arXiv subjects

Wenting Zhou

Publications and source records attributed to Wenting Zhou.

At least 19 recordsLinked to original sources

arXiv2026

Frequency-domain extended-effective-source gravitational self-force for eccentric Schwarzschild orbits

Effective-source formulations provide a practical route to gravitational self-force (GSF) calculations when singular retarded fields cannot be handled directly mode by mode. Up to now, only frequency-domain first-order GSF for circular orbits in Schwarzschild spacetime was obtained with the effective-source method. We construct and implement a frequency-domain extended effective-source (EES) formulation for first-order Lorenz-gauge GSF on eccentric orbits in Schwarzschild spacetime. The central obstacle is that radial libration forces the physical puncture and effective source to switch between interior and exterior branches, limiting their differentiability and producing slow spectral convergence. We overcome this obstruction by analytically extending both branches across the libration region and solving the coupled Lorenz-gauge perturbation equations for the resulting smooth extended sources. This provides an end-to-end gravitational EES implementation for eccentric motion and validates a computational ingredient needed for extending frequency-domain effective-source calculations toward second order.

gr-qc↗

arXiv2026

The point-particle-limit effective-source approach for computing gravitational self-force in the Lorenz gauge

The traditional effective-source method is hampered by complex analytical expressions and the inherent smoothness limit, which incur high computational costs and complicate implementation. To overcome these limitations, we introduce the point-particle-limit effective source method, which analytically takes the size of the effective source to zero, thereby transforming the problem into a well-defined jump condition of retarded metric field at the particle position governed by the local singular field. This formulation naturally pairs with a discontinuous Galerkin scheme, whose inherent capacity for accommodating solution discontinuities enables highly accurate enforcement of the jump conditions. We apply both the traditional and point-particle-limit effective source method to calculate the time-domain gravitational metric perturbation and gravitational self-force in the Lorenz gauge on a point particle in a circular orbit around a Schwarzschild black hole. The comparison of numerical results shows the excellent advantage of the point-particle-limit effective source method, which validates the correctness and efficiency of the point-particle-limit effective source method and thereby establishes a numerical foundation for computing generic geodesic orbits or long-time self-consistent orbital evolution.

gr-qc↗

arXiv2026

Signatures of the Israel Junction II: Double Photon Rings in Slowly Rotating Kerr Spacetime with Thin Shell

Applying the junction conditions to the slowly rotating Kerr spacetime with a thin shell, when higher order terms in the spin parameter a can be neglected, we find that while the angular momentum $L$ and Carter constant $C$ of the ray remain unchanged upon crossing the shell, its energy $E$ does not. Consequently, the impact parameters $\eta=L/E$ and $\xi=C/E^2$ of the ray are discontinued at the shell. Utilizing this transformation, we study the shadow of this spacetime and the corresponding images from an equatorial thin accretion disk. The presence of the shell gives rise to distinctive features in the observed images. Notably, we observe distinct double photon rings in the images, which can gradually merge into a single ring. Moreover, the shadow boundaries and the photon rings do not exhibit a one-to-one correspondence. The abrupt changes in redshift factor and the truncated photon regions profoundly influence the image, producing distinctive features such as the step-like structures. These features in shell-equipped spacetimes can help evaluate, through future astronomical observations, the applicability of the Israel junction condition and the shell model in real astrophysical systems.

gr-qc↗

arXiv2026

RelayGR: Scaling Long-Sequence Generative Recommendation via Cross-Stage Relay-Race Inference

Real-time recommender systems execute multi-stage cascades (retrieval, pre-processing, fine-grained ranking) under strict tail-latency SLOs, leaving only tens of milliseconds for ranking. Generative recommendation (GR) models can improve quality by consuming long user-behavior sequences, but in production their online sequence length is tightly capped by the ranking-stage P99 budget. We observe that the majority of GR tokens encode user behaviors that are independent of the item candidates, suggesting an opportunity to pre-infer a user-behavior prefix once and reuse it during ranking rather than recomputing it on the critical path. Realizing this idea at industrial scale is non-trivial: the prefix cache must survive across multiple pipeline stages before the final ranking instance is determined, the user population implies cache footprints far beyond a single device, and indiscriminate pre-inference would overload shared resources under high QPS. We present RelayGR, a production system that enables in-HBM relay-race inference for GR. RelayGR selectively pre-infers long-term user prefixes, keeps their KV caches resident in HBM over the request lifecycle, and ensures the subsequent ranking can consume them without remote fetches. RelayGR combines three techniques: 1) a sequence-aware trigger that admits only at-risk requests under a bounded cache footprint and pre-inference load, 2) an affinity-aware router that co-locates cache production and consumption by routing both the auxiliary pre-infer signal and the ranking request to the same instance, and 3) a memory-aware expander that uses server-local DRAM to capture short-term cross-request reuse while avoiding redundant reloads. We implement RelayGR on Huawei Ascend NPUs and evaluate it with real queries. Under a fixed P99 SLO, RelayGR supports up to 1.5$\times$ longer sequences and improves SLO-compliant throughput by up to 3.6$\times$.

cs.DC↗

arXiv2025

Experimental investigation of uncertainty relations for non-Hermitian operators

Uncertainty relations for Hermitian operators have been confirmed through many experiments. However, previous experiments have only tested the special case of non-Hermitian operators, i.e., uncertainty relations for unitary operators. In this study, we explore uncertainty relations for general non-Hermitian operators, which include Hermitian and unitary operators as special cases. We perform experiments with both real and complex non-Hermitian operators for qubit states, and confirm the validity of the uncertainty relations within the experimental error. Our results provide experimental evidence of uncertainty relations for non-Hermitian operators. Furthermore, our methods for realizing and measuring non-Hermitian operators are valuable in characterizing open-system dynamics and enhancing parameter estimation.

quant-ph↗

arXiv2025

Images and photon regions of continuous photon sphere spacetime

We study images of spacetimes containing continuous photon spheres (CPS). For a self-gravitating, isotropic, spherically symmetric spacetime with CPS, we find that a thin accretion disk produces images that closely resemble those of a Schwarzschild black hole, despite significant differences in photon dynamics. More generally, for any static, pherically symmetric spacetime with a luminous CPS core, the image profile is universal: members of this class produce identical image shapes, differing only by an overall normalization factor. This universality is, however, sensitive to the nature of the accretion flow and breaks down for spherically symmetric infalling accretion, where Doppler shifts and non-static emission introduce image features that depend on the flow dynamics and the metric. Finally, we investigate photon regions in a rotating CPS spacetime and find that unlike in Kerr spacetime, the photon region appears as one or two angular sectors in a constant-$\phi$ cross section. These distinctive photon region properties could produce observable signatures that distinguish rotating CPS spacetimes from the Kerr one.

gr-qc↗

arXiv2025

Effects of acceleration on interatomic interactions

The Unruh effect establishes a fundamental equivalence between acceleration and thermality by demonstrating that a uniformly accelerated ground-state detector undergoes excitation as if immersed in a thermal bath. In this paper, we investigate how acceleration influences the interaction between two ground-state atoms that are synchronously and uniformly accelerated in vacuum with proper acceleration $a$ and coupled to a fluctuating electromagnetic field. We find that the resulting interaction potential comprises both diagonal components $(\delta E)^{jk}$ with $j=k$, which are present in both inertial and acceleration cases, and off-diagonal components $(\delta E)^{jk}$ with $j\neq k$, which arise exclusively due to acceleration and vanish in the inertial case. The dependence of each component on acceleration and interatomic separation $L$ generally differs. For small accelerations, the leading-order diagonal components of the van der Waals (vdW) and Casimir-Polder (CP) interaction potentials remain unchanged from their inertial counterparts, exhibiting the standard scaling behaviors $\sim L^{-6}$ and $\sim L^{-7}$, respectively. In contrast, the off-diagonal components scale as $\sim a^2L^{-4}$ in the vdW subregions and $\sim a^2L^{-5}$ in the CP subregion. However, when the acceleration becomes sufficiently large, both diagonal and off-diagonal components of the vdW and CP interaction potentials are significantly modified, giving rise to entirely new interaction behaviors that deviate from those observed in the inertial case, whether in vacuum or thermal environments, indicating a breakdown of the acceleration-thermality equivalence established by the Unruh effect for single detectors.

gr-qc↗

arXiv2025

Generic effective sources for first-order in mass-ratio gravitational self-force calculations in Schwarzschild spacetime

The numerical calculation of gravitational self-force in extreme mass ratio inspiral systems is fundamentally challenging due to the singular nature of point-particle sources. To overcome these difficulties, the effective source method offers an innovative alternative by replacing traditional regularization techniques with a reformulation of the problem. In this paper, we present the first fully analytic framework for constructing effective sources to compute the gravitational self-force for generic orbits in Schwarzschild spacetime. By reformulating the singular field through angular modulation in terms of a tetrad decomposition, the effective source can be constructed with the linear combination of scalar modes. The derived effective source is continuous across the particle's worldline, enabling efficient numerical implementation in $1+1$ dimensions.

gr-qc↗

arXiv2024

Interaction between Unruh-Dewitt detectors exclusively due to acceleration: A Parallel to the FDU Effect

We have discovered an interaction between two detectors in a vacuum that emerges exclusively due to acceleration, akin to the spontaneous excitation of a single detector as predicted by the Fulling-Davies-Unruh (FDU) effect. However, this interaction contrasts sharply with the FDU effect, which suggests that a uniformly accelerated detector behaves as if it were in a thermal bath, as the discovered interaction does not manifest in a thermal environment. The novel interaction displays unique dependencies on the separation between detectors: it can be either attractive or repulsive, with the potential to transition between these behaviors as the inter-detector separation changes. More intriguingly, it exhibits a surprising large-small duality in its dependence on acceleration, suggesting the existence of an optimal acceleration at which the interaction is strongest, in contrast to the monotonic acceleration-dependence of the FDU effect.

gr-qc↗

arXiv2024

P/D-Serve: Serving Disaggregated Large Language Model at Scale

Serving disaggregated large language models (LLMs) over tens of thousands of xPU devices (GPUs or NPUs) with reliable performance faces multiple challenges. 1) Ignoring the diversity (various prefixes and tidal requests), treating all the prompts in a mixed pool is inadequate. To facilitate the similarity per scenario and minimize the inner mismatch on P/D (prefill and decoding) processing, fine-grained organization is required, dynamically adjusting P/D ratios for better performance. 2) Due to inaccurate estimation on workload (queue status or maintained connections), the global scheduler easily incurs unnecessary timeouts in prefill. 3) Block-fixed device-to-device (D2D) KVCache transfer over cluster-level RDMA (remote direct memory access) fails to achieve desired D2D utilization as expected. To overcome previous problems, this paper proposes an end-to-end system P/D-Serve, complying with the paradigm of MLOps (machine learning operations), which models end-to-end (E2E) P/D performance and enables: 1) fine-grained P/D organization, mapping the service with RoCE (RDMA over converged ethernet) as needed, to facilitate similar processing and dynamic adjustments on P/D ratios; 2) on-demand forwarding upon rejections for idle prefill, decoupling the scheduler from regular inaccurate reports and local queues, to avoid timeouts in prefill; and 3) efficient KVCache transfer via optimized D2D access. P/D-Serve is implemented upon Ascend and MindSpore, has been deployed over tens of thousands of NPUs for more than eight months in commercial use, and further achieves 60\%, 42\% and 46\% improvements on E2E throughput, time-to-first-token (TTFT) SLO (service level objective) and D2D transfer time. As the E2E system with optimizations, P/D-Serve achieves 6.7x increase on throughput, compared with aggregated LLMs.

cs.DC↗

arXiv2022

Understanding thermal nature of de Sitter spacetime via inter-detector interaction

The seminar discovery by Gibbons and Hawking that a freely falling detector observes an isotropic background of thermal radiation reveals that de Sitter space is equivalent to a thermal bath at the Gibbons-Hawking temperature in Minkowski space, as far as the response rate of the detector is concerned. Meanwhile, for a static detector which is endowed with a proper acceleration with respect to the local freely-falling detectors, the temperature becomes the square root of the sum of the squared Gibbons-Hawking temperature and the squared Unruh temperature associated with the proper acceleration of the detector. Here, we demonstrate, by examining the interaction of two static detectors in the de Sitter invariant vacuum, that de Sitter space in regard to its thermal nature is unique on its own right in the sense that it is even neither equivalent to the thermal bath in Minkowski space when the static detectors become freely-falling nor to the Unruh thermal bath at the cosmological horizon where the Unruh effect dominates, insofar as the behavior of the inter-detector interaction in de Sitter space dramatically differs both from that in the Minkowski thermal bath and the Unruh thermal bath.

gr-qc↗

arXiv2022

Quantum thermal field fluctuation induced corrections to the interaction between two ground-state atoms

We generalize the formalism proposed by Dalibard, Dupont-Roc, and Cohen-Tannoudji [the DDC formalism] in the fourth order for two atoms in interaction with scalar fields in vacuum to a thermal bath at finite temperature $T$, and then calculate the interatomic interaction energy of two ground-state atoms separately in terms of the contributions of thermal fluctuations and the radiation reaction of the atoms and analyze in detail the thermal corrections to the van der Waals and Casimir-Polder interactions. We discover a particular region, i.e., $\sqrt[4]{λ^3β}\ll L\ll λ$ with $L$, $β$ and $λ$ denoting the interatomic separation, the wavelength of thermal photons and the transition wavelength of the atoms respectively, where the thermal corrections remarkably render the van der Waals force, which is usually attractive, repulsive, leading to an interesting crossover phenomenon of the interatomic interaction from attractive to repulsive as the temperature increases. We also find that the thermal corrections cause significant changes to the Casimir-Polder force when the temperature is sufficiently high, resulting in an attractive force proportional to $TL^{-3}$ in the $λ\llβ\ll L$ region, and a force which can be either attractive or repulsive and even vanishing in the $ β\llλ\ll L$ region depending on the interatomic separation.

hep-th↗

arXiv2022

Probing long-range properties of vacuum altered by uniformly accelerating two spatially separated Unruh-DeWitt detectors

In a quantum sense, vacuum is not an empty void but full of virtual particles (fields). It may have long-range properties, be altered, and even undergo phase transitions. It is suggested that long-range properties of a quantum vacuum may be probed by distributing matter over a large spatial volume. Here, we study a simplest example of such, i.e., two uniformly accelerated Unruh-DeWitt detectors which are spatially separated, and examine the inter-detector interaction energy arising from the coupling between the detectors and fluctuating fields to see if novel phenomena related to the long-range properties emerge of a vacuum altered by uniformly accelerating two spatially separated detectors through it. Our results show that when the inter-detector separation is much larger than the thermal wavelength of the Unruh thermal bath, the inter-detector interaction displays a completely new behavior, which, as compared with that of the inertial detectors, is surprisingly exclusively acceleration-dependent, signaling a new phase of the vacuum in which its imprint as seen by two inertial observers seems to be completely wiped out. Moreover, we demonstrate that the inter-detector interaction in the near region can be significantly enhanced by the accelerated motion in certain circumstances, and with two Rydberg atoms as the detectors, the acceleration required for an experimentally detectable enhancement of the interaction energy can be $10^5$ times smaller than that required for the detection of the Unruh effect.

hep-th↗

arXiv2020

Interatomic interaction of two ground-state atoms in vacuum: contributions of vacuum fluctuations and radiation reaction

We generalize the formalism proposed by Dalibard, Dupont-Roc and Cohen-Tannoudji [the DDC formalism] to the fourth order of the coupling constant, which can be used to study the interatomic interaction of two ground-state atoms coupled with the vacuum scalar fields. We show that the interatomic potential can be attributed to the joint effect of both vacuum fluctuations and the radiation reaction of atoms. Remarkably, the formulae we derived for the contributions of vacuum fluctuations and the radiation reaction to the interatomic potential upon which future research on fourth-order effects in particular circumstances can be based differ from those in the existing literature [Phys. Rev. D 95, 085014 (2017)].

quant-ph↗

arXiv2020

Radiation-reaction-induced transitions of two maximally entangled atoms in non-inertial motion

We apply the DDC formalism [proposed by Dalibard, Dupont-Roc and Cohen-Tannoudji] to study the average rate of change of energy of two identical two-level atoms interacting with the vacuum massless scalar field in synchronized motion along stationary trajectories. By separating the contributions of vacuum fluctuations and atomic radiation reaction, we first show that for the two-atom system initially prepared in the factorizable eigenstates $|g_Ag_B\rangle$ and $|e_Ae_B\rangle$, where $g$ and $e$ represent the ground state and the excited state of a single atom respectively, both vacuum fluctuations and atomic radiation reaction contribute to the average rate of change of energy of the two-atom system, and the contribution of vacuum fluctuations is independent of the interatomic separation while that of atomic radiation reaction is dependent on it. This is contrary to the existing results in the literature where vacuum fluctuations are interatomic-separation dependent. However, if the two-atom system is initially prepared in the unfactorizable symmetric/antisymmetric entangled state, the average rate of change of energy of the two-atom system is never perturbed by the vacuum fluctuations, but is totally a result of the atomic radiation reaction. We then consider two special cases of motion of the two-atom system which is initially prepared in the symmetric/antisymmetric entangled state, i.e., synchronized inertial motion and synchronized uniform acceleration. In contrast to the average rate of change of energy of a single uniformly accelerated atom, the average rate of change of energy of the uniformly accelerated two-atom system is nonthermal-like. The effects of noninertial motion on the transitions of states of the two correlated atoms are also discussed.

quant-ph↗

arXiv2018

Time-dependent resonance interaction energy between two entangled atoms under non-equilibrium conditions

We consider the time-dependent resonance interaction energy between two identical atoms, one in the ground state and the other in an excited state, and interacting with the vacuum electromagnetic field, during a nonequilibrium situation such as the dynamical atomic self-dressing process. We suppose the two atoms prepared in a correlated, symmetric or antisymmetric, state. Since the atoms start from a nonequilibrium conditions, their interaction energy is time dependent. We obtain, at second order in the atom-field coupling, an analytic expression for the time-dependent resonance interaction energy between the atoms. We show that this interaction vanishes when the two atoms are outside the light-cone of each other, in agreement with relativistic causality, while it instantaneously settles to its stationary value after time $t=R/c$ ($R$ being the interatomic distance), as obtained in a time-independent approach. We also investigate the time-dependent electric energy density in the space around the two correlated atoms, in both cases of antisymmetric (subradiant) and symmetric (superradiant) states, during the dressing process of our two-atom system. We show that the field energy density vanishes in points outside the light-cone of both atoms, thus preserving relativistic causality. On the other hand, inside the light-cone of both atoms, the energy density instantaneously settles to its stationary value. Specifically, for points at equal distance from the two atoms, we find that it vanishes if the two atoms are prepared in the antisymmetric (subradiant) state, while it is enhanced, with respect to the case of atoms in a factorized state, in the symmetric (superradiant) state. The physical meaning of these results is discussed in detail in terms of interference effects of the field emitted by the two atoms.

quant-ph↗

arXiv2018

Resonance dipole-dipole interaction between two accelerated atoms in the presence of a reflecting plane boundary

We study the resonant dipole-dipole interaction energy between two uniformly accelerated identical atoms, one excited and the other in the ground state, prepared in a correlated {\em Bell-type} state, and interacting with the scalar field or the electromagnetic field nearby a perfectly reflecting plate. We suppose the two atoms moving with the same uniform acceleration, parallel to the plane boundary, and that their separation is constant during the motion. We separate the contributions of vacuum fluctuations and radiation reaction field to the resonance energy shift of the two-atom system, and show that Unruh thermal fluctuations do not affect the resonance interaction, which is exclusively related to the radiation reaction field. However, nonthermal effects of acceleration in the radiation-reaction contribution, beyond the Unruh acceleration-temperature equivalence, affect the resonance interaction energy. By considering specific geometric configurations of the two-atom system relative to the plate, we show that the presence of the mirror significantly modifies the resonance interaction energy between the two accelerated atoms. In particular, we find that new and different features appear with respect to the case of atoms in the free space, related to the presence of the boundary and to the peculiar structure of the quantum electromagnetic field vacuum in the locally inertial frame. Our results suggest the possibility to exploit the resonance interaction between accelerated atoms, as a probe for detecting the elusive effects of atomic acceleration on radiative processes.

quant-ph↗

arXiv2018

Vacuum fluctuations and radiation reaction contributions to the resonance dipole-dipole interaction between two atoms near a reflecting boundary

We investigate the resonance dipole-dipole interaction energy between two identical atoms, one in the ground state and the other in the excited state, interacting with the electromagnetic field in the presence of a perfectly reflecting plane boundary. The atoms are prepared in a correlated (symmetric or anti-symmetric) Bell-type state. Following a procedure due to Dalibard et. al. [J. Dalibard et. al., J. Phys. (Paris) {\bf 43}, 1617 (1982); {\bf 45}, 637 (1984)], we separate the contributions of vacuum fluctuations and radiation reaction (source) field to the resonance interaction energy between the two atoms, and show that only the source field contributes to the interatomic interaction, while vacuum field fluctuations do not. By considering specific geometric configurations of the two-atom-system with respect to the mirror and specific choices of dipole orientations, we show that the presence of the mirror significantly affects the resonance interaction energy and that different features appear with respect to the case of atoms in free space, for example a change in the spatial dependence of the interaction. Our findings also suggest that the presence of a boundary can be exploited to tailor and control the resonance interaction between two atoms, as well as the related energy transfer process. The possibility of observing these phenomena is also discussed.

quant-ph↗