{"id":"6ebc0fb8-090d-4b66-b516-46df54d4f6be","arxiv_id":"2412.14733","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A superconducting qutrit-resonator system experimentally realizes eigenvalue braiding around a third-order exceptional point using three control parameters, plus chiral state transfer when no exceptional point is encircled.","lead":"This paper reports experiments on a superconducting circuit that realizes a three-level non-Hermitian system, showing how its eigenvalues braid as control parameters loop around a third-order exceptional point. The work demonstrates that only three tunable parameters are needed to generate the full braid group B3, and that chiral state transfer can occur without encircling an exceptional point.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Braid-group claim depends on exact zero coupling between |e,0> and |g,1>; residual coupling would alter exceptional arcs and braid labels.","rationale":"The reader's weakest assumption identifies the same load-bearing point: the pseudo-chirality symmetry at δef=0 is exact only if the Hamiltonian has the precise form of Eq. (1), with no coupling between |e,0> and |g,1>. This symmetry is the enabling condition for reducing the parameter space from four to three dimensions while still realizing B3. The physical device, however, has a finite transmon-resonator coupling that can generate a residual η, and the manuscript neither bounds η nor analyzes how calibration errors or higher-order effects shift the exceptional arcs. Since the experiment's eigenvalue data are not measured directly but are recovered by fitting population dynamics to the very same Hamiltonian that assumes η=0, the braid labels in Fig. 3 are not independently verified. A concrete re-analysis with η as a free parameter would settle whether the symmetry assumption is controlled. My concern does not change the reader's conditional verdict; it sharpens the condition that must be met for the central claim to hold.","tokens_in":11654,"tokens_out":13299,"duration_ms":108408,"concrete_test":"Re-fit the raw time-domain population data used for Fig. 2c and for the red/blue loops in Fig. 3 with an extended Hamiltonian H = Eq. (1) + η(|e,0><g,1| + |g,1><e,0|), treating η (and optionally a small δef offset) as a free parameter. Report the best-fit η and its confidence interval, and compute the discriminant locus for that η range. If the best-fit η is consistent with zero and the homotopy class of the red and blue loops (i.e., which exceptional arcs they encircle) is unchanged for all η within the 95% confidence interval, the three-parameter B3 claim survives; if η is poorly constrained or the loop labels change, the central claim is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that three independent parameters suffice to realize the entire braid group B3 relies on the pseudo-chirality symmetry of the Hamiltonian in Eq. (1), which holds only at δef = 0 and only if the coupling between |e,0> and |g,1> is exactly zero. In the physical device, the transmon-resonator coupling J = 48 MHz can induce a small off-resonant coupling η between |e,0> and |g,1>; the paper does not report any bound on η. If η is nonzero, the discriminant locus changes: the two EP2 arcs that meet at the EP3 cusp in Fig. 2a generically separate or form a closed loop, so the loops labeled σ1 and σ2 in Fig. 3 may no longer encircle distinct arcs. Because all eigenvalues in Fig. 3 are retrieved by fitting the population dynamics to the same model (Eq. (1)) that assumes η = 0, the reported braids could be an artifact of the fit. The braid-group generation, the Yang-Baxter relation, and the 'three parameters are sufficient' assertion would fail if η exceeds a small fraction of κ, and no error analysis connects calibration uncertainties to the homotopy class of the loops.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an experimental study of a three-level non-Hermitian system realized in a superconducting transmon-resonator device. The central claims are (i) the observation of a third-order exceptional point (EP3) whose surrounding parameter-space topology yields eigenvalue braiding described by the full braid group B3 using only three independent control parameters, and (ii) the observation of chiral quantum state transfer when dynamical loops enclose no exceptional points, with time-symmetric behavior when EPs are enclosed. The theoretical model is the three-level Hamiltonian in Eq. (1), whose pseudo-chirality symmetry at δef=0 is invoked to reduce the parameter-space dimension. The experimental evidence for the EP structure is based on fitting measured population dynamics to this model and then computing eigenvalues from the fitted Hamiltonian, while the state-transfer results are direct overlap measurements compared with numerical simulations.","tokens_in":11913,"tokens_out":5368,"duration_ms":47668,"significance":"If fully supported, the braid-group result would be a notable advance in quantum simulation of non-Hermitian topology, being the first demonstration of non-Abelian B3 braiding near an EP3 in a solid-state quantum system and showing that symmetry can reduce the number of tuning parameters from four to three. The theoretical analysis, including the Cardano-form discriminant and the pseudo-chirality/anti-PT symmetry arguments, is clean and explicit. The state-overlap measurements in Fig. 4 are a genuine experimental result and match the numerics well. However, the eigenvalue-braid data are not directly measured, and the central claim therefore rests on a model-dependent extraction that needs to be made falsifiable. The paper lacks a quantitative sensitivity analysis of the zero-coupling assumption on which the symmetry and the braid labels rely.","major_comments":[{"comment":"The eigenvalue braids shown in Fig. 3 are not measured directly. The paper's procedure (described for the EP characterization and presumably used for the braid data) is to fit the measured population dynamics to the model Hamiltonian in Eq. (1), extract parameters (δ_ef, Ω, G), and then compute eigenvalues from that same Hamiltonian. The 'experimental data' in Fig. 3 therefore reduce to a re-evaluation of the theoretical model with fitted parameters; they cannot independently confirm the existence of the EP2 arcs or the braid group B3. To make the central claim falsifiable, the authors should either extract the complex eigenvalues directly from the time-resolved complex amplitudes (including phases) without assuming Eq. (1), or, at minimum, perform a parameter-uncertainty analysis showing that the homotopy class of each loop is stable under the calibration and fitting errors. Without such an analysis, the statement 'we experimentally investigate the eigenvalues braiding' is not supported by the data as presented.","section":"Results, 'Complex eigenvalues braiding' (Fig. 3); also Results, EP characterization (Fig. 2)"},{"comment":"The pseudo-chirality symmetry that enables the EP3 and the reduction to three parameters holds only for the exact Hamiltonian of Eq. (1), in which the coupling between |e,0> and |g,1> is exactly zero. In the physical device, the transmon-resonator coupling J/2π=48 MHz is always present; although the direct coupling is far off-resonant (resonator at 6.697 GHz versus transmon at 5.684 GHz), it will generate a small but finite effective coupling η between |e,0> and |g,1> through higher-order processes. The paper neither estimates η nor provides an upper bound. A nonzero η generically deforms the discriminant locus: the two EP2 arcs in Fig. 2a may no longer meet at the EP3 cusp, and the loops labeled σ1 and σ2 in Fig. 3 may encircle different arcs or none at all. The authors should quantify η for their device parameters and demonstrate that the braid-group claim and the specific braid labels are robust to this coupling and to the calibration uncertainties in δ_ef, Ω, and G. This is a load-bearing assumption that currently lacks any error analysis.","section":"Methods, 'Symmetries of non-Hermitian Hamiltonian'; Eq. (1)"}],"minor_comments":[{"comment":"The text contains a typo: 'ac Start shift' should read 'ac Stark shift'.","section":"Results, EP characterization paragraph after Eq. (1)"},{"comment":"The sign convention for δef and the precise definition of the rotating frame are not fully specified; please define the frame and list all terms (including the ac Stark compensation) so that Eq. (1) can be reproduced unambiguously.","section":"Eq. (1) and the preceding paragraph"},{"comment":"Two exceptional lines with Ω=0, G=±κ/4 are mentioned as being omitted because 'they do not impact our results'; a one-sentence justification of why they do not affect the homotopy classes of the shown loops would improve clarity.","section":"Fig. 3a caption and discussion"},{"comment":"The phrase 'entire braid group B3' is stronger than what the experiment verifies; the paper demonstrates generators σ1 and σ2 and two concatenations. Please qualify the claim as 'generated by the demonstrated loops in the model' until an independent measurement of the braids is provided.","section":"Abstract and Conclusion"},{"comment":"The statement that the state transfer vanishes when the starting point is set at (Ω_m,0) is presented without an experimental data point; please indicate whether this is a numerical prediction or an experimental observation.","section":"Conclusion, last sentence before acknowledgements"},{"comment":"The claim that the standard error of the mean is smaller than the plotted points is not supported by any quantitative number; please provide the actual error values for the overlap measurements in Fig. 4 and for the fitted eigenvalues in Figs. 2–3.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is relevant to the quantum non-Hermitian physics community and the theoretical model is clean, but the central experimental claim of eigenvalue braiding needs to be made directly falsifiable. The model-dependent retrieval of eigenvalues is a serious circularity concern that should be resolved before publication. The state-transfer measurement is the strongest experimental part and may be sufficient to support a revised, more carefully worded version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a legitimate experimental advance—EP3 in a driven superconducting qutrit, braid-group B3 operations via three parameters, and chiral state transfer—and the data support the central narrative. The main caveat is that the eigenvalue braids are not directly measured; they come from fitting the population dynamics to the same 3x3 Hamiltonian and then diagonalizing it. That makes the braid plots model-dependent, and the paper gives no error bars that propagate calibration uncertainties into the homotopy class of the loops.\n\nWhat's actually new: prior EP3 experiments were classical (Patil et al., Tang et al.) and prior quantum EP work was mostly EP2 (Liu et al., Abbasi et al.). This paper does EP3 braiding and the symmetry-driven reduction from four to three parameters in a quantum circuit, plus a clear demonstration that loops outside EP2 arcs give chiral transfer and loops enclosing them give time-symmetric transfer. The theoretical scaffold—pseudo-chirality at delta_ef=0, discriminant analysis, braid concatenation, Yang-Baxter identity—is standard but correctly handled. The population dynamics in Fig. 4 match simulation well, and the methods for state preparation and tomography are plausible.\n\nSoft spots, in order. First, model-dependent eigenvalue extraction. Fig. 3 \"experimental data\" are eigenvalues of Eq. (1) computed with parameters fitted to the populations. This is an indirect measurement, not an independent spectral probe. It is not fatal—fitting all three populations is a reasonable way to infer the effective Hamiltonian—but it means the braid plots inherit every assumption of Eq. (1). The paper should say this explicitly and quantify fit uncertainties. Second, the symmetry-breaking coupling. The entire \"three parameters suffice\" argument sits on the absence of a direct |e,0>-|g,1> coupling at delta_ef=0. The manuscript mentions this and gives an argument about phase invariance, but no bound on residual off-resonant coupling is reported. If eta were non-negligible relative to kappa, the EP2 arcs would separate and the loop homotopy could change. This may be negligible in the actual device given the ~1 GHz detuning, but the authors should put a number on it. Third, no data or code release. For a claim that depends on fitted parameters and simulation comparisons, releasing raw populations and fit procedures would let a referee check the braid inference.\n\nThe citation pattern looks fine; the relevant prior EP3 and EP2 experiments are cited. This deserves a serious referee. I'd send it out, but the referee should be asked to focus on the model-dependence of the braid extraction and on the residual-coupling question. If those are answered, the paper is a solid PRL/PRX-type result. Readership: non-Hermitian topology and circuit-QED quantum control.","headline":"A real first: EP3 braid-group control in a superconducting qutrit, but the braid plots are inferred from a fitted model rather than measured spectrally, so the referee should focus on that model-dependence.","tokens_in":12409,"tokens_out":3646,"would_cite":true,"duration_ms":29149,"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 claims that a three-dimensional non-Hermitian superconducting system can realize the full braid group $B_3$ of eigenvalues with only three independent parameters, and that fast driving near the third-order exceptional point…","keywords":["third-order exceptional point","braid group","eigenvalue braiding","non-Hermitian quantum system","superconducting circuit","chiral state transfer","anti-PT symmetry","exceptional point encirclement"],"falsifier":"Run the concatenated loops $\\sigma_1\\sigma_2$ and $\\sigma_2\\sigma_1$ of Fig. 3e-f with the same initial eigenstate and reconstruct the eigenvalue braids from the measured time traces; if the two resulting braids are topologically equivalent, the loop space does not generate the non-Abelian group $B_3$ and the central claim fails.","tokens_in":11487,"feed_emoji":"🌀","tokens_out":7274,"duration_ms":62525,"temperature":0.7,"pith_summary":"The paper reports a superconducting-circuit experiment in which a three-level non-Hermitian quantum system is controllably steered around a third-order exceptional point, a degeneracy where three eigenvalues and their eigenstates coalesce. It claims that, thanks to a pseudo-chirality symmetry of the Hamiltonian at zero detuning, three independent control parameters are sufficient to realize the full braid group $B_3$ of the three complex eigenvalues, instead of the generic four parameters. The authors verify the braid generators by quasistatic loops, show that $\\sigma_1\\sigma_2 \\neq \\sigma_2\\sigma_1$, and that the Yang-Baxter identity $\\sigma_1\\sigma_2\\sigma_1 = \\sigma_2\\sigma_1\\sigma_2$ holds, and then observe chiral state transfer under fast driving when no EP is enclosed and time-symmetric transfer when one is. The result matters because it lowers the experimental cost of accessing high-order exceptional-point topology and demonstrates directional quantum-state manipulation in a dissipative quantum processor.","feed_headline":"Three parameters are enough to braid eigenvalues around an EP3","feed_subtitle":"A superconducting qutrit experiment maps parameter-space loops to the non-Abelian braid group B3.","key_machinery":"The load-bearing object is the $3\\times 3$ non-Hermitian Hamiltonian in Eq. (1), whose entries are the detuning $\\delta_{\\rm ef}$, the Rabi amplitude $\\Omega$, the cavity-assisted coupling $G$, and the photon-loss rate $\\kappa$. At $\\delta_{\\rm ef}=0$ the Hamiltonian obeys pseudo-chirality and anti-PT symmetry, which reduces the number of independent parameters needed to reach the EP3 from four to three. The paper tracks the discriminant $\\Delta$ of the cubic characteristic polynomial: $\\Delta=0$ gives the exceptional arcs of EP2s, and the additional condition $p=0$ locates the EP3 at their cusp. The braid group $B_3$ enters through the fundamental group of the eigenvalue-space complement, with each control loop mapping to a permutation of the three eigenvalue strands; this mapping is what lets the authors generate braid words by concatenating loops.","core_discovery":"The central claim is that the Hamiltonian of Eq. (1), built from a transmon qutrit coupled to a lossy resonator, hosts an EP3 at $\\delta_{\\rm ef}=0$ where all three eigenvalues coalesce, and that the exceptional structure in the three-parameter space $(\\delta_{\\rm ef}, \\Omega, G)$ is rich enough to generate the entire braid group $B_3$. Quasistatic loops around different exceptional arcs produce the generators $\\sigma_1$ and $\\sigma_2$; concatenating loops yields $\\sigma_1^2$, $\\sigma_2\\sigma_1$, and $\\sigma_1\\sigma_2$, with the non-commutativity of $\\sigma_1$ and $\\sigma_2$ directly measured. The same system, driven rapidly along rectangular loops in a plane of fixed $G$, shows directional (chiral) state transfer between the right eigenstates only when the loop encloses no EP, while loops enclosing one or two EPs exhibit time-symmetric transfer. The authors support these claims with eigenvalue coalescence data, retrieved eigenvalues along the control loops, and state-overlap measurements from the superconducting processor.","pith_inferences":["One natural extension beyond the paper would be a four-level symmetry-protected system, where pseudo-chirality might realize a subset of $B_4$ with fewer than the generic six parameters; the paper does not address this case.","Because the braid group is non-Abelian, concatenated loops could in principle implement braid-based operations on states encoded in the eigenstate manifold, but the authors stop at demonstrating the braids.","The direction-dependent chiral transfer without EP encirclement may serve as a fast dissipative state-preparation tool, a use the paper leaves implicit.","A time-domain test comparing final states after $\\sigma_1\\sigma_2\\sigma_1$ and $\\sigma_2\\sigma_1\\sigma_2$ would dynamically confirm the Yang-Baxter identity, going beyond the paper's topological demonstration."],"forward_implications":["Full $B_3$ braiding of a three-level non-Hermitian system becomes accessible with only three tunable parameters, eliminating the generic four-parameter requirement whenever the pseudo-chirality symmetry is present.","Concatenating control loops supplies a physical implementation of the braid generators and of the Yang-Baxter relation $\\sigma_1\\sigma_2\\sigma_1=\\sigma_2\\sigma_1\\sigma_2$, tying circuit control to non-Abelian topology.","Chiral state transfer without an enclosed exceptional point shows that fast loops near, but not around, an EP can select the final eigenstate by direction and starting point.","Loops that enclose one or two EPs suppress the chirality, confirming that the number of enclosed EPs, not just their proximity, determines the time asymmetry.","The same platform can realize higher-order EP physics with engineered dissipation, since the resonator loss provides a tunable non-Hermitian channel."],"supporting_citations":[{"why":"establishes the standard result that a quasistatic loop around an EP permutes the N complex eigenvalues, which the braid construction builds on.","marker":"[10, 11]"},{"why":"supplies the higher-dimensional parameter space $G_N$ whose fundamental group is the basis for the braid-group bijection.","marker":"[13]"},{"why":"provides the homotopy characterization identifying the fundamental group of the eigenvalue-space complement with the braid group.","marker":"[14]"},{"why":"lays out the generic four-parameter requirement and the trefoil-knot exceptional structure that the present three-parameter scheme bypasses.","marker":"[17]"},{"why":"shows that additional symmetries can reduce the degrees of freedom needed for high-order EPs, the route the paper exploits.","marker":"[18]"},{"why":"is the prior realization of dynamical EP encirclement in a real quantum system, providing the experimental baseline for the state-transfer part.","marker":"[32]"},{"why":"demonstrates topological quantum state control through EP proximity and chiral transfer without encirclement in a two-level system, which the paper extends to three levels.","marker":"[33]"},{"why":"gives the symmetry-protection framework for higher-order exceptional points that underlies the pseudo-chirality reduction.","marker":"[41]"}],"fun_headline_variants":["Braid eigenvalues around a third-order exceptional point","Superconducting qutrit experiments map EP3 braiding","Three parameters control eigenvalue braiding near an EP3","Chiral state transfer and braiding at a third-order EP","Topological braiding near a third-order exceptional point"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire construction assumes the system is exactly described by Eq. (1) with $\\delta_{\\rm ef}=0$ pseudo-chirality, meaning no residual coupling between $|e,0\\rangle$ and $|g,1\\rangle$ and a single loss channel on $|g,1\\rangle$; if these idealizations fail, the exceptional arcs shift and the measured braids would not be the predicted $B_3$ words.","fun_headline_variants_meta":{"raw":{"variants":["Braid eigenvalues around a third-order exceptional point","Superconducting qutrit experiments map EP3 braiding","Three parameters control eigenvalue braiding near an EP3","Chiral state transfer and braiding at a third-order EP","Topological braiding near a third-order exceptional point"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000212,"raw_usage":{"total_tokens":1447,"prompt_tokens":1003,"completion_tokens":444,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":619,"completion_tokens_details":{"reasoning_tokens":365}},"tokens_in":619,"tokens_out":444,"duration_ms":4077,"temperature":1.0,"reasoning_tokens":365,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T11:56:56.869143+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the concatenated loops $\\sigma_1\\sigma_2$ and $\\sigma_2\\sigma_1$ of Fig. 3e-f with the same initial eigenstate and reconstruct the eigenvalue braids from the measured time traces; if the two resulting braids are topologically equivalent, the loop space does not generate the non-Abelian group $B_3$ and the central claim fails.","supporting_citations":[{"cited_title":"Guria, Q","cited_arxiv_id":null,"evidence_quote":"supplies the higher-dimensional parameter space $G_N$ whose fundamental group is the basis for the braid-group bijection."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the homotopy characterization identifying the fundamental group of the eigenvalue-space complement with the braid group."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"lays out the generic four-parameter requirement and the trefoil-knot exceptional structure that the present three-parameter scheme bypasses."},{"cited_title":"Mandal and E","cited_arxiv_id":null,"evidence_quote":"gives the symmetry-protection framework for higher-order exceptional points that underlies the pseudo-chirality reduction."}],"review_version":1}