arXiv subjects

Song Li

Publications and source records attributed to Song Li.

At least 37 records · Page 2Linked to original sources

arXiv2025

Large Language Models Cannot Reliably Detect Vulnerabilities in JavaScript: The First Systematic Benchmark and Evaluation

Researchers have proposed numerous methods to detect vulnerabilities in JavaScript, especially those assisted by Large Language Models (LLMs). However, the actual capability of LLMs in JavaScript vulnerability detection remains questionable, necessitating systematic evaluation and comprehensive benchmarks. Unfortunately, existing benchmarks suffer from three critical limitations: (1) incomplete coverage, such as covering a limited subset of CWE types; (2) underestimation of LLM capabilities caused by unreasonable ground truth labeling; and (3) overestimation due to unrealistic cases such as using isolated vulnerable files rather than complete projects. In this paper, we introduce, for the first time, three principles for constructing a benchmark for JavaScript vulnerability detection that directly address these limitations: (1) comprehensiveness, (2) no underestimation, and (3) no overestimation. Guided by these principles, we propose FORGEJS, the first automatic benchmark generation framework for evaluating LLMs' capability in JavaScript vulnerability detection. Then, we use FORGEJS to construct ARENAJS-the first systematic benchmark for LLM-based JavaScript vulnerability detection-and further propose JUDGEJS, an automatic evaluation framework. We conduct the first systematic evaluation of LLMs for JavaScript vulnerability detection, leveraging JUDGEJS to assess seven popular commercial LLMs on ARENAJS. The results show that LLMs not only exhibit limited reasoning capabilities, but also suffer from severe robustness defects, indicating that reliable JavaScript vulnerability detection with LLMs remains an open challenge.

cs.CR↗

arXiv2025

LongCat-Flash-Omni Technical Report

We introduce LongCat-Flash-Omni, a state-of-the-art open-source omni-modal model with 560 billion parameters, excelling at real-time audio-visual interaction. By adopting a curriculum-inspired progressive training strategy that transitions from simpler to increasingly complex modality sequence modeling tasks, LongCat-Flash-Omni attains comprehensive multimodal capabilities while maintaining strong unimodal capability. Building upon LongCat-Flash, which adopts a high-performance Shortcut-connected Mixture-of-Experts (MoE) architecture with zero-computation experts, LongCat-Flash-Omni integrates efficient multimodal perception and speech reconstruction modules. Despite its immense size of 560B parameters (with 27B activated), LongCat-Flash-Omni achieves low-latency real-time audio-visual interaction. For training infrastructure, we developed a modality-decoupled parallelism scheme specifically designed to manage the data and model heterogeneity inherent in large-scale multimodal training. This innovative approach demonstrates exceptional efficiency by sustaining over 90% of the throughput achieved by text-only training. Extensive evaluations show that LongCat-Flash-Omni achieves state-of-the-art performance on omni-modal benchmarks among open-source models. Furthermore, it delivers highly competitive results across a wide range of modality-specific tasks, including text, image, and video understanding, as well as audio understanding and generation. We provide a comprehensive overview of the model architecture design, training procedures, and data strategies, and open-source the model to foster future research and development in the community.

cs.MM↗

arXiv2025

LongCat-Audio-Codec: An Audio Tokenizer and Detokenizer Solution Designed for Speech Large Language Models

This paper presents LongCat-Audio-Codec, an audio tokenizer and detokenizer solution designed for industrial grade end-to-end speech large language models. By leveraging a decoupled model architecture and a multistage training strategy, LongCat-Audio-Codec exhibits robust semantic modeling capabilities, flexible acoustic feature extraction capabilities, and low-latency streaming synthesis capabilities. It encodes speech at an ultra-low frame rate of 16.67 Hz, with a minimum bitrate of 0.43 kbps and a maximum bitrate of 0.87 kbps. Evaluation results demonstrate that LongCat-Audio-Codec achieves strong speech intelligibility and is capable of synthesizing highquality speech at low bitrate, thus effectively balancing coding efficiency and decoding quality. The inference code and model checkpoints of LongCat-Audio-Codec are available at: https://github.com/meituan-longcat/LongCat-Audio-Codec.

eess.AS↗

arXiv2025

Unraveling dark Higgs mechanism via dark photon production at an e^+ e^- collider

In the phenomenology study of dark photon, its mass origin is usually not under concern. However, in theory construction its mass is often generated via a dark Higgs mechanism, which leads to the presence of a light (non-decoupled) dark Higgs particle. In this work, we study the impact of such a dark Higgs particle in the collider detection of the dark photon. We focus on the process of final state dark photon radiating dark Higgs, which is called dark final state radiation (FSR). Considering the effects on both the signal cross section and the distribution of the missing mass square, the invisible dark photon search at BaBar is reanalyzed and a new exclusion limit for invisible dark photon is presented.

hep-ph↗

arXiv2025

On a matrix constrained CKP hierarchy

The algebraic structures of integrable hierarchies play an important role in the study of soliton equations. In this paper, we use splitting theory to give a matrix representation of a constrained CKP hierarchy, which can be considered as a generalization of the $\hat{A}_{2n}^{(2)}$-KdV hierarchy and the constrained KP hierarchy. An equivalent construction in terms of the pseudo-differential operator is discussed. Darboux transformations, scaling transformation and tau functions $\ln τ_f$ for this constrained hierarchy are studied. Moreover, we present formulas for the Virasoro vector fields on $\ln τ_f$ for the $\hat{A}_{2 n}^{(2)}$-KdV hierarchy.

nlin.SI↗

arXiv2025

Interpreting deep learning-based stellar mass estimation via causal analysis and mutual information decomposition

End-to-end deep learning models fed with multi-band galaxy images are powerful data-driven tools used to estimate galaxy physical properties in the absence of spectroscopy. However, due to a lack of interpretability and the associational nature of such models, it is difficult to understand how the information that is included in addition to integrated photometry (e.g., morphology) contributes to the estimation task. Improving our understanding in this field would enable further advances into unraveling the physical connections among galaxy properties and optimizing data exploitation. Therefore, our work is aimed at interpreting the deep learning-based estimation of stellar mass via two interpretability techniques: causal analysis and mutual information decomposition. The former reveals the causal paths between multiple variables beyond nondirectional statistical associations, while the latter quantifies the multicomponent contributions (i.e., redundant, unique, and synergistic) of different input data to the stellar mass estimation. Using data from the Sloan Digital Sky Survey (SDSS) and the Wide-field Infrared Survey Explorer (WISE), we obtained meaningful results that provide physical interpretations for image-based models. Our work demonstrates the gains from combining deep learning with interpretability techniques, and holds promise in promoting more data-driven astrophysical research (e.g., astrophysical parameter estimations and investigations on complex multivariate physical processes).

astro-ph.IM↗

arXiv2025

Stable Phase Retrieval: Optimal Rates in Poisson and Heavy-tailed Models

We investigate stable recovery guarantees for phase retrieval under two realistic and challenging noise models: the Poisson model and the heavy-tailed model. Our analysis covers both nonconvex least squares (NCVX-LS) and convex least squares (CVX-LS) estimators. For the Poisson model, we demonstrate that in the high-energy regime where the true signal $pmb{x}$ exceeds a certain energy threshold, both estimators achieve a signal-independent, minimax optimal error rate $\mathcal{O}(\sqrt{\frac{n}{m}})$, with $n$ denoting the signal dimension and $m$ the number of sampling vectors. In contrast, in the low-energy regime, the NCVX-LS estimator attains an error rate of $\mathcal{O}(\|\pmb{x}\|^{1/4}_2\cdot(\frac{n}{m})^{1/4})$, which decreases as the energy of signal $\pmb{x}$ diminishes and remains nearly optimal with respect to the oversampling ratio. This demonstrates a signal-energy-adaptive behavior in the Poisson setting. For the heavy-tailed model with noise having a finite $q$-th moment ($q>2$), both estimators attain the minimax optimal error rate $\mathcal{O}( \frac{\| ξ\|_{L_q}}{\| \pmb{x} \|_2} \cdot \sqrt{\frac{n}{m}} )$ in the high-energy regime, while the NCVX-LS estimator further achieves the minimax optimal rate $\mathcal{O}( \sqrt{\|ξ\|_{L_q}}\cdot (\frac{n}{m})^{1/4} )$ in the low-energy regime. Our analysis builds on two key ideas: the use of multiplier inequalities to handle noise that may exhibit dependence on the sampling vectors, and a novel interpretation of Poisson noise as sub-exponential in the high-energy regime yet heavy-tailed in the low-energy regime. These insights form the foundation of a unified analytical framework, which we further apply to a range of related problems, including sparse phase retrieval, low-rank PSD matrix recovery, and random blind deconvolution.

math.ST↗

arXiv2025

Deformation-Driven Enhancement of Spin Defect Emission in Hexagonal Boron Nitride

The negatively charged boron vacancy (VB-) in hexagonal boron nitride (hBN) has been extensively investigated as it offers a novel playground for two-dimensional quantum sensing, with ultimate proximity to target samples. However, its practical sensitivity is limited by the intrinsically weak photoluminescence of the spin ensemble. Here, we report a photoluminescence enhancement of up to 30 times from VB- centers in suspended regions of hBN compared to those in substrate-supported areas. The key spin properties, such as the optically detected magnetic resonance (ODMR) contrast and linewidth, as well as the spin lifetime, of the VB- centers in this region are well preserved. Detailed investigations, including measurements of zero-field ODMR, Raman spectroscopy, and Kelvin probe force microscopy, reveal a correlation between emission enhancement and local deformation in the sample. It is concluded that the suspended regions exhibit higher local deformation compared to the supported areas, breaking the local symmetry and thereby activating otherwise forbidden or weak optical transitions of the VB- centers.

quant-ph↗

arXiv2025

Optimizing the depth-dependent nitrogen-vacancy center quantum sensor in diamane

Negatively charged nitrogen-vacancy (NV) center in diamond is the representative solid state defect qubit for quantum information science, offering long coherence time at room temperature. To achieve high sensitivity and spatial resolution, shallow NV center near the surface are preferred. However, shallow NV center suffers from surface states and spin noise which reduce the photostability and coherence time. In this work, we systematically study the NV center in recently reported two-dimensional diamond, known as diamane--using first-principles calculations. We show that the quantum confinement in finite-layer diamane, with appropriate surface passivation, could significantly modify the band structure. In particular, we identify oxygen surface termination capable of hosting NV center in diamane while optimizing photostability compared to bulk diamond. Furthermore, layer-dependent NV center demonstrates tunable zero-phonon-line and suppressed phonon side band, while retaining long coherence time. Our findings highlight diamane as a promising platform for NV-based quantum information processing with improved optical properties and stability

cond-mat.mtrl-sci↗

arXiv2025

Indication for a compact object next to a LIGO-Virgo binary black hole merger

The astrophysical origin of binary black hole (BBH) mergers remains uncertain, although many events have been observed by the LIGO-Virgo-KAGRA network. Such mergers are potentially originated in the vicinity of massive black holes (MBHs). GW190814, due to its secondary mass and mass ratio being beyond the expectations of isolated stellar evolution theories, is a promising event that has occurred in an active galactic nucleus (AGN) disk. In this model, a compact object resides in the vicinity of a merging BBH. Here we report multiple pieces of evidence suggesting that GW190814 is a BBH merging near a compact object. The orbital motion of BBHs around a third body produces a line-of-sight acceleration (LSA) and induces a varying Doppler shift. Using a waveform template that considers LSA, we perform Bayesian inference on a few BBH events with a high signal-to-noise ratio in the gravitational-wave (GW) transient catalog. Compared to the model for isolated BBH mergers, we obtain significantly higher network signal-to-noise ratios for GW190814 with the inclusion of LSA, constraining the LSA to $a = 0.0015^{+0.0008}_{-0.0008} ~c~\mathrm{s}^{-1}$ at a $90 \%$ confidence level. Additionally, the Bayes factor for the LSA case over the isolated case is $58/1$, indicating that the GW data strongly prefer the LSA model. We conclude that this is the first indication showing merging BBHs are located near a third compact object.

astro-ph.HE↗

arXiv2025

Collimated Hard X-Rays from Hybrid Laser and Plasma Wakefield Accelerators

We report a synergistic enhancement of betatron radiation based on the hybrid laser and plasma wakefield acceleration scheme. Quasi-phase-stable acceleration in an up-ramp plasma density first generates GeV-energy electron beams that act as a drive beam for PWFA, which then further accelerates the witness beam to GeV energies, enhancing both photon energy and flux. A full width at half maximum divergence $(6.1 \pm 1.9)\times(5.8\pm 1.6) $ mrad$^2$ of betatron radiation, a critical energy of $71 \pm 8$ keV, and an average flux of more than $10^{14}$ photons per steradian above 5 keV were all experimentally obtained thanks to this scheme, which was an order of magnitude higher than the previous reports. Quasi-three-dimensional particle-in-cell simulations were used to model the acceleration and radiation of the electrons in our experimental conditions, establishing a new paradigm for compact collimated hard X-ray sources.

physics.plasm-ph↗

arXiv2025

CogMath: Assessing LLMs' Authentic Mathematical Ability from a Human Cognitive Perspective

Although large language models (LLMs) show promise in solving complex mathematical tasks, existing evaluation paradigms rely solely on a coarse measure of overall answer accuracy, which are insufficient for assessing their authentic capabilities. In this paper, we propose \textbf{CogMath}, which comprehensively assesses LLMs' mathematical abilities through the lens of human cognition. Specifically, inspired by psychological theories, CogMath formalizes human reasoning process into 3 stages: \emph{problem comprehension}, \emph{problem solving}, and \emph{solution summarization}. Within these stages, we investigate perspectives such as numerical calculation, knowledge, and counterfactuals, and design a total of 9 fine-grained evaluation dimensions. In each dimension, we develop an ``\emph{Inquiry}-\emph{Judge}-\emph{Reference}'' multi-agent system to generate inquiries that assess LLMs' mastery from this dimension. An LLM is considered to truly master a problem only when excelling in all inquiries from the 9 dimensions. By applying CogMath on three benchmarks, we reveal that the mathematical capabilities of 7 mainstream LLMs are overestimated by 30\%-40\%. Moreover, we locate their strengths and weaknesses across specific stages/dimensions, offering in-depth insights to further enhance their reasoning abilities.

cs.AI↗

arXiv2025

Immunogenicity Prediction with Dual Attention Enables Vaccine Target Selection

Immunogenicity prediction is a central topic in reverse vaccinology for finding candidate vaccines that can trigger protective immune responses. Existing approaches typically rely on highly compressed features and simple model architectures, leading to limited prediction accuracy and poor generalizability. To address these challenges, we introduce VenusVaccine, a novel deep learning solution with a dual attention mechanism that integrates pre-trained latent vector representations of protein sequences and structures. We also compile the most comprehensive immunogenicity dataset to date, encompassing over 7000 antigen sequences, structures, and immunogenicity labels from bacteria, virus, and tumor. Extensive experiments demonstrate that VenusVaccine outperforms existing methods across a wide range of evaluation metrics. Furthermore, we establish a post-hoc validation protocol to assess the practical significance of deep learning models in tackling vaccine design challenges. Our work provides an effective tool for vaccine design and sets valuable benchmarks for future research. The implementation is at https://github.com/songleee/VenusVaccine.

cs.LG↗

arXiv2025

Audio Turing Test: Benchmarking the Human-likeness of Large Language Model-based Text-to-Speech Systems in Chinese

Recent advances in large language models (LLMs) have significantly improved text-to-speech (TTS) systems, enhancing control over speech style, naturalness, and emotional expression, which brings TTS Systems closer to human-level performance. Although the Mean Opinion Score (MOS) remains the standard for TTS System evaluation, it suffers from subjectivity, environmental inconsistencies, and limited interpretability. Existing evaluation datasets also lack a multi-dimensional design, often neglecting factors such as speaking styles, context diversity, and trap utterances, which is particularly evident in Chinese TTS evaluation. To address these challenges, we introduce the Audio Turing Test (ATT), a multi-dimensional Chinese corpus dataset ATT-Corpus paired with a simple, Turing-Test-inspired evaluation protocol. Instead of relying on complex MOS scales or direct model comparisons, ATT asks evaluators to judge whether a voice sounds human. This simplification reduces rating bias and improves evaluation robustness. To further support rapid model development, we also finetune Qwen2-Audio-Instruct with human judgment data as Auto-ATT for automatic evaluation. Experimental results show that ATT effectively differentiates models across specific capability dimensions using its multi-dimensional design. Auto-ATT also demonstrates strong alignment with human evaluations, confirming its value as a fast and reliable assessment tool. The white-box ATT-Corpus and Auto-ATT can be found in ATT Hugging Face Collection (https://huggingface.co/collections/meituan/audio-turing-test-682446320368164faeaf38a4).

cs.SD↗

arXiv2025

VenusMutHub: A systematic evaluation of protein mutation effect predictors on small-scale experimental data

In protein engineering, while computational models are increasingly used to predict mutation effects, their evaluations primarily rely on high-throughput deep mutational scanning (DMS) experiments that use surrogate readouts, which may not adequately capture the complex biochemical properties of interest. Many proteins and their functions cannot be assessed through high-throughput methods due to technical limitations or the nature of the desired properties, and this is particularly true for the real industrial application scenario. Therefore, the desired testing datasets, will be small-size (~10-100) experimental data for each protein, and involve as many proteins as possible and as many properties as possible, which is, however, lacking. Here, we present VenusMutHub, a comprehensive benchmark study using 905 small-scale experimental datasets curated from published literature and public databases, spanning 527 proteins across diverse functional properties including stability, activity, binding affinity, and selectivity. These datasets feature direct biochemical measurements rather than surrogate readouts, providing a more rigorous assessment of model performance in predicting mutations that affect specific molecular functions. We evaluate 23 computational models across various methodological paradigms, such as sequence-based, structure-informed and evolutionary approaches. This benchmark provides practical guidance for selecting appropriate prediction methods in protein engineering applications where accurate prediction of specific functional properties is crucial.

q-bio.QM↗

arXiv2025

Stable recovery guarantees for blind deconvolution under random mask assumption

This study addresses the blind deconvolution problem with modulated inputs, focusing on a measurement model where an unknown blurring kernel $\boldsymbol{h}$ is convolved with multiple random modulations $\{\boldsymbol{d}_l\}_{l=1}^{L}$(coded masks) of a signal $\boldsymbol{x}$, subject to $\ell_2$-bounded noise. We introduce a more generalized framework for coded masks, enhancing the versatility of our approach. Our work begins within a constrained least squares framework, where we establish a robust recovery bound for both $\boldsymbol{h}$ and $\boldsymbol{x}$, demonstrating its near-optimality up to a logarithmic factor. Additionally, we present a new recovery scheme that leverages sparsity constraints on $\boldsymbol{x}$. This approach significantly reduces the sampling complexity to the order of $L=O(\log n)$ when the non-zero elements of $\boldsymbol{x}$ are sufficiently separated. Furthermore, we demonstrate that incorporating sparsity constraints yields a refined error bound compared to the traditional constrained least squares model. The proposed method results in more robust and precise signal recovery, as evidenced by both theoretical analysis and numerical simulations. These findings contribute to advancing the field of blind deconvolution and offer potential improvements in various applications requiring signal reconstruction from modulated inputs.

cs.IT↗

arXiv2025

Native antisite defects in h-BN

Hexagonal boron nitride (hBN) is an excellent host for solid-state single phonon emitters. Experimental observed emission ranges from infrared to ultraviolet. The emission centers are generally attributed to either intrinsic or extrinsic point defects embedded into hBN. Nevertheless, the microscopic structure of most of these defect emitters is uncertain. Here, through density-functional theory calculations we studied the native antisite defects in hBN. We find that the neutral boron antisite might be a nonmagnetic single photon source with zero-phonon-line (ZPL) at 1.58 eV and such a lineshape that is often observed in experiments. Furthermore, the positively charged nitrogen antisite might be associated with a dim color center recently observed as a blue emitter with ZPL at 2.63 eV. These simple single substitution defects indicate the existence of out-of-plane phonon mode which significantly affects the optical properties. Our results could provide useful information for identification of quantum emitters in hBN.

cond-mat.mtrl-sci↗

arXiv2025

Stark Shift from Quantum Defects in Hexagonal Boron Nitride

Color centers in hexagonal boron nitride have emerged as promising candidates for quantum information applications, owing to their efficient and bright single photon emission. Despite the challenges in directly characterizing these emitters, the interaction between external fields and defects, such as the Stark shift, offers valuable insights into their local geometric configurations. In this study, we focus on clarifying the possible origin of the distinct Stark shift characteristics observed experimentally, particularly in the emission range around 2 eV. We find that the local symmetry of the defects plays a crucial role in determining the nature of the Stark shift, which can be either linear or quadratic. Additionally, the local dielectric environment significantly influences the Stark shift response. Our calculations not only enhance the understanding of the micro-structure of these hitherto unknown emitters but also pave the way for their more effective utilization as single-photon sources and qubits in quantum technologies.

cond-mat.mtrl-sci↗