arXiv subjects

Pengfei Zhang

Publications and source records attributed to Pengfei Zhang.

At least 91 records · Page 5Linked to original sources

arXiv2026

Exact Relation Between Wehrl-Rényi Entropy and Many-Body Entanglement

Quantum entanglement is key to understanding correlations and emergent phenomena in quantum many-body systems. For $N$ qubits (distinguishable spin-$1/2$ particles) in a pure quantum state, many-body entanglement can be characterized by the purity of the reduced density matrix of a subsystem, defined as the trace of the square of this reduced density matrix. Nevertheless, this approach depends on the choice of subsystem. In this letter, we establish an exact relation between the Wehrl-Rényi entropy (WRE) $S_W^{(2)}$, which is the 2nd Rényi entropy of the Husimi function of the entire system, and the purities of all possible subsystems. Specifically, we prove the relation $e^{-S_W^{(2)}} = (6π)^{-N} \sum_A \mathrm{Tr}({{\hat ρ}_A}^2)$, where $A$ denotes a subsystem with reduced density matrix ${\hat ρ}_A$, and the summation runs over all $2^N$ possible subsystems. Furthermore, we show that the WRE can be experimentally measured via a concrete scheme. Therefore, the WRE is a subsystem-independent and experimentally measurable characterization of the overall entanglement in pure states of $N$ qubits. It can be applied to the study of strongly correlated spin systems, particularly those with all-to-all couplings that do not have a natural subsystem division, such as systems realized with natural atoms in optical tweezer arrays or superconducting quantum circuits. We also analytically derive the WRE for several representative many-body states, including Haar-random states, the Greenberger-Horne-Zeilinger (GHZ) state, and the W state.

quant-ph↗

arXiv2026

Distinguishing Coherent and Incoherent Errors in Multi-Round Time-Reversed Dynamics via Scramblons

Despite the rapid development of quantum science and technology, errors are inevitable and play a crucial role in quantum simulation and quantum computation. In quantum chaotic systems, coherent errors arising from imperfect Hamiltonian control and incoherent errors induced by coupling to the environment are both exponentially amplified during time evolution due to information scrambling. A fundamental question is how these two classes of errors imprint distinct signatures on the emergent irreversibility of many-body dynamics. In this Letter, we address this question by investigating multi-round time-reversed dynamics in the presence of both coherent and incoherent errors. By applying scramblon theory, we obtain closed-form expressions for the Loschmidt echo over different rounds of time-reversed evolution. For incoherent errors, the error accumulates linearly with the number of rounds, whereas coherent errors exhibit a crossover from quadratic to linear accumulation. These predictions are explicitly verified using the solvable Sachdev-Ye-Kitaev model. Our results provide a theoretical foundation for characterizing and calibrating coherent and incoherent errors in reversed dynamics, with particular relevance to nuclear magnetic resonance systems.

quant-ph↗

arXiv2025

Emergent Universality Class in Dissipative Quantum Systems with Dipole Moment Conservation

Understanding the non-equilibrium dynamics of quantum many-body systems remains one of the grand challenges of modern physics. In particular, increasing attention has been devoted to the emergence of non-equilibrium universality classes that have no equilibrium counterparts. A prominent example is the Kardar-Parisi-Zhang universality class realized in dissipative Bose-Einstein condensates. In this Letter, motivated by recent experimental advances, we investigate the universal dynamics of dissipative quantum systems with dipole moment conservation. We develop an effective field theory description, supported by a concrete quantum spin model, to capture the resulting universal behaviors. Our analysis unveils a novel strongly interacting non-equilibrium fixed point that governs the equal-time phase fluctuations in systems with either strong or weak dipole symmetries. Moreover, charge transport becomes subdiffusive in the presence of strong dipole symmetry, while it remains diffusive in the weakly symmetric case. Our results reveal the intricate interplay between kinetic constraints and dissipation in quantum many-body systems.

quant-ph↗

arXiv2025

An extragalactic gamma-ray binary formed in supernova 2022jli

On May 5 2022, a type Ic supernova (SN) explosion SN~2022jli was discovered. This SN showed additional optical emissions, which were found to exhibit 12.4-day periodic undulations and concordant periodic velocity shifts. These key features likely indicate a compact object in a binary system was formed. A faint $γ$-ray source was also detected at the position of the SN and upon checking the $γ$-ray photons' arrival times, it was revealed that the same 12.4-day periodicity was likely present. Here we report our detailed analysis results for the $γ$-ray source. Not only was the $γ$-ray emission detectable for $\sim$1.5\,years since the discovery time, but a strong modulation at period 12.5\,day was also clearly determined. Considering the newly formed compact object to be a neutron star or a stellar-mass black hole, the putative binary, having an orbital period of 12.5\,day, is likely the first extragalactic high-energy system detected. The system may serve as a valuable example for the formation of many such binaries observed in the Milky Way and nearby galaxies.

astro-ph.HE↗

arXiv2025

Language-driven Fine-grained Retrieval

Existing fine-grained image retrieval (FGIR) methods learn discriminative embeddings by adopting semantically sparse one-hot labels derived from category names as supervision. While effective on seen classes, such supervision overlooks the rich semantics encoded in category names, hindering the modeling of comparability among cross-category details and, in turn, limiting generalization to unseen categories. To tackle this, we introduce LaFG, a Language-driven framework for Fine-Grained Retrieval that converts class names into attribute-level supervision using large language models (LLMs) and vision-language models (VLMs). Treating each name as a semantic anchor, LaFG prompts an LLM to generate detailed, attribute-oriented descriptions. To mitigate attribute omission in these descriptions, it leverages a frozen VLM to project them into a vision-aligned space, clustering them into a dataset-wide attribute vocabulary while harvesting complementary attributes from related categories. Leveraging this vocabulary, a global prompt template selects category-relevant attributes, which are aggregated into category-specific linguistic prototypes. These prototypes supervise the retrieval model to steer

cs.CV↗

arXiv2025

Revisiting $γ$-Ray Orbital Modulation in the Redback Millisecond Pulsar PSR J2039-5617

PSR J2039-5617 is a redback millisecond pulsar binary system consisting of a compact star with a mass of 1.1-1.6 $M_\odot$ and a low-mass companion of 0.15-0.22 $M_\odot$. For this binary, we performed a timing analysis using 16 years of data from the Fermi Large Area Telescope, covering the period from 2008 August to 2024 October. Our analysis detected an orbital modulation with a period of 0.2279781 days at a significance level of $\sim4σ$, which is in good agreement with previous findings. However, unlike previous reports, we identified a transition in the orbital modulation around 2021 August, after which the orbital signal disappeared. We speculate that the system may be undergoing a transition from a rotation-powered to an accretion-powered state at this epoch. Additionally, we conducted the phase-resolved and spectral analyses, and in the phase-resolved results, we observed an anti-correlation between its $γ$-ray and X-ray emissions, which consistent with the predictions of high-energy radiation models for such systems. We provide some predictive discussions based on the results of $γ$-ray data analysis, and future Fermi-LAT observations will determine whether these predictions hold true.

astro-ph.HE↗

arXiv2025

High Dimensional Distributed Gradient Descent with Arbitrary Number of Byzantine Attackers

Adversarial attacks pose a major challenge to distributed learning systems, prompting the development of numerous robust learning methods. However, most existing approaches suffer from the curse of dimensionality, i.e. the error increases with the number of model parameters. In this paper, we make a progress towards high dimensional problems, under arbitrary number of Byzantine attackers. The cornerstone of our design is a direct high dimensional semi-verified mean estimation method. The idea is to identify a subspace with large variance. The components of the mean value perpendicular to this subspace are estimated using corrupted gradient vectors uploaded from worker machines, while the components within this subspace are estimated using auxiliary dataset. As a result, a combination of large corrupted dataset and small clean dataset yields significantly better performance than using them separately. We then apply this method as the aggregator for distributed learning problems. The theoretical analysis shows that compared with existing solutions, our method gets rid of $\sqrt{d}$ dependence on the dimensionality, and achieves minimax optimal statistical rates. Numerical results validate our theory as well as the effectiveness of the proposed method.

cs.LG↗

arXiv2025

Multimodal 3D Genome Pre-training

Deep learning techniques have driven significant progress in various analytical tasks within 3D genomics in computational biology. However, a holistic understanding of 3D genomics knowledge remains underexplored. Here, we propose MIX-HIC, the first multimodal foundation model of 3D genome that integrates both 3D genome structure and epigenomic tracks, which obtains unified and comprehensive semantics. For accurate heterogeneous semantic fusion, we design the cross-modal interaction and mapping blocks for robust unified representation, yielding the accurate aggregation of 3D genome knowledge. Besides, we introduce the first large-scale dataset comprising over 1 million pairwise samples of Hi-C contact maps and epigenomic tracks for high-quality pre-training, enabling the exploration of functional implications in 3D genomics. Extensive experiments show that MIX-HIC can significantly surpass existing state-of-the-art methods in diverse downstream tasks. This work provides a valuable resource for advancing 3D genomics research.

cs.LG↗

arXiv2025

Universal Relations in Long-range Quantum Spin Chains

Understanding the emergence of novel collective behaviors in strongly interacting systems lies at the heart of quantum many-body physics. Valuable insight comes from examining how few-body correlations manifest in many-body systems, embodying the ``from few to many'' philosophy. An intriguing example is the set of universal relations in ultracold atomic gases, which connect a wide range of observables to a single quantity known as the contact. In this Letter, we demonstrate that universal relations manifest in a distinct class of quantum many-body systems, long-range quantum spin chains, which belong to a completely new universality class. Using effective field theory and the operator product expansion, we establish connections between the asymptotic behavior of equal-time spin correlation functions, the dynamical structure factor, and the contact density. The theoretical predictions for equal-time correlators are explicitly verified through numerical simulations based on matrix product states. Our results could be readily tested in state-of-the-art trapped-ion systems.

cond-mat.quant-gas↗

arXiv2025

Training data membership inference via Gaussian process meta-modeling: a post-hoc analysis approach

Membership inference attacks (MIAs) test whether a data point was part of a model's training set, posing serious privacy risks. Existing methods often depend on shadow models or heavy query access, which limits their practicality. We propose GP-MIA, an efficient and interpretable approach based on Gaussian process (GP) meta-modeling. Using post-hoc metrics such as accuracy, entropy, dataset statistics, and optional sensitivity features (e.g. gradients, NTK measures) from a single trained model, GP-MIA trains a GP classifier to distinguish members from non-members while providing calibrated uncertainty estimates. Experiments on synthetic data, real-world fraud detection data, CIFAR-10, and WikiText-2 show that GP-MIA achieves high accuracy and generalizability, offering a practical alternative to existing MIAs.

cs.LG↗

arXiv2025

Hybrid Brownian SYK-Hubbard Model: from Spectral Function to Quantum Chaos

Understanding the emergence of complex correlations in strongly interacting systems remains a fundamental challenge in quantum many-body physics. One fruitful approach is to develop solvable toy models that encapsulate universal properties shared by realistic systems. In this work, we introduce the Brownian SYK-Hubbard model, which combines the all-to-all random interactions of the Sachdev-Ye-Kitaev (SYK) model with on-site Hubbard-type interactions. This hybrid construction enables the study of the interplay between nonlocal random dynamics and local correlation effects: (1) As the interaction strength increases, the single-particle spectrum exhibits a transition from a single peak to a two-peak structure, signaling the onset of Mottness. (2) The spectral form factor undergoes a sequence of dynamical transitions as the evolution time increases before reaching the plateau in the long-time limit under strong Hubbard interactions. (3) The out-of-time-order correlator is computed by summing a series of modified ladder diagrams, which determines the quantum Lyapunov exponent and reveals a violation of the bound on branching time. Our results establish a new analytically tractable platform for exploring the effects of Hubbard interactions in chaotic many-body systems.

quant-ph↗

arXiv2025

Contextual Gesture: Co-Speech Gesture Video Generation through Context-aware Gesture Representation

Co-speech gesture generation is crucial for creating lifelike avatars and enhancing human-computer interactions by synchronizing gestures with speech. Despite recent advancements, existing methods struggle with accurately identifying the rhythmic or semantic triggers from audio for generating contextualized gesture patterns and achieving pixel-level realism. To address these challenges, we introduce Contextual Gesture, a framework that improves co-speech gesture video generation through three innovative components: (1) a chronological speech-gesture alignment that temporally connects two modalities, (2) a contextualized gesture tokenization that incorporate speech context into motion pattern representation through distillation, and (3) a structure-aware refinement module that employs edge connection to link gesture keypoints to improve video generation. Our extensive experiments demonstrate that Contextual Gesture not only produces realistic and speech-aligned gesture videos but also supports long-sequence generation and video gesture editing applications, shown in Fig.1.

cs.CV↗

arXiv2025

Quantum many-body scars through the lens of correlation matrix

Quantum many-body scars (QMBS) -- rare eigenstates that evade thermalization -- are typically characterized by their low entanglement entropies compared to surrounding thermal eigenstates. However, due to finite-size effects in systems accessible via exact diagonalization, this measure can be ambiguous. To address this limitation, we propose using the correlation matrix spectrum as an alternative probe to identify QMBS. In cases of exact QMBS that either have known analytic expressions or are captured by various frameworks of QMBS, we find that the dimensionality of the null space of the correlation matrix -- an integer value, and thus immune to finite-size effects -- can qualitatively identify QMBS. Beyond serving as a diagnostic tool, the correlation matrix method enables the manipulation of the QMBS subspace. For approximate QMBS, such as those in the PXP model, we observe that the correlation matrix spectrum features numerous approximate zero eigenvalues, thereby distinguishing these states. We demonstrate the effectiveness and utility of this method with several paradigmatic QMBS examples.

cond-mat.quant-gas↗

arXiv2025

Designing Shadow Tomography Protocols by Natural Language Processing

Quantum circuits form a foundational framework in quantum science, enabling the description, analysis, and implementation of quantum computations. However, designing efficient circuits, typically constructed from single- and two-qubit gates, remains a major challenge for specific computational tasks. In this work, we introduce a novel artificial intelligence-driven protocol for quantum circuit design, benchmarked using shadow tomography for efficient quantum state readout. Inspired by techniques from natural language processing (NLP), our approach first selects a compact gate dictionary by optimizing the entangling power of two-qubit gates. We identify the iSWAP gate as a key element that significantly enhances sample efficiency, resulting in a minimal gate set of {I, SWAP, iSWAP}. Building on this, we implement a recurrent neural network trained via reinforcement learning to generate high-performing quantum circuits. The trained model demonstrates strong generalization ability, discovering efficient circuit architectures with low sample complexity beyond the training set. Our NLP-inspired framework offers broad potential for quantum computation, including extracting properties of logical qubits in quantum error correction.

quant-ph↗

arXiv2025

Combinatorial optimization enhanced by shallow quantum circuits with 104 superconducting qubits

A pivotal task for quantum computing is to speed up solving problems that are both classically intractable and practically valuable. Among these, combinatorial optimization problems have attracted tremendous attention due to their broad applicability and natural fitness to Ising Hamiltonians. Here we propose a quantum sampling strategy, based on which we design an algorithm for accelerating solving the ground states of Ising model, a class of NP-hard problems in combinatorial optimization. The algorithm employs a hybrid quantum-classical workflow, with a shallow-circuit quantum sampling subroutine dedicated to navigating the energy landscape. Using up to 104 superconducting qubits, we demonstrate that this algorithm outputs favorable solutions against even a highly-optimized classical simulated annealing (SA) algorithm. Furthermore, we illustrate the path toward quantum speedup based on the time-to-solution metric against SA running on a single-core CPU with just 100 qubits. Our results indicate a promising alternative to classical heuristics for combinatorial optimization, a paradigm where quantum advantage might become possible on near-term superconducting quantum processors with thousands of qubits and without the assistance of error correction.

quant-ph↗

arXiv2025

Heat Diffusion Models -- Interpixel Attention Mechanism

Denoising Diffusion Probabilistic Models (DDPM) process images as a whole. Since adjacent pixels are highly likely to belong to the same object, we propose the Heat Diffusion Model (HDM) to further preserve image details and generate more realistic images. HDM essentially is a DDPM that incorporates an attention mechanism between pixels. In HDM, the discrete form of the two-dimensional heat equation is integrated into the diffusion and generation formulas of DDPM, enabling the model to compute relationships between neighboring pixels during image processing. Our experiments demonstrate that HDM can generate higher-quality samples compared to models such as DDPM, Consistency Diffusion Models (CDM), Latent Diffusion Models (LDM), and Vector Quantized Generative Adversarial Networks (VQGAN).

cs.CV↗

arXiv2025

Towards Controllable Speech Synthesis in the Era of Large Language Models: A Systematic Survey

Text-to-speech (TTS) has advanced from generating natural-sounding speech to enabling fine-grained control over attributes like emotion, timbre, and style. Driven by rising industrial demand and breakthroughs in deep learning, e.g., diffusion and large language models (LLMs), controllable TTS has become a rapidly growing research area. This survey provides the first comprehensive review of controllable TTS methods, from traditional control techniques to emerging approaches using natural language prompts. We categorize model architectures, control strategies, and feature representations, while also summarizing challenges, datasets, and evaluations in controllable TTS. This survey aims to guide researchers and practitioners by offering a clear taxonomy and highlighting future directions in this fast-evolving field. One can visit https://github.com/imxtx/awesome-controllabe-speech-synthesis for a comprehensive paper list and updates.

cs.CL↗

arXiv2025

$p$-orbital self-organization of ultracold atoms coupled to optical cavities

Atoms coupled to optical cavities provide a novel platform for understanding high-orbital exotic phenomena in strongly correlated materials. In this study, we investigate strongly correlated ultracold bosonic gases that are coupled to two orthogonally arranged optical cavities and driven by a blue-detuned running-wave laser field. Our results demonstrate that atoms initially in the $s$-orbital state can be scattered into $p_x$- and $p_y$-orbital states in either a symmetric or asymmetric manner, depending on the frequencies of the two cavities. For the symmetric configuration, we observe that atoms are scattered into the $p_x$- and $p_y$-orbitals equally. In the asymmetric case, photons emitted into one cavity mode suppress the scattering into the orthogonal mode. Notably, the coupling of atoms with multiple cavity modes leads to the emergence of high-orbital self-organized phases, accompanied by orbital-density waves that break different symmetries.

cond-mat.quant-gas↗