection background-free by mapping preparation errors onto the same low-count events as the ground-state signal.","core_discovery":"The central claim is that a control qubit encoded in the ground-to-Rydberg transition can be coherently mapped onto a mesoscopic target ensemble by a qubit-controlled collective phase. With the control in $|g\\rangle$, each dressed target spin in $|0\\rangle$ accumulates a phase $\\phi = \\delta_g t_T$ with $\\delta_g = \\Omega_P^2/4\\Delta$; with the control in the Rydberg state $|S\\rangle$, the control-target interaction shifts the dressing transition out of resonance and reduces the shift to $\\delta_S \\ll \\delta_g$. A Ramsey sequence converts the accumulated phase into a population difference, and a threshold on the number of target atoms found in $|0\\rangle$ assigns the control state. The per-atom phase is independent of the number $N$ of target atoms because residual intra-ensemble Rydberg interactions scale as $\\beta^4\\Delta$ at small dressing admixture $\\beta = \\Omega_P/2\\Delta$, so the collective signal grows with $\\sqrt{N}$ while remaining insensitive to atom loss. The demonstrated single-Rydberg-excitation detection fidelity is $\\mathcal{F} = 99.81^{+0.17}_{-1.47}\\%$ with a threshold of at least three detected target spins.","pith_inferences":["Beyond the eight-atom demonstration, the same threshold logic should work with a smaller ensemble at lower fidelity but with much simpler optical access; the scheme is effectively a deterministic, tunable amplifier for a single qubit's state.","An implicit consequence is that the protocol could be used for repeated interrogation: as long as the target ensemble can be re-prepared or refreshed, the control qubit's state can be checked multiple times without directly measuring it, which would suit feed-forward quantum computation.","A testable extension for other platforms is to replace the final push-out readout with a cavity-enhanced or non-destructive collective measurement; the binomial fidelity model in the paper predicts that the detection threshold should then shift with the improved readout noise."],"forward_implications":["Non-destructive mid-circuit readout of Rydberg excitations in neutral-atom quantum simulators becomes practical without requiring a second atomic species.","Adding more target atoms increases the signal-to-noise ratio as $\\sqrt{N}$ while leaving the per-atom phase unchanged, so detection fidelity can be improved simply by enlarging the ensemble.","The coherent control-ensemble coupling can in principle prepare entangled states such as $(|g\\rangle_C|0\\rangle_T^{\\otimes N} + |S\\rangle_C|1\\rangle_T^{\\otimes N})/\\sqrt{2}$, subject to a readout that preserves both subsystems.","Because number fluctuations do not alter the accumulated phase per atom, the interaction is a candidate building block for loss-tolerant multi-qubit phase gates and for faster Rydberg detection through increased photon-collection rates from the ensemble."],"supporting_citations":[{"why":"The supplementary material containing the pair-wise interaction derivation, atom-loss rates, and the full fidelity estimation that yields $\\mathcal{F} = 99.81\\%$.","marker":"[26]"},{"why":"Supplies the single-site-addressed subwavelength atomic array and the Rydberg-switched array platform on which the protocol is built.","marker":"[18]"},{"why":"Provides the dual-species ensemble-readout and multiqubit-gate proposals that motivate the approach and that the present single-species scheme is compared against.","marker":"[23]"},{"why":"Gives the Raman driving scheme and the perturbative AC-Stark shift formula used to compute the dressing phase and its higher-order corrections.","marker":"[30]"},{"why":"Supplies the Rydberg interaction potentials used for the control-target shift and its distance dependence.","marker":"[44]"},{"why":"Supports the physics of Rydberg-dressed spin lattices, including suppression of higher-order dressing interactions and the enhanced-loss channels cited in the supplementary analysis.","marker":"[34]"},{"why":"Provides the diagnostic-test index used to define the state-assignment fidelity in the first detection sequence.","marker":"[47]"},{"why":"Provides the binomial confidence-interval formula that bounds the per-target fidelity used in the final fidelity estimate.","marker":"[48]"}],"fun_headline_variants":["Collective phase turns one atom into a full ensemble readout","Qubit state mapped onto 8-atom ensemble with 99.81% fidelity","Phase imprinting reads out Rydberg qubit without destroying it","One atom's state read by eight friends at 99.81% accuracy","Mesoscopic ensemble detects single qubit with loss-proof phase"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The target atoms must behave as independent spectators during the dressing pulse, so each one accumulates the same phase no matter how many neighbors are present; if correlated interactions or correlated losses among the targets become significant, the per-atom phase and the binomial fidelity estimate break down.","fun_headline_variants_meta":{"raw":{"variants":["Collective phase turns one atom into a full ensemble readout","Qubit state mapped onto 8-atom ensemble with 99.81% fidelity","Phase imprinting reads out Rydberg qubit without destroying it","One atom's state read by eight friends at 99.81% accuracy","Mesoscopic ensemble detects single qubit with loss-proof phase"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000557,"raw_usage":{"total_tokens":2721,"prompt_tokens":1085,"completion_tokens":1636,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":701,"completion_tokens_details":{"reasoning_tokens":1541}},"tokens_in":701,"tokens_out":1636,"duration_ms":11644,"temperature":1.0,"reasoning_tokens":1541,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T17:22:44.453915+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the per-target Ramsey oscillation frequency for $N = 8$ and for larger ensembles at fixed nearest-neighbor spacing and identical dressing parameters. The paper reports essentially equal rates for $N=8$ and $N=4$; a statistically significant drift of $\\delta_g$ with $N$, or a histogram of detected $|0\\rangle$ atoms showing multi-atom correlations that violate independent Bernoulli trials, would falsify the robustness claim.","supporting_citations":[{"cited_title":"Srakaew, P","cited_arxiv_id":null,"evidence_quote":"Supplies the single-site-addressed subwavelength atomic array and the Rydberg-switched array platform on which the protocol is built."},{"cited_title":"Petrosyan, S","cited_arxiv_id":null,"evidence_quote":"Provides the dual-species ensemble-readout and multiqubit-gate proposals that motivate the approach and that the present single-species scheme is compared against."},{"cited_title":"Srakaew, Rydberg interactions in subwavelength atomic arrays and Hubbard systems, Doctoral disserta- tion, Ludwig-Maximilians-Universit¨ at M¨ unchen (2024)","cited_arxiv_id":null,"evidence_quote":"Gives the Raman driving scheme and the perturbative AC-Stark shift formula used to compute the dressing phase and its higher-order corrections."},{"cited_title":"Weber, C","cited_arxiv_id":null,"evidence_quote":"Supplies the Rydberg interaction potentials used for the control-target shift and its distance dependence."},{"cited_title":"Zeiher, R","cited_arxiv_id":null,"evidence_quote":"Supports the physics of Rydberg-dressed spin lattices, including suppression of higher-order dressing interactions and the enhanced-loss channels cited in the supplementary analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the diagnostic-test index used to define the state-assignment fidelity in the first detection sequence."},{"cited_title":"pair- interaction","cited_arxiv_id":null,"evidence_quote":"Provides the binomial confidence-interval formula that bounds the per-target fidelity used in the final fidelity estimate."}],"review_version":1}