Searcharxiv⌕ Search

arXiv subjects

Peng Xu

Publications and source records attributed to Peng Xu.

At least 73 records · Page 4Linked to original sources
arXiv2026

Artificial-atom arrays in moire superlattices for quantum optics

Solid-state platforms are particularly attractive for quantum optics because they facilitate on-chip integration and are compatible with established semiconductor and photonic technologies. However, a major challenge in solid-state quantum optics is the fabrication of arrays of identical emitters, such as quantum dots. In this work, we propose moire superlattices as a novel solid-state platform for manipulating light at the single-photon level. Moire superlattices form arrays of artificial-atom states characterized by nearly identical optical transition energies, tunable spacing, and highly adjustable electronic structures. They naturally operate as atomically thin, scalable, periodic emitters, making them ideal for quantum applications. Additionally, the extensive materials database of moire superlattices offers spectral coverage spanning a broad range of optical wavelengths.

physics.optics↗
arXiv2026

Fast and Coherent Transfer of Atomic Qubits in Optical Tweezers using Fiber Array Architecture

Programmable neutral-atom arrays offer a promising route toward scalable quantum computing, where coherent qubit transfer enables non-local connectivity and reduces resource overhead. However, transfer speed and motional heating remain key bottlenecks for fast and deep quantum circuits. Here, we employ a fiber array neutral-atom quantum computing architecture with site-resolved control of trap depths to realize smooth amplitude exchange between static and moving traps, thereby enabling fast and coherent qubit transfer with ultralow motional heating. With a 10 $μ$s in situ transfer between static and moving traps, we obtain a per-cycle heating rate of 0.156(9) $μ$K, sustain over 500 cycles with negligible atom loss, and achieve a quantum state fidelity of 0.99992(5) per cycle. For inter-site transfer between two separated static traps, the operation takes 120 $μ$s with 0.783(17) $μ$K heating per transfer, and remains negligible atom loss for up to 100 repeated cycles with a fidelity of 0.9998(1) per transfer. Furthermore, through experimental studies of parallel transfer, we establish a model that elucidates the relationship between array inhomogeneity and the transfer heating rate. This fast, low-heating coherent transfer capability provides a practical route for improving both speed and fidelity in atom-shuttling based quantum computing.

quant-ph↗
arXiv2026

A Dynamical Equilibrium Linking Nanohertz Stochastic Gravitational Wave Background to Cosmic Structure Formation

The stochastic gravitational wave background (SGWB) is conventionally treated as a passive relic of its astrophysical and cosmological sources, with negligible back-reaction on the matter content of the Universe. Here we show that this assumption needs to be modified once the SGWB and matter are treated as a dynamically coupled non-equilibrium system. Combining linearized general relativity with the fluctuation-dissipation theorem, we derive a generalized Langevin framework that drives the coupled system toward a dynamical equilibrium, which is characterized by a distinctive strain spectrum with a high-frequency cutoff $\mathcal{W}$, and a scale-dependent coupling parameter that screens gravity progressively for the most massive structures. Three findings support this framework. Fitting the equilibrium spectrum to the NANOGrav 15-year dataset yields a Bayes factor of $48\pm 3.8$ over the supermassive black hole binary baseline, achieved entirely within general relativity and the Standard Model. The PTA-calibrated screening mass scale $m_{c}\sim 10^{12}\text{--}10^{14}\,M_{\odot}$ overlaps, with no free cosmological parameter, the $Λ$CDM-derived linear-to-nonlinear transition mass $M_{\rm NL}$ of cosmic structure at $\sim 8\,h^{-1}\,\mathrm{Mpc}$. Most strikingly, promoting this concordance to a structural identification expresses $\mathcal{W}$ entirely in terms of $M_{\rm NL}$, and its inverse acquires a transparent physical reading as a coherence threshold for SGWB-matter coupling. $\mathcal{W}$ is thereby a derived quantity linking nanohertz gravitational-wave observables to the late-time cosmological sector. The framework makes distinctive scale-dependent predictions testable by forthcoming large-scale structure surveys and space-borne gravitational-wave observatories.

astro-ph.CO↗
arXiv2026

FluxMC: Rapid and High-Fidelity Inference for Space-Based Gravitational-Wave Observations

Bayesian inference in the physical sciences faces a fundamental challenge: the imperative for high-fidelity physical modeling often clashes with the intrinsic limitations of stochastic sampling algorithms. Complex, high-dimensional parameter spaces expose the universal vulnerability of conventional methods, e.g., Markov Chain Monte Carlo (MCMC), which struggle with the prohibitive costs of likelihood evaluations and the risk of entrapment in local optima. To resolve this impasse, we introduce FluxMC (Flow-guided Unbiased eXploration Monte Carlo), a machine learning-enhanced framework designed to shift the inference paradigm from blind local search to globally guided transport. It integrates Flow Matching with Parallel Tempering MCMC, effectively combining the global foresight of generative AI with the rigorous asymptotic convergence and local robustness of temperature-based sampling. We showcase the efficacy of this framework through the lens of space-based gravitational-wave (GW) astronomy -- a field representing the frontier of challenging parameter inversion. In the analysis of massive black hole binaries using high-fidelity waveforms (IMRPhenomHM), FluxMC achieves robust convergence in under five hours, whereas traditional Parallel Tempering MCMC fails to converge even after hundreds of hours, yielding high Jensen-Shannon divergences (JSD) of $O(10^{-1})$. Our method reduces the distributional error by two to three orders of magnitude. Furthermore, for computationally efficient models (IMRPhenomD), it eliminates systematic biases caused by local-optima entrapment. Ultimately, FluxMC removes the necessity to compromise between model accuracy and analysis speed, establishing a new computational foundation where scientific discovery is limited only by observational data quality, not by algorithmic capacity.

astro-ph.IM↗
arXiv2026

Detectability of axion-like dark matter for different time-delay interferometry combinations in space-based gravitational wave detectors

In the space-based gravitational wave detections, the axion-like dark matter would alter the polarization state of the laser link between spacecrafts due to the birefringence effect. However, current designs of space-based laser interferometer are insensitive to variations in the polarization angle. Thus, the additional wave plates are employed to enable the response of the axion-induced birefringence effect. We calculate and compare the sensitivities of different space-based detectors, accounting for three time-delay interferometry combinations, including Monitor, Beacon, and Relay. We find that the Monitor and Beacon combinations have better sensitivity in the high-frequency range, and the optimal sensitivity reaches $g_{aγ}\sim 10^{-13}\text{GeV}^{-1}$, while the Sagnac combination is superior in the low-frequency range. We also find that ASTROD-GW can cover the detection range of axion-like dark matter mass down to $10^{-20}\text{eV}$.

gr-qc↗
arXiv2026

Universal Dark-matter Density Profiles of Cosmic Filaments

We present a comprehensive analysis of the radial dark-matter (DM) density profiles of cosmic filaments in the hydrodynamical simulation TNG50. The cosmic web is extracted from high-resolution density grids at redshifts $z =$ 0, 0.5, 1, 2 and 3 using the DisPerSE algorithm. We show that the filament spine locations returned directly by DisPerSE do not accurately reflect the true density ridges. To address this issue, we introduce a "shrinking-cylinder" re-centering algorithm, which significantly increases the inferred central densities and restores the inner power-law behavior of the profiles. When the radial coordinate is scaled by the virial radii of the terminal nodes, the filament density profiles exhibit a nearly universal form that depends only weakly on redshift, node mass, and filament length. This result suggests that cosmic filaments, much like dark-matter halos, obey a form of structural self-similarity once an appropriate characteristic scale is introduced. By repeating the measurement using only smoothly distributed, unbound DM particles, we find that the apparent central cusp of the full profile is primarily produced by low-mass halos embedded along the filament spines, while the smooth component develops a flat core within $R/R_{\rm vir}\lesssim0.1$. The redshift evolution of this smooth component further suggests a transition from predominantly smooth filamentary accretion at high redshift to increasingly clumpy accretion at late times. Finally, we show that the universal filament profile is accurately described by a generalized triple-power-law model.

astro-ph.CO↗
arXiv2026

Version Control System for Data with MatrixOne

The rapid advancement of artificial intelligence has elevated data to a cornerstone of modern software systems. As data projects become increasingly complex and dynamic, version control for data has become essential rather than merely convenient. Existing version control systems designed for source code are inadequate for large-scale data management, as they often require loading entire datasets into memory for diff and merge operations. Database systems, while providing robust data management capabilities, lack native support for version control operations such as diff and merge between data forks. We present a version control system for data implemented in MatrixOne, a cloud-native relational database system. Our system leverages MatrixOne's immutable storage architecture and multi-version concurrency control (MVCC) to enable git-like operations on database tables at scale. The system supports the complete spectrum of version control operations: clone, tag/branch, diff, merge, and revert, on terabyte-scale datasets with near-instantaneous performance. This version control system enables data engineers to adopt established software engineering workflows: creating branches for isolated experimentation, submitting pull requests for change review, and running CI/CD pipelines efficiently and safely. Changes in the development environment are isolated from production in both data integrity and computing resources. Verified changes can be published to production in atomic transactions, ensuring data consistency and avoiding service disruptions.

cs.DB↗
arXiv2026

On the mapping between bound states and black hole quasinormal modes via analytic continuation: a spectral instability perspective

In this work, we investigate the relation between bound states and quasinormal modes within black hole perturbation theory in the context of spectral instability. Our analysis indicates that the reliability of such spectral mapping stretches beyond the domain of validity of the analytic continuation employed to connect the perturbative bound-state problem to the corresponding open-system dynamics. However, for the numerical scheme proposed by Völkel to work, the transformations of the metric parameters must be carried out in a region where the underlying Taylor expansion is convergent. As analytically accessible explicit examples, we explore the perturbed delta-function and Pöschl-Teller potential barriers. For the latter, we construct two distinct perturbative setups for which the convergence of the series expansion involved in the perturbation theory can be rigorously controlled. When the deformation is placed near the potential's extremum, the resulting corrections to the bound-state energies can be analytically continued to yield perturbed quasinormal frequencies, in agreement with known semi-analytic results. In contrast, when the perturbation is localized asymptotically far from the compact object, the bound states are only mildly modified and are accurately described by a perturbative expansion to the first order. However, the associated analytic continuation yields a strongly deformed spectrum that shows no clear connection to the quasinormal modes. These findings contribute to the effort to scrutinize the conditions under which bound states faithfully encode quasinormal spectra and to shed light on the underlying physics of black hole spectral instability.

gr-qc↗
arXiv2026

ScoutAttention: Efficient KV Cache Offloading via Layer-Ahead CPU Pre-computation for LLM Inference

Large language models encounter critical GPU memory capacity constraints during long-context inference, where KV cache memory consumption severely limits decode batch sizes. While existing research has explored offloading KV cache to DRAM, these approaches either demand frequent GPU-CPU data transfers or impose extensive CPU computation requirements, resulting in poor GPU utilization as the system waits for I/O operations or CPU processing to complete. We propose ScoutAttention, a novel KV cache offloading framework that accelerates LLM inference through collaborative GPU-CPU attention computation. To prevent CPU computation from bottlenecking the system, ScoutAttention introduces GPU-CPU collaborative block-wise sparse attention that significantly reduces CPU load. Unlike conventional parallel computing approaches, our framework features a novel layer-ahead CPU pre-computation algorithm, enabling the CPU to initiate attention computation one layer in advance, complemented by asynchronous periodic recall mechanisms to maintain minimal CPU compute load. Experimental results demonstrate that ScoutAttention maintains accuracy within 2.4% of baseline while achieving 2.1x speedup compared to existing offloading methods.

cs.LG↗
arXiv2026

Calibration of key parameters during the in-orbit phase for the Taiji-2 gravitational reference sensor

The Taiji mission, a pioneering Chinese space-borne gravitational wave observatory, requires ultra-precise calibration of its gravitational reference sensors (GRSs) to achieve its targeted sensitivity of $3\times10^{-15} \mathrm{\ m\ s^{-2}\ Hz^{-1/2}}$. Maintaining this precision is challenged by time-varying scale factors drifts and dynamic center-of-mass (c.m.) offsets between the test masses (TMs) and spacecraft, driven by factors such as propellant consumption, thermal effects and aging electronics. This paper develops an advanced in-orbit calibration framework that simultaneously estimates the GRS scale factors and c.m. offsets between TMs and spacecraft through a combination of spacecraft maneuvers and Kalman filter. By applying periodic torque signals to induce controlled spacecraft angular accelerations, we leverage star tracker and GRS readouts to disentangle coupled disturbances and achieve dual-parameter calibration with unprecedented precision, with scale factors errors below 0.2\% and c.m. offsets residuals within 100 $\mathrmμ$m, satisfies the Taiji-2 calibration requirements. This method is robust across different satellite configurations. The results not only ensure the feasibility of Taiji-2's scientific objectives but also establish a scalable calibration paradigm for future missions such as Taiji-3, where sub-micrometer c.m. stability and ultra-low noise gravitational reference will be essential.

astro-ph.IM↗
arXiv2026

KCLNet: Electrically Equivalence-Oriented Graph Representation Learning for Analog Circuits

Digital circuits representation learning has made remarkable progress in the electronic design automation domain, effectively supporting critical tasks such as testability analysis and logic reasoning. However, representation learning for analog circuits remains challenging due to their continuous electrical characteristics compared to the discrete states of digital circuits. This paper presents a direct current (DC) electrically equivalent-oriented analog representation learning framework, named \textbf{KCLNet}. It comprises an asynchronous graph neural network structure with electrically-simulated message passing and a representation learning method inspired by Kirchhoff's Current Law (KCL). This method maintains the orderliness of the circuit embedding space by enforcing the equality of the sum of outgoing and incoming current embeddings at each depth, which significantly enhances the generalization ability of circuit embeddings. KCLNet offers a novel and effective solution for analog circuit representation learning with electrical constraints preserved. Experimental results demonstrate that our method achieves significant performance in a variety of downstream tasks, e.g., analog circuit classification, subcircuit detection, and circuit edit distance prediction.

cs.LG↗
arXiv2026

An Opacity-Free Test of the Cosmic Distance Duality Relation Using Strongly Lensed Gravitational Wave Signals with Space-Based Detector Networks

The cosmic distance duality relation (CDDR), expressed as $d_L(z) = (1+z)^2 D_A(z)$, is a fundamental relation in modern cosmology. In this work, we apply a method to test the CDDR using simulated strongly lensed gravitational-wave (SLGW) signals from massive binary black holes (MBBH) as observed by proposed space-based detector networks. Our analysis is conducted under the point-mass lens model, considering the strong lensing scenario that produces two images. We generate 90 days of simulated SLGW data for 10 events based on the Population III stellar formation model, with source redshifts in the range $z_s \in [2,6]$ and lens redshifts in $z_L \in [0.2,1]$. The deviation of CDDR is parameterized by $η_1(z) = 1 + η_0 z$ and $η_2(z) = 1 + η_0 z/(1+z)$, and we incorporate the deviation parameter $η_0$ directly into the waveform model. Parameter estimation is performed within a Bayesian statistical framework, combining simulated data from both Taiji and LISA. For a single lensed event, the joint Taiji+LISA analysis improves the measurement precision of $η_0$ by roughly a factor of two compared with Taiji-only observations. By combining 10 simulated events, the population-level constraints on $η_0$, quantified by the half width of the $95\%$ credible interval, reach approximately $2.61\times10^{-4}$ ($1.72\times10^{-4}$) for the $η_1(z)$ parameterization and $1.22\times10^{-3}$ ($6.86\times10^{-4}$) for $η_2(z)$ in the Taiji-only (Taiji+LISA) scenario, respectively. The inferred values of $η_0$ remain consistent with $η_0 = 0$ within the estimated uncertainties, with no statistically significant evidence for deviations from the CDDR at the achieved precision. These results demonstrate the significant advantage of joint space-based observations for high-precision tests of the CDDR.

astro-ph.CO↗
arXiv2026

Symphony: A Cognitively-Inspired Multi-Agent System for Long-Video Understanding

Despite rapid developments and widespread applications of MLLM agents, they still struggle with long-form video understanding (LVU) tasks, which are characterized by high information density and extended temporal spans. Recent research on LVU agents demonstrates that simple task decomposition and collaboration mechanisms are insufficient for long-chain reasoning tasks. Moreover, directly reducing the time context through embedding-based retrieval may lose key information of complex problems. In this paper, we propose Symphony, a multi-agent system, to alleviate these limitations. By emulating human cognition patterns, Symphony decomposes LVU into fine-grained subtasks and incorporates a deep reasoning collaboration mechanism enhanced by reflection, effectively improving the reasoning capability. Additionally, Symphony provides a VLM-based grounding approach to analyze LVU tasks and assess the relevance of video segments, which significantly enhances the ability to locate complex problems with implicit intentions and large temporal spans. Experimental results show that Symphony achieves state-of-the-art performance on LVBench, LongVideoBench, VideoMME, and MLVU, with a 5.0% improvement over the prior state-of-the-art method on LVBench. Code is available at https://github.com/Haiyang0226/Symphony.

cs.CV↗
arXiv2026

AerialVLA: A Vision-Language-Action Model for UAV Navigation via Minimalist End-to-End Control

Vision-Language Navigation (VLN) for Unmanned Aerial Vehicles (UAVs) demands complex visual interpretation and continuous control in dynamic 3D environments. Existing hierarchical approaches rely on dense oracle guidance or auxiliary object detectors, creating semantic gaps and limiting genuine autonomy. We propose AerialVLA, a minimalist end-to-end Vision-Language-Action framework mapping raw visual observations and fuzzy linguistic instructions directly to continuous physical control signals. First, we introduce a streamlined dual-view perception strategy that reduces visual redundancy while preserving essential cues for forward navigation and precise grounding, which additionally facilitates future simulation-to-reality transfer. To reclaim genuine autonomy, we deploy a fuzzy directional prompting mechanism derived solely from onboard sensors, completely eliminating the dependency on dense oracle guidance. Ultimately, we formulate a unified control space that integrates continuous 3-Degree-of-Freedom (3-DoF) kinematic commands with an intrinsic landing signal, freeing the agent from external object detectors for precision landing. Extensive experiments on the TravelUAV benchmark demonstrate that AerialVLA achieves state-of-the-art performance in seen environments. Furthermore, it exhibits superior generalization in unseen scenarios by achieving nearly three times the success rate of leading baselines, validating that a minimalist, autonomy-centric paradigm captures more robust visual-motor representations than complex modular systems.

cs.CV↗
arXiv2026

Fast CZ Gate via Energy-Level Engineering in Superconducting Qubits with a Tunable Coupler

In superconducting quantum circuits, decoherence errors in qubits constitute a critical factor limiting quantum gate performance. To mitigate decoherence-induced gate infidelity, rapid implementation of quantum gates is essential. Here we propose a scheme for rapid controlled-Z (CZ) gate implementation through energy-level engineering, which leverages Rabi oscillations between the $\left|11\right\rangle$ state and the non-computational state in a tunable-coupler architecture. Numerical simulations achieved a $\mathrm{22~ns}$ nonadiabatic CZ gate with fidelity over $99.99\%$. We further investigated the performance of the CZ gate in the presence of anharmonicity offsets. The results demonstrate that a high-fidelity CZ gate with an error rate below $10^{-4}$ remains achievable even with finite anharmonicity variations. Furthermore, the detrimental impact of spectator qubits in different quantum states on the fidelity of CZ gate is effectively suppressed by incorporating a tunable coupler. This scheme exhibits potential for extending the circuit execution depth constrained by coherence time limitations.

quant-ph↗
arXiv2026

Scalable Second-order Riemannian Optimization for $K$-means Clustering

Clustering is a hard discrete optimization problem. Nonconvex approaches such as low-rank semidefinite programming (SDP) have recently demonstrated promising statistical and local algorithmic guarantees for cluster recovery. Due to the combinatorial structure of the $K$-means clustering problem, current relaxation algorithms struggle to balance their constraint feasibility and objective optimality, presenting tremendous challenges in computing the second-order critical points with rigorous guarantees. In this paper, we provide a new formulation of the $K$-means problem as a smooth unconstrained optimization over a submanifold and characterize its Riemannian structures to allow it to be solved using a second-order cubic-regularized Riemannian Newton algorithm. By factorizing the $K$-means manifold into a product manifold, we show how each Newton subproblem can be solved in linear time. Our numerical experiments show that the proposed method converges significantly faster than the state-of-the-art first-order nonnegative low-rank factorization method, while achieving similarly optimal statistical accuracy.

cs.LG↗
arXiv2026

Beyond Masks: Efficient, Flexible Diffusion Language Models via Deletion-Insertion Processes

While Masked Diffusion Language Models (MDLMs) relying on token masking and unmasking have shown promise in language modeling, their computational efficiency and generation flexibility remain constrained by the masking paradigm. In this paper, we propose Deletion-Insertion Diffusion language models (DID) that rigorously formulate token deletion and insertion as discrete diffusion processes, replacing the masking and unmasking processes in current MDLMs. DID improves training and inference efficiency by eliminating two major sources of computational overhead in MDLMs: the computations on non-informative 1) tokens inherent to the paradigm, and 2) tokens introduced in variable-length settings. Furthermore, DID offers greater flexibility by: 1) natively supporting variable-length sequences without requiring fixed-length padding, and 2) an intrinsic self-correction mechanism during generation due to insertion that dynamically adjusts token positions. To train DID, we design a score-based approach that assigns scores to token insertion operations and derive appropriate training objectives. The objectives involve subsequence counting problems, which we efficiently solve via a parallelized dynamic programming algorithm. Our experiments across fixed and variable-length settings demonstrate the advantage of DID over baselines of MDLMs and existing insertion-based LMs, in terms of modeling performance, sampling quality, and training/inference speed, without any hyperparameter tuning.

cs.CL↗
arXiv2026

Towards Realistic Detection Pipelines of Taiji: New Challenges in Data Analysis and High-Fidelity Simulations of Space-Based Gravitational Wave Antenna

Taiji, a Chinese space-based gravitational wave (GW) detection project, aims to explore the millihertz GW universe with unprecedented sensitivity. By observing astrophysical and cosmological sources, including Galactic binaries, massive black hole binaries, extreme mass-ratio inspirals, and stochastic gravitational wave backgrounds, etc., Taiji is expected to deliver transformative insights into astrophysics, cosmology, and fundamental physics. However, Taiji's data analysis faces unique challenges compared to ground-based detectors like LIGO-Virgo-KAGRA, such as the overlap of numerous signals, extended data durations, more rigorous accuracy requirements for the waveform templates, incompletely characterized noise spectra, non-stationary noises, and various data anomalies. Taking Taiji as a representative example, this paper reviews the data characteristics and data analysis challenges of space-based GW detection, and introduces the second round of Taiji Data Challenge, a collection of simulation datasets designed as a shared platform for resolving these critical issues. This platform distinguishes itself from previous works by the systematic integration of orbital dynamics based on a full drag-free and attitude control simulation, extended noise sources, more complicated and overlapping GW signals, second-generation time-delay interferometry, and the coupling effect of time-varying arm-lengths, etc. Concurrently released is the open-source toolkit Triangle, which offers the capabilities for customized simulation of signals, noises, and other instrumental effects. By taking a step further towards realistic detection, Taiji Data Challenge II and Triangle altogether serve as a new testbed, supporting the development of Taiji's global analysis and end-to-end pipelines, and ultimately bridging the gaps between observation and scientific objectives.

gr-qc↗