arXiv subjects

Jing Zou

Publications and source records attributed to Jing Zou.

16 recordsLinked to original sources

arXiv2026

ContiguousKV: Accelerating LLM Prefill with Granularity-Aligned KV Cache Management

Efficiently serving Large Language Models (LLMs) with persistent Prefix Key-Value (KV) Cache is critical for applications like conversational search and multi-turn dialogue. Serving a request requires loading the pre-computed prefix KV cache and generating the first token, defined as the Re-Prefill Phase. Offloading this shared prefix cache to secondary storage is essential for memory scalability. Re-Prefill with offloading suffers from severe I/O bottlenecks in two aspects. First, semantic-aware KV cache pruning algorithms select important tokens in fine granularity, while systems manage I/O in coarse, fixed-size blocks, causing severe read amplification. Second, the sequential dependency between identifying important tokens and loading KV cache creates idle I/O and compute bubbles, under-utilizing system resources. This paper proposes \textit{ContiguousKV}, a high-performance prefix KV cache offloading system that bridges algorithmic semantics with I/O efficiency to accelerate the Re-Prefill phase. We first introduce \textit{ContiguousChunk}, a unified data management granularity that aligns KV cache pruning with I/O operations. All the mechanisms critical for I/O performance are performed at the granularity of ContiguousChunk, thereby eliminating read amplification. By exploiting the high similarity in important ContiguousChunk indices across layers, we propose intra- and inter-period asynchronous prefetching to break the sequential dependency between I/O and compute, effectively eliminating idle bubbles. Finally, we propose attention-guided cache management to retain semantically critical prefix data in memory. Evaluations on Qwen2.5 series models show that ContiguousKV achieves a 3.85x speedup in the Re-Prefill phase over the state-of-the-art offloading system IMPRESS, while maintaining high output quality.

cs.OS↗

arXiv2026

Upper limits on microhertz gravitational waves from supermassive black-hole binaries using PSR J1909-3744 data from the second IPTA data release

We present the results of a search for gravitational waves (GWs) from individual sources using high-cadence observations of PSR J1909$-$3744 obtained during an intensive observing campaign with the International Pulsar Timing Array second data release (IPTA-DR2) between July 2010 and November 2012. The observations, conducted at three different radio frequencies with the Nan\c{c}ay Radio Telescope (NRT) and Parkes Telescope (PKS) and five frequencies with the Green Bank Telescope (GBT), enabled precise corrections for dispersion measure effects and scattering variations. After these corrections, the timing residuals showed an unmodeled periodic noise component with an amplitude of 340 ns. Our analysis yields upper limits on the GW strain from individual sources, constraining it to be below $1.9 \times 10^{-14}$ at 71 nHz and $2.3 \times 10^{-13}$ at 1 \textmu Hz for average sky locations, while for optimal source locations the limits improve to $6.2 \times 10^{-15}$ and $8.9 \times 10^{-14}$ at the same frequencies, respectively. Our new limits are about a factor of 1.52 more stringent than those of Perera et al. based on an earlier EPTA data.

astro-ph.HE↗

arXiv2025

Study of Four nulling pulsars with FAST

We present an analysis of 4 nulling pulsars with the Five-hundred-meter Aperture Spherical radio Telescope (FAST). For PSR J1649+2533, our results suggest mode changing rather than subpulse drifting as previously reported at lower frequencies. For PSR J1752+2359, we confirm its quasi-periodic switching between distinct emission states, but further show that the so-called "quasi-null" or "RRAT-like" state actually consists of persistent low-level emission superposed with occasional bright pulses. For PSR J1819+1305, our data confirm the modulation reported earlier, while additional weaker features are also seen. For PSR J1916+1023, we detect both nulling and subpulse drifting, but find no clear evidence of direct interaction between them. These results provide new insights into the diverse manifestations of pulsar nulling, highlight the capability of FAST to detect subtle emission states, and add to the growing body of work on pulsar emission variability.

astro-ph.HE↗

arXiv2025

DARD: Dice Adversarial Robustness Distillation against Adversarial Attacks

Deep learning models are vulnerable to adversarial examples, posing critical security challenges in real-world applications. While Adversarial Training (AT ) is a widely adopted defense mechanism to enhance robustness, it often incurs a trade-off by degrading performance on unperturbed, natural data. Recent efforts have highlighted that larger models exhibit enhanced robustness over their smaller counterparts. In this paper, we empirically demonstrate that such robustness can be systematically distilled from large teacher models into compact student models. To achieve better performance, we introduce Dice Adversarial Robustness Distillation (DARD), a novel method designed to transfer robustness through a tailored knowledge distillation paradigm. Additionally, we propose Dice Projected Gradient Descent (DPGD), an adversarial example generalization method optimized for effective attack. Our extensive experiments demonstrate that the DARD approach consistently outperforms adversarially trained networks with the same architecture, achieving superior robustness and standard accuracy.

cs.LG↗

arXiv2025

FAST Observations of the Microstructure in Interpulse Pulsars

In this study, we investigate the microstructure properties of four pulsars (PSRs J0953+0755 (B0950+08), J0627+0706, J0826+2637 (B0823+26) and J1946+1805 (B1944+17)) using the Five-hundred-meter Aperture Spherical radio Telescope (FAST), with particular emphasis on identifying microstructure within interpulse (IP). Through the application of autocorrelation function (ACF) analysis and fast Fourier transform (FFT) techniques, we have systematically examined the periodicity of microstructure in these pulsars. Our findings represent the first successful detection of microstructure within IP. Furthermore, we conducted a comprehensive statistical analysis comparing the characteristic timescales ($\tau_{\mu}$) and the characteristic periods $P_{\mu}$ of quasi-periodic microstructure between the main pulse (MP) and IP, and our results indicate that the $\tau_{\mu}$ and $P_{\mu}$ of microstructure across components appear consistent within measurement errors for PSR J0627+0706, but microstructure in IP are relatively smaller than those in MP for PSR J0953+0755. Furthermore, the relationship between $P_{\mu}$ of microstructure and the rotation period in neutron star populations was reconfirmed: $P_{\mu}(\text{ms})=(1.337\pm0.114)\times P(\text{s})^{(1.063\pm0.038)}$.

astro-ph.HE↗

arXiv2025

CorBenchX: Large-Scale Chest X-Ray Error Dataset and Vision-Language Model Benchmark for Report Error Correction

AI-driven models have shown great promise in detecting errors in radiology reports, yet the field lacks a unified benchmark for rigorous evaluation of error detection and further correction. To address this gap, we introduce CorBenchX, a comprehensive suite for automated error detection and correction in chest X-ray reports, designed to advance AI-assisted quality control in clinical practice. We first synthesize a large-scale dataset of 26,326 chest X-ray error reports by injecting clinically common errors via prompting DeepSeek-R1, with each corrupted report paired with its original text, error type, and human-readable description. Leveraging this dataset, we benchmark both open- and closed-source vision-language models,(e.g., InternVL, Qwen-VL, GPT-4o, o4-mini, and Claude-3.7) for error detection and correction under zero-shot prompting. Among these models, o4-mini achieves the best performance, with 50.6 % detection accuracy and correction scores of BLEU 0.853, ROUGE 0.924, BERTScore 0.981, SembScore 0.865, and CheXbertF1 0.954, remaining below clinical-level accuracy, highlighting the challenge of precise report correction. To advance the state of the art, we propose a multi-step reinforcement learning (MSRL) framework that optimizes a multi-objective reward combining format compliance, error-type accuracy, and BLEU similarity. We apply MSRL to QwenVL2.5-7B, the top open-source model in our benchmark, achieving an improvement of 38.3% in single-error detection precision and 5.2% in single-error correction over the zero-shot baseline.

cs.AI↗

arXiv2025

Semiparametric Causal Discovery and Inference with Invalid Instruments

Learning causal relationships among a set of variables, as encoded by a directed acyclic graph, from observational data is complicated by the presence of unobserved confounders. Instrumental variables (IVs) are a popular remedy for this issue, but most existing methods either assume the validity of all IVs or postulate a specific form of relationship, such as a linear model, between the primary variables and the IVs. To overcome these limitations, we introduce a partially linear structural equation model for causal discovery and inference that accommodates potentially invalid IVs and allows for general dependence of the primary variables on the IVs. We establish identification under this semiparametric model by constructing surrogate valid IVs, and develop a finite-sample procedure for estimating the causal structures and effects. Theoretically, we show that our procedure consistently learns the causal structures, yields asymptotically normal estimates, and effectively controls the false discovery rate in edge recovery. Simulation studies demonstrate the superiority of our method over existing competitors, and an application to inferring gene regulatory networks in Alzheimer's disease illustrates its usefulness.

stat.ME↗

arXiv2025

Long-Term Timing Analysis of PSR J1741-3016: Efficient Noise Characterization Using PINT

The stable rotation of young pulsars is often interrupted by two non-deterministic phenomena: glitches and red timing noise. Timing noise provides insights into plasma and nuclear physics under extreme conditions. The framework leverages rotational symmetry in pulsar spin-down models and temporal symmetry in noise processes to achieve computational efficiency, aligning with the journal's focus on symmetry principles in physical systems. In this paper, we apply a novel frequentist framework developed within the PINT software package (v0.9.8) to analyze single-pulsar noise processes. Using 17.5 years of pulse time-of-arrival (TOA) data for the young pulsar PSR J1741-3016, observed with the Nanshan 26 m radio telescope, we investigate its timing properties. In this study, we employed the Downhill Weighted Least-Squares Fitter to estimate the pulsar's spin parameters and position. The Akaike Information Criterion (AIC) was used for model parameter selection. The results obtained with PINT were compared to those from ENTERPRISE and TEMPONEST, two Bayesian-based frameworks. We demonstrate that PINT achieves comparable results with significantly reduced computational costs. Additionally, the adequacy of the noise model can be readily verified through visual inspection tools. Future research will utilize this framework to analyze timing noise across a large sample of young pulsars.

astro-ph.HE↗

arXiv2025

Searching for gravitational-wave bursts from cosmic string cusps with the Parkes Pulsar Timing Array third data release

Pulsar timing arrays (PTAs) are designed to detect nanohertz-frequency gravitational waves (GWs). Since GWs are anticipated from cosmic strings, PTAs offer a viable approach to testing their existence. We present the results of the first Bayesian search for gravitational-wave bursts from cosmic string cusps (GWCS) using the third PPTA data release for 30 millisecond pulsars. In this data collection, we find no evidence for GWCS signals. We compare a model with a GWCS signal to one with only noise, including a common spatially uncorrelated red noise (CURN), and find that our data is more consistent with the noise-only model.We then establish upper limits on the strain amplitude of GWCS at the pulsar term, based on the analysis of 30 millisecond pulsars, after finding no compelling evidence. We find the addition of a CURN with different spectral indices into the noise model has a negligible impact on the upper limits. And the upper limit range of the amplitude of the pulsar-term GWCS is concentrated between 10^{-12} and 10^{-11}. Finally, we set upper limits on the amplitude of GWCS events, parametrized by width and event epoch, for a single pulsar PSR J1857+0943. Moreover, we derive upper limit on the cosmic string tension as a function of burst width and compare it with previous results.

gr-qc↗

arXiv2024

The Lipschitz continuity of the solution to branched rough differential equations

Based on an isomorphism between Grossman Larson Hopf algebra and Tensor Hopf algebra, we apply a sub-Riemannian geometry technique to branched rough differential equations and obtain the explicit Lipschitz continuity of the solution with respect to the initial value, the vector field and the driving rough path.

math.PR↗

arXiv2024

MMR-Mamba: Multi-Modal MRI Reconstruction with Mamba and Spatial-Frequency Information Fusion

Multi-modal MRI offers valuable complementary information for diagnosis and treatment; however, its utility is limited by prolonged scanning times. To accelerate the acquisition process, a practical approach is to reconstruct images of the target modality, which requires longer scanning times, from under-sampled k-space data using the fully-sampled reference modality with shorter scanning times as guidance. The primary challenge of this task is comprehensively and efficiently integrating complementary information from different modalities to achieve high-quality reconstruction. Existing methods struggle with this: 1) convolution-based models fail to capture long-range dependencies; 2) transformer-based models, while excelling in global feature modeling, struggle with quadratic computational complexity. To address this, we propose MMR-Mamba, a novel framework that thoroughly and efficiently integrates multi-modal features for MRI reconstruction, leveraging Mamba's capability to capture long-range dependencies with linear computational complexity while exploiting global properties of the Fourier domain. Specifically, we first design a Target modality-guided Cross Mamba (TCM) module in the spatial domain, which maximally restores the target modality information by selectively incorporating relevant information from the reference modality. Then, we introduce a Selective Frequency Fusion (SFF) module to efficiently integrate global information in the Fourier domain and recover high-frequency signals for the reconstruction of structural details. Furthermore, we devise an Adaptive Spatial-Frequency Fusion (ASFF) module, which mutually enhances the spatial and frequency domains by supplementing less informative channels from one domain with corresponding channels from the other.

eess.IV↗

arXiv2024

Revitalizing Multivariate Time Series Forecasting: Learnable Decomposition with Inter-Series Dependencies and Intra-Series Variations Modeling

Predicting multivariate time series is crucial, demanding precise modeling of intricate patterns, including inter-series dependencies and intra-series variations. Distinctive trend characteristics in each time series pose challenges, and existing methods, relying on basic moving average kernels, may struggle with the non-linear structure and complex trends in real-world data. Given that, we introduce a learnable decomposition strategy to capture dynamic trend information more reasonably. Additionally, we propose a dual attention module tailored to capture inter-series dependencies and intra-series variations simultaneously for better time series forecasting, which is implemented by channel-wise self-attention and autoregressive self-attention. To evaluate the effectiveness of our method, we conducted experiments across eight open-source datasets and compared it with the state-of-the-art methods. Through the comparison results, our Leddam (LEarnable Decomposition and Dual Attention Module) not only demonstrates significant advancements in predictive performance, but also the proposed decomposition strategy can be plugged into other methods with a large performance-boosting, from 11.87% to 48.56% MSE error degradation.

cs.LG↗

arXiv2023

Learning Homeomorphic Image Registration via Conformal-Invariant Hyperelastic Regularisation

Deformable image registration is a fundamental task in medical image analysis and plays a crucial role in a wide range of clinical applications. Recently, deep learning-based approaches have been widely studied for deformable medical image registration and achieved promising results. However, existing deep learning image registration techniques do not theoretically guarantee topology-preserving transformations. This is a key property to preserve anatomical structures and achieve plausible transformations that can be used in real clinical settings. We propose a novel framework for deformable image registration. Firstly, we introduce a novel regulariser based on conformal-invariant properties in a nonlinear elasticity setting. Our regulariser enforces the deformation field to be smooth, invertible and orientation-preserving. More importantly, we strictly guarantee topology preservation yielding to a clinical meaningful registration. Secondly, we boost the performance of our regulariser through coordinate MLPs, where one can view the to-be-registered images as continuously differentiable entities. We demonstrate, through numerical and visual experiments, that our framework is able to outperform current techniques for image registration.

eess.IV↗

arXiv2022

AMDET: Attention based Multiple Dimensions EEG Transformer for Emotion Recognition

Affective computing is an important branch of artificial intelligence, and with the rapid development of brain computer interface technology, emotion recognition based on EEG signals has received broad attention. It is still a great challenge to effectively explore the multi-dimensional information in the EEG data in spite of a large number of deep learning methods. In this paper, we propose a deep model called Attention-based Multiple Dimensions EEG Transformer (AMDET), which can exploit the complementarity among the spectral-spatial-temporal features of EEG data by employing the multi-dimensional global attention mechanism. We transformed the original EEG data into 3D temporal-spectral-spatial representations and then the AMDET would use spectral-spatial transformer encoder layer to extract effective features in the EEG signal and concentrate on the critical time frame with a temporal attention layer. We conduct extensive experiments on the DEAP, SEED, and SEED-IV datasets to evaluate the performance of AMDET and the results outperform the state-of-the-art baseline on three datasets. Accuracy rates of 97.48%, 96.85%, 97.17%, 87.32% were achieved in the DEAP-Arousal, DEAP-Valence, SEED, and SEED-IV datasets, respectively. We also conduct extensive experiments to explore the possible brain regions that influence emotions and the coupling of EEG signals. AMDET can perform as well even with few channels which are identified by visualizing what learned model focus on. The accuracy could achieve over 90% even with only eight channels and it is of great use and benefit for practical applications.

eess.SP↗

arXiv2005

A Sublinear Algorithm of Sparse Fourier Transform for Nonequispaced Data

We present a sublinear randomized algorithm to compute a sparse Fourier transform for nonequispaced data. Suppose a signal S is known to consist of N equispaced samples, of which only L<N are available. If the ratio p=L/N is not close to 1, the available data are typically non-equispaced samples. Then our algorithm reconstructs a near-optimal B-term representation R with high probability 1-delta, in time and space poly(B,log(L),log p, log(1/delta), epsilon^{-1}, such that ||S-R||^2 < (1+epsilon) ||S-R_{opt}^B||^2, where R_{opt}^B is the optimal B-term Fourier representation of signal S. The sublinear poly(logL) time is compared to the superlinear O(Nlog N+L) time requirement of the present best known Inverse Nonequispaced Fast Fourier Transform (INFFT) algorithms. Numerical experiments support the advantage in speed of our algorithm over other methods for sparse signals: it already outperforms INFFT for large but realistic size N and works well even in the situation of a large percentage of missing data and in the presence of noise.

math.NA↗

arXiv2004

Theoretical and Experimental Analysis of a Randomized Algorithm for Sparse Fourier Transform Analysis

We analyze a sublinear RAlSFA (Randomized Algorithm for Sparse Fourier Analysis) that finds a near-optimal B-term Sparse Representation R for a given discrete signal S of length N, in time and space poly(B,log(N)), following the approach given in \cite{GGIMS}. Its time cost poly(log(N)) should be compared with the superlinear O(N log N) time requirement of the Fast Fourier Transform (FFT). A straightforward implementation of the RAlSFA, as presented in the theoretical paper \cite{GGIMS}, turns out to be very slow in practice. Our main result is a greatly improved and practical RAlSFA. We introduce several new ideas and techniques that speed up the algorithm. Both rigorous and heuristic arguments for parameter choices are presented. Our RAlSFA constructs, with probability at least 1-delta, a near-optimal B-term representation R in time poly(B)log(N)log(1/delta)/ epsilon^{2} log(M) such that ||S-R||^{2}<=(1+epsilon)||S-R_{opt}||^{2}. Furthermore, this RAlSFA implementation already beats the FFTW for not unreasonably large N. We extend the algorithm to higher dimensional cases both theoretically and numerically. The crossover point lies at N=70000 in one dimension, and at N=900 for data on a N*N grid in two dimensions for small B signals where there is noise.

math.NA↗