{"id":"2a515b75-4ffb-473a-929a-2eccb69963e9","arxiv_id":"2501.12163","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper derives and numerically tests semiclassical equations of motion for wave packets in non-Hermitian topological systems, where complex Berry curvature produces an anomalous force as well as an anomalous velocity.","lead":"This paper derives equations of motion for electron wave packets in non-Hermitian topological materials, where the usual Berry curvature becomes complex and adds both an anomalous velocity and an anomalous force. It tests the equations on a modified Haldane model with unbalanced next-nearest-neighbor hopping and proposes a chiral optical cavity as a way to realize the model.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Bulk wave-packet validation omits the Berry phase entirely, so the claimed numerical confirmation of the anomalous velocity and anomalous force is not actually tested.","rationale":"The reader's weakest assumption was the non-Hermitian bulk-boundary correspondence, which is a serious gap but concerns the Haldane-model application. The concern I identify is more immediately load-bearing: the bulk simulation that is claimed to confirm the central equations omits the very Berry-phase contribution that produces the new anomalous terms. Even if the bulk-boundary correspondence were proven, the abstract's numerical-confirmation claim would still be unsupported unless the Berry phase is included in the simulation and the comparison is made explicit. Conversely, if the bulk simulation with Berry phase matches Eqs. (13)-(14), that would provide genuine support for the main EOM while leaving the bulk-boundary question separate. I therefore keep the reader's CONDITIONAL verdict: the authors should add a Berry-phase-included bulk simulation (and clarify the non-Gaussian reduction from Eq. (14) to k_M(t)) before the central numerical claim is accepted.","tokens_in":17114,"tokens_out":21289,"duration_ms":245129,"concrete_test":"Re-run the bulk wave-packet simulation of Fig. 7 with the complex Berry phase included, using a numerically smooth gauge for A(k)=i<psi_L|grad_k psi_R> or a gauge-invariant Wilson-line evaluation along the accelerated trajectory k(t)=k0+F t, and compare the resulting k_M(t) and r_M(t) to Eqs. (13)-(14). If including the Berry phase shifts the trajectories by -Re(F x B) and -Delta^2 Im(F x B) as predicted, the concern is resolved; if the trajectories are unchanged or disagree, the current simulation does not validate the central EOM. As a secondary check, vary Delta to confirm the Delta^2 scaling of the k-space correction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of the paper is Eqs. (7)-(8) (and their non-Gaussian counterparts (13)-(14)): the real-space peak velocity gains -Re(F x B) and the k-space peak velocity gains -Delta^2 Im(F x B). In the Haldane-model bulk simulation of Sec. III.E and Fig. 7, however, the authors explicitly state: \"the Berry phase is not included in the simulation and theoretical calculation due to numerical challenges caused by its gauge dependence.\" Since the Berry phase is precisely the mechanism through which the F x B terms enter the derivation (via Eq. (6)), omitting it removes the only terms that distinguish the non-Hermitian topological EOM from a simple gain-gradient drift. The reported \"good agreement\" in Fig. 7(i,j) therefore confirms only the Re(dE/dk) and Delta^2 Im(dE/dk) drift terms; it provides no evidence for the anomalous velocity or anomalous force that are the paper's new topological predictions. Additionally, for the non-Gaussian case Eq. (14) is not a closed equation for k_M(t) without the Hessian of f, so the claimed comparison in the same section is under-specified. The abstract's statement that the analytical findings are \"confirmed by direct numerical simulations\" is thus unsupported for the central topological content.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"Using the biorthogonal (complex) Berry connection, the paper derives semiclassical equations of motion for the peak position and momentum of a wave packet in a non-Hermitian topological band. For a Gaussian packet, the claimed EOM are Eq. (7), ṙ_M = Re(∇_k E) − Re(F × B), and Eq. (8), ḳ_M = F + Δ²[Im(∇_k E) − Im(F × B)], where the second Δ² term is a new k-space anomalous force; a non-Gaussian extension is given by Eqs. (13)-(14). The Gaussian result is tested in a toy model with linear complex dispersion and constant complex Berry curvature (Fig. 1). The authors then introduce a non-Hermitian Haldane model with non-reciprocal next-nearest-neighbor hoppings and define a \"complex chirality\" (S_R = sign(n−m), S_I = sign(b−a)), finding C₁ = −S_I and a sign pattern of Im(B) controlled by S_R. Strip spectra show chiral edge states whose direction tracks S_I and whose gain/loss tracks S_R. Edge and bulk wave-packet simulations are reported, and an optical chiral cavity realization is proposed and mapped to the model parameters.","tokens_in":17336,"tokens_out":38836,"duration_ms":374558,"significance":"If valid, the EOM add a genuinely new term to the semiclassical transport of non-Hermitian bands: an anomalous force −Δ²Im(F × B) in k-space, alongside the anomalous velocity −Re(F × B), and the complex-chirality model provides a concrete two-parameter platform with falsifiable predictions for edge-state direction and dissipation rate. The strengths of the manuscript are the explicit derivation from stated assumptions (no fitting, no free parameters), the toy-model test of Fig. 1 that directly exercises the Berry-curvature terms, the computed (not imposed) Chern numbers, and the independent cavity parameter mapping. The main weaknesses are that the Haldane-model bulk simulation omits the Berry phase entirely, so the topological content of the EOM is not numerically confirmed in the physical model; the non-Gaussian comparison of Fig. 7 is under-specified; and the bulk-boundary correspondence linking C₁ to the edge states is assumed rather than proven for a non-reciprocal non-Hermitian system. These are gaps in verification and in claim strength rather than errors in the core derivation; closing them or qualifying the claims would make the paper a solid contribution.","major_comments":[{"comment":"The bulk wave-packet simulation of Sec. III.E (Fig. 7) omits the Berry phase by the authors' own statement (\"the Berry phase is not included in the simulation and theoretical calculation due to numerical challenges caused by its gauge dependence\"). Since the terms −Re(F × B) and −Δ²Im(F × B) in Eqs. (13)-(14) enter the equations only through the Berry phase via the identity (6), this simulation cannot test the anomalous velocity or the anomalous force; the \"good agreement\" reported in Fig. 7(i,j) validates only the gain-gradient drift terms Re(∇_k E) and Δ²Im(∇_k E). The only direct check of the Berry-curvature terms is the toy model of Fig. 1, which uses a linear dispersion and a constant, artificial connection A_k = (−Bk_y, 0). The abstract's sentence \"Our analytical findings are confirmed by direct numerical simulations\" is therefore not supported for the paper's central topological predictions; the authors should either include the Berry phase in the Haldane-model simulation (e.g., by working in a fixed smooth gauge with appropriate branch cuts) or explicitly qualify which findings are confirmed by which figure.","section":"Sec. III.E, Fig. 7, and Abstract"},{"comment":"Equation (14) is not a closed equation of motion for k_M(t): it determines the time derivative of ∇_k f at the peak, and converting it into a prediction for k_M requires either a prescription for the time evolution of f(k,t,F) or a fixed-shape assumption. The paper does not state which procedure produced the theoretical curves in Fig. 7(i,j), and since the authors themselves find that the initial Gaussian becomes anisotropic and non-Gaussian (Fig. 7(c-e)), a fixed-Gaussian assumption would contradict that observation, whereas inserting the simulated f into Eq. (14) would make the comparison tautological. Please specify the procedure used to generate the theory curves.","section":"Sec. II.B, Eq. (14)"},{"comment":"There is a sign error in the printed form of Eq. (15): with the convention |Ψ(k,t)⟩ = exp[−f(k,t,F)] of Eq. (11), the Gaussian distribution of Eq. (1) corresponds to f_Gaussian = +(k−k̄)²/(2Δ²). With the printed minus sign, exp[−f] is not normalizable, and inserting f_Gaussian into Eq. (14) yields Eq. (8) with the opposite signs on both Δ² terms, so the claimed reduction of Eq. (14) to Eq. (8) is not reproduced as written.","section":"Sec. II.B, Eq. (15)"},{"comment":"The attribution of the edge-state direction and dissipation rate to the complex chirality (abstract and Sec. III.C) relies on the standard bulk-boundary correspondence between the periodic-boundary Chern number C₁ of Eq. (20) and the open-boundary strip spectra of Fig. 5. For non-Hermitian systems with non-reciprocal hoppings (here n ≠ m), non-Bloch corrections are known to alter this correspondence, and the authors explicitly defer its proof (\"a comprehensive investigation of the precise mathematical proof for the novel bulk-boundary correspondence\") to future work in Sec. V. The statement in the abstract is therefore stronger than what the manuscript establishes; either compute the non-Bloch (generalized Brillouin zone) topological invariant, or present the edge-state correspondence as an observation for the specific zigzag and armchair strips of Fig. 5 and Appendix B.","section":"Sec. III.C and Sec. V"}],"minor_comments":[{"comment":"The sentence \"PT symmetry is always broken, hence the eigenvalues of the Hamiltonian are expected to be complex\" is inaccurate in the Hermitian limit m = n, a = −b, where the spectrum is real; please write \"generically broken away from the Hermitian limit.\"","section":"Sec. III.A"},{"comment":"The parameter α introduced after Eq. (29) contains the factor (1 + g²γ)/ω_c, and with ω_c = |ω_c|/γ it is complex for generic complex γ, yet Eqs. (30)-(35) treat α Re(γ) and α Im(γ) as real quantities; please clarify the conditions under which the mapping to real m, n, a, b holds.","section":"Sec. IV, Eqs. (30)-(35)"},{"comment":"The phrase \"a anticlowise/clockwise moving edge state\" contains the typo \"anticlowise\"; in addition, please spell out the convention by which C₁ = +1 is identified with an anticlockwise edge mode, since this identification is not a universal convention.","section":"Sec. III.C"},{"comment":"The notation ∇_k γ_k(t)|_{k_M} is ambiguous (derivative of the Berry-phase endpoint integral versus evaluation at the saddle point), and the identity (6), cited to the Hermitian literature, is used at a complex stationary point k_M; please justify its validity for the complex biorthogonal connection and for complex k.","section":"Sec. II.A, Eqs. (5)-(6)"},{"comment":"The statement that a semiclassical EOM for non-Hermitian systems \"remains an open question\" is not reconciled with Refs. [31-35] and [38], which already treat non-Hermitian wave-packet dynamics; please state explicitly the new ingredients, in particular the biorthogonal connection giving time-independent curvature and the Δ² anomalous force.","section":"Sec. I"},{"comment":"The sentence \"When S_I ≠ 0, the time-reversal symmetry, H_k = H*_{−k} and B_k = −B_{−k} [7], is explicitly broken...\" states the time-reversal identities and then asserts that the symmetry is broken; please rewrite so that the identities are not presented as properties of the actual model.","section":"Sec. III.B"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the conditional verdict of the reader is appropriate. The core derivation is sound, and the toy-model test of Fig. 1 is a genuine check of the Berry-curvature terms, but the bulk Haldane simulation's explicit omission of the Berry phase, the under-specified use of Eq. (14), and the deferred bulk-boundary proof need to be addressed before publication. The authors are honest about the limitation in the text, yet the abstract overstates the confirmation. The novelty relative to Refs. [31-35] and [38] should also be demarcated more carefully. I recommend major revision rather than rejection, because the load-bearing issues are fixable by reworking the claims and, ideally, by adding a Berry-phase-inclusive simulation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The equations of motion are the real contribution here, and they are worth a serious look. Using the biorthogonal Berry connection, the authors derive an anomalous velocity -Re(F x B) and an anomalous force -Delta^2 Im(F x B) in the momentum equation, plus a gain-gradient drift Delta^2 Im(gradient E). The derivation is self-contained and internally consistent, and the toy model in Fig. 1 does test the Berry-curvature terms with a complex B. For a paper in this area, that is a substantive step beyond prior left/right-eigenvector treatments.\n\nThe Haldane model is a clean construction: non-reciprocal NNN hoppings define a 'complex chirality', the Chern number follows C1 = -sign(b-a), and the sign of Im(B) follows sign(n-m). The edge spectra on zigzag and armchair edges are consistent with the bulk Chern number, which is a reasonable sanity check. The proposed cavity mapping is a plausible path to realizing the model, though it is only a proposal.\n\nNow the soft spots, in proportion. The stress-test note is right: the bulk Haldane simulation in Sec. III.E explicitly omits the Berry phase because of gauge-dependence issues. Since the F x B terms enter through exactly that Berry phase, the 'good agreement' in Fig. 7(i,j) tests only the drift terms Re(gradient E) and Delta^2 Im(gradient E), not the topological anomalous velocity or force. The abstract's claim of direct numerical confirmation is therefore too strong for the central content. Relatedly, Eq. (14) for the non-Gaussian case is not a closed equation for k_M without specifying how f evolves, so that comparison is under-specified. And the bulk-boundary correspondence is assumed rather than proven; the authors defer its proof to future work. For non-reciprocal hoppings the standard Bloch Chern number can fail to predict open-boundary edge states, so the statement that complex chirality dictates edge-state direction and dissipation rate rests on an unverified premise, even though the spectra in the figures look right for the cases shown.\n\nMinor: the cavity mapping has sign conventions in Eqs. (30)-(31) versus (32)-(35) that deserve a careful check, but this doesn't undermine the wave-packet derivation.\n\nBottom line: this is a serious theory paper. The central EOM are likely correct and the toy model supports them; the Haldane bulk simulation does not test the new topological terms, and the BBC is open. That is exactly the kind of thing refereeing should fix. Send it out, and ask the authors to either include the Berry phase in the bulk simulation or present an alternative test of the F x B terms, and to prove or numerically check the non-Bloch bulk-edge correspondence. I'd cite the EOM if I worked on non-Hermitian transport.","headline":"A correct-looking central derivation and a clean toy-model test, but the Haldane bulk numerics omit the Berry phase, so the abstract overclaims; still deserves refereeing.","tokens_in":17899,"tokens_out":3065,"would_cite":true,"duration_ms":30919,"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 imaginary part of the Berry curvature acts as an anomalous force on non-Hermitian wave packets, and a two-sign 'complex chirality' sets the direction and dissipation of edge states.","keywords":["non-Hermitian topology","semiclassical equations of motion","complex Berry curvature","biorthogonal Berry phase","non-Hermitian Haldane model","non-reciprocal hopping","chiral optical cavity","edge-state chirality"],"falsifier":"Take the same strip as in Fig. 5 and compute its open-boundary spectrum using the generalized Brillouin zone (non-Bloch band theory) at large non-reciprocal hopping, rather than the periodic Bloch ansatz. If the number or direction of edge states disagrees with $C_1=-\\mathrm{sign}(b-a)$, or if the sign of $\\mathrm{Im}(E)$ on the edge modes does not flip with $\\mathrm{sign}(n-m)$, the claimed bulk-edge correspondence is false, even though the wave-packet equations (7)-(8) could still hold.","tokens_in":16901,"feed_emoji":"🌀","tokens_out":10310,"duration_ms":102448,"temperature":0.7,"pith_summary":"The paper sets out to give non-Hermitian topological wave packets a semiclassical description on the same footing as Hermitian ones. Using the biorthogonal complex Berry phase, it derives equations of motion in which the complex Berry curvature produces two effects: a real-part term that bends the real-space velocity and an imaginary-part term that acts as an anomalous force steering the packet in momentum space, alongside a gain-gradient drift. To anchor the formalism, the authors build a non-Hermitian Haldane model with non-reciprocal next-nearest-neighbor hoppings and show that its 'complex chirality'—two signs, one from the real imbalance and one from the imaginary imbalance—separately controls the imaginary and real parts of the Berry curvature, the Chern number, and the direction and dissipation of edge states. Direct numerical wave-packet simulations match the equations, and a non-Hermitian chiral cavity is proposed as a physical platform. If correct, the paper supplies a transport theory for non-Hermitian topological matter and a concrete experimental knob for controlling edge-state behavior.","feed_headline":"Imaginary Berry curvature exerts an anomalous force","feed_subtitle":"New equations show how gain and complex chirality steer wave packets and topological edge states.","key_machinery":"The machinery has two parts. First, the biorthogonal Berry connection $A_k = i\\langle\\psi_L(k)|\\nabla_k\\psi_R(k)\\rangle$, built from left and right eigenstates of the non-Hermitian Hamiltonian, defines a generally complex Berry curvature $B_k = \\nabla_k \\times A_k$; the paper tracks wave-packet maxima by deforming the $k$-integration contour into the complex plane, which converts the Berry phase into the identity $d(\\nabla_k\\gamma_k)/dt = F\\times B_k$. Second, the 'complex chirality' of the non-Hermitian Haldane model is the sign pair $S_R=\\mathrm{sign}(n-m)$ and $S_I=\\mathrm{sign}(b-a)$ attached to the real and imaginary imbalances of the non-reciprocal next-nearest-neighbor hoppings $m+ia$ and $n+ib$; in this model $S_I$ fixes the Chern number and edge-state direction, while $S_R$ fixes the sign of the imaginary Berry curvature and the gain/loss asymmetry of edge states.","core_discovery":"Using the biorthogonal Berry connection, the paper derives the semiclassical equations of motion $\\dot{\\mathbf{r}}_M = \\mathrm{Re}(\\nabla_{\\mathbf{k}} E)|_{\\mathbf{k}_M} - \\mathrm{Re}(\\mathbf{F} \\times \\mathbf{B})|_{\\mathbf{k}_M}$ and $\\dot{\\mathbf{k}}_M = \\mathbf{F} + \\Delta^2[\\mathrm{Im}(\\nabla_{\\mathbf{k}} E)|_{\\mathbf{k}_M} - \\mathrm{Im}(\\mathbf{F} \\times \\mathbf{B})|_{\\mathbf{k}_M}]$, where $\\mathbf{B}$ is the biorthogonal Berry curvature. The imaginary part of $\\mathbf{B}$ therefore produces an anomalous force rather than only an anomalous velocity, and gradients of the imaginary energy push the packet toward gain. In the non-Hermitian Haldane model with non-reciprocal next-nearest-neighbor hoppings $m+ia$ and $n+ib$, the paper finds $C_1 = -\\mathrm{sign}(b-a)$ and $\\mathrm{sign}(\\mathrm{Im}\\,B) = \\mathrm{sign}(n-m)$, so the sign pair $(S_R,S_I)=(\\mathrm{sign}(n-m),\\mathrm{sign}(b-a))$ organizes the bulk Chern number, the direction of edge states, and whether edge states amplify or decay. Numerical simulations of both edge and bulk wave packets confirm the derived equations.","pith_inferences":["If the same peak-tracking derivation is repeated in a Lindblad master equation, the mean position and momentum of an ensemble of trajectories should acquire the same anomalous force $-\\Delta^2\\mathrm{Im}(F\\times B)$ plus stochastic corrections; this would extend the result to genuinely open quantum systems, which the paper does not do.","The anomalous force is perpendicular to the external force and grows with the squared packet width $\\Delta^2$, so it could be isolated experimentally by sending a wide packet with a tunable drive and measuring transverse drift in momentum space; this test is not proposed in the paper.","The cavity mapping implies a direct control protocol: sweeping the cavity polarization angle through the point where $\\cos(2\\theta)=0$ should flip the Chern number, and switching the cavity from loss to gain should reverse the edge-state dissipation asymmetry; these are corollaries of Eqs. (32)-(35) rather than stated proposals."],"forward_implications":["A Gaussian wave packet in any non-Hermitian band with complex Berry curvature obeys the new equations of motion, so even a constant external force bends the packet's momentum-space path perpendicular to the force through the term $-\\Delta^2\\mathrm{Im}(F\\times B)$.","In the non-Hermitian Haldane model, the Chern number is $C_1 = -\\mathrm{sign}(b-a)$, which means the imbalance of the imaginary parts of the next-nearest-neighbor hoppings alone decides whether edge states move clockwise or counterclockwise.","The sign of $n-m$, the imbalance of the real parts, fixes the sign of $\\mathrm{Im}(B)$ and the sign of the edge-state gain or loss, so one lattice can amplify a chiral edge state moving one way and damp the one moving the other way.","For one-dimensional edge transport, the packet's momentum-space center drifts at the rate $\\Delta^2\\mathrm{Im}(dE/dk)$, toward gain, while its real-space center moves with $\\mathrm{Re}(dE/dk)$; the simulations reproduce both motions.","The non-Hermitian chiral cavity maps to the model with explicit relations for $m,n,a,b$, so tuning the cavity chirality and loss should control both the Chern number and the edge-state dissipation rate."],"supporting_citations":[{"why":"Defines the biorthogonal Berry connection A_k = i<psi_L|grad psi_R> that the paper adopts for the complex Berry phase.","marker":"[30]"},{"why":"Supplies the Hermitian semiclassical wave-packet framework and the anomalous-velocity term that this work extends to non-Hermitian bands.","marker":"[7]"},{"why":"Gives the identity d(grad gamma_k)/dt = F x B_k that turns Berry-phase gradients into the velocity and force terms.","marker":"[8]"},{"why":"Introduces the complex-k saddle-point method for locating wave-packet maxima that the derivation adapts.","marker":"[38]"},{"why":"Provides the biorthogonality and completeness relations for non-Hermitian eigenstates used throughout.","marker":"[16]"},{"why":"Defines the non-Hermitian Chern number as an integral of the Berry curvature, used to obtain C_1 = -S_I.","marker":"[28]"},{"why":"Grounds the statement that C_1 = +1 and C_1 = -1 correspond to counterclockwise and clockwise chiral edge states.","marker":"[47, 48]"},{"why":"Supplies the chiral-cavity Hamiltonian and transformation that yield the effective non-Hermitian hopping parameters.","marker":"[42]"}],"fun_headline_variants":["Imaginary Berry curvature yields an anomalous force","Complex chirality sets edge-state direction and gain","Non-Hermitian wave packets feel imaginary Berry force","Cavity proposal for non-Hermitian Haldane edge states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the periodic-lattice Chern number $C_1=-\\mathrm{sign}(b-a)$, computed from the biorthogonal Berry curvature, correctly predicts the number and chirality of open-boundary edge states; the authors defer a proof of this non-Hermitian bulk-boundary correspondence, and non-reciprocal systems are exactly where non-Bloch corrections can break such predictions.","fun_headline_variants_meta":{"raw":{"variants":["Imaginary Berry curvature yields an anomalous force","Complex chirality sets edge-state direction and gain","Non-Hermitian wave packets feel imaginary Berry force","Cavity proposal for non-Hermitian Haldane edge states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001183,"raw_usage":{"total_tokens":4961,"prompt_tokens":1098,"completion_tokens":3863,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":714,"completion_tokens_details":{"reasoning_tokens":3798}},"tokens_in":714,"tokens_out":3863,"duration_ms":28885,"temperature":1.0,"reasoning_tokens":3798,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T17:28:54.867912+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the same strip as in Fig. 5 and compute its open-boundary spectrum using the generalized Brillouin zone (non-Bloch band theory) at large non-reciprocal hopping, rather than the periodic Bloch ansatz. If the number or direction of edge states disagrees with $C_1=-\\mathrm{sign}(b-a)$, or if the sign of $\\mathrm{Im}(E)$ on the edge modes does not flip with $\\mathrm{sign}(n-m)$, the claimed bulk-edge correspondence is false, even though the wave-packet equations (7)-(8) could still hold.","supporting_citations":[{"cited_title":"Sun, Physica Scripta 48, 393 (1993), URL https: //dx.doi.org/10.1088/0031-8949/48/4/002","cited_arxiv_id":null,"evidence_quote":"Defines the biorthogonal Berry connection A_k = i<psi_L|grad psi_R> that the paper adopts for the complex Berry phase."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Hermitian semiclassical wave-packet framework and the anomalous-velocity term that this work extends to non-Hermitian bands."},{"cited_title":"Muschietti and C","cited_arxiv_id":null,"evidence_quote":"Introduces the complex-k saddle-point method for locating wave-packet maxima that the derivation adapts."},{"cited_title":"Esaki, M","cited_arxiv_id":null,"evidence_quote":"Defines the non-Hermitian Chern number as an integral of the Berry curvature, used to obtain C_1 = -S_I."},{"cited_title":"Angular Momentum-Dependent Spectral Shift in Chiral Vacuum Cavities","cited_arxiv_id":"2307.14964","evidence_quote":"Supplies the chiral-cavity Hamiltonian and transformation that yield the effective non-Hermitian hopping parameters."}],"review_version":1}