arXiv subjects

Jie Li

Publications and source records attributed to Jie Li.

At least 235 records · Page 13Linked to original sources

arXiv2025

Distinct orbital contributions to electronic and magnetic structures in La$_{4}$Ni$_{3}$O$_{10}$

High-T$_c$ superconductivity has recently been discovered in Ruddlesden-Popper phase nickelates under pressure, where the low-energy electronic structure is dominated by Ni $d_{x^2 - y^2}$ and $d_{z^2}$ orbitals. However, the respective roles of these orbitals in superconductivity remain unclear. Here, by combining X-ray absorption, electron energy loss spectroscopy, and density functional theory calculations on La$_{4}$Ni$_{3}$O$_{10}$ single crystals, we identify ligand holes in the $p_{x,y}$ orbitals of planar oxygen and the $p_z$ orbitals of apical oxygen, which hybridize with the Ni $d_{x^2-y^2}$ and $d_{z^2}$ orbitals, respectively. These ligand holes enable orbital-selective O K-edge resonant inelastic X-ray scattering (RIXS) study, which reveals that $d_{x^2-y^2}$ states dominate the low-energy charge excitations and are more itinerant. We also observe a $\sim$0.1 eV bimagnon through RIXS and Raman spectroscopy, which leads to an interlayer superexchange interaction J$_z$ of $\sim$50 meV. Our results reveal distinct contributions of Ni $d_{x^2-y^2}$ and $d_{z^2}$ orbitals to the electronic and magnetic structure and provide direct experimental insights to understand the RP-phase nickelate superconductors.

cond-mat.supr-con↗

arXiv2025

FreezeVLA: Action-Freezing Attacks against Vision-Language-Action Models

Vision-Language-Action (VLA) models are driving rapid progress in robotics by enabling agents to interpret multimodal inputs and execute complex, long-horizon tasks. However, their safety and robustness against adversarial attacks remain largely underexplored. In this work, we identify and formalize a critical adversarial vulnerability in which adversarial images can "freeze" VLA models and cause them to ignore subsequent instructions. This threat effectively disconnects the robot's digital mind from its physical actions, potentially inducing inaction during critical interventions. To systematically study this vulnerability, we propose FreezeVLA, a novel attack framework that generates and evaluates action-freezing attacks via min-max bi-level optimization. Experiments on three state-of-the-art VLA models and four robotic benchmarks show that FreezeVLA attains an average attack success rate of 76.2%, significantly outperforming existing methods. Moreover, adversarial images generated by FreezeVLA exhibit strong transferability, with a single image reliably inducing paralysis across diverse language prompts. Our findings expose a critical safety risk in VLA models and highlight the urgent need for robust defense mechanisms.

cs.CV↗

arXiv2025

WenetSpeech-Chuan: A Large-Scale Sichuanese Corpus with Rich Annotation for Dialectal Speech Processing

The scarcity of large-scale, open-source data for dialects severely hinders progress in speech technology, a challenge particularly acute for the widely spoken Sichuanese dialects of Chinese. To address this critical gap, we introduce WenetSpeech-Chuan, a 10,000-hour, richly annotated corpus constructed using our novel Chuan-Pipeline, a complete data processing framework for dialectal speech. To facilitate rigorous evaluation and demonstrate the corpus's effectiveness, we also release high-quality ASR and TTS benchmarks, WenetSpeech-Chuan-Eval, with manually verified transcriptions. Experiments show that models trained on WenetSpeech-Chuan achieve state-of-the-art performance among open-source systems and demonstrate results comparable to commercial services. As the largest open-source corpus for Sichuanese dialects, WenetSpeech-Chuan not only lowers the barrier to research in dialectal speech processing but also plays a crucial role in promoting AI equity and mitigating bias in speech technologies. The corpus, benchmarks, models, and receipts are publicly available on our project page.

cs.CL↗

arXiv2025

Qianfan-VL: Domain-Enhanced Universal Vision-Language Models

We present Qianfan-VL, a series of multimodal large language models ranging from 3B to 70B parameters, achieving state-of-the-art performance through innovative domain enhancement techniques. Our approach employs multi-stage progressive training and high-precision data synthesis pipelines, which prove to be critical technologies for enhancing domain-specific capabilities while maintaining strong general performance. Qianfan-VL achieves comparable results to leading open-source models on general benchmarks, with state-of-the-art performance on benchmarks such as CCBench, SEEDBench IMG, ScienceQA, and MMStar. The domain enhancement strategy delivers significant advantages in OCR and document understanding, validated on both public benchmarks (OCRBench 873, DocVQA 94.75%) and in-house evaluations. Notably, Qianfan-VL-8B and 70B variants incorporate long chain-of-thought capabilities, demonstrating superior performance on mathematical reasoning (MathVista 78.6%) and logical inference tasks. All models are trained entirely on Baidu's Kunlun P800 chips, validating the capability of large-scale AI infrastructure to train SOTA-level multimodal models with over 90% scaling efficiency on 5000 chips for a single task. This work establishes an effective methodology for developing domain-enhanced multimodal models suitable for diverse enterprise deployment scenarios.

cs.CV↗

arXiv2025

A zero-dead-time strontium lattice clock with a stability at $10^{-19}$ level

Optical atomic clocks play a crucial role in fundamental physics, relativistic geodesy, and the future redefinition of the SI second. Standard operation relies on cyclic interrogation sequences, which alternate between atomic interrogation and dead time used for state preparation and readout. This approach introduces the Dick effect, where laser frequency noise aliases onto the atomic transition frequency. Although reducing laser noise improves clock stability, the Dick effect remains a key limitation. In this work, we demonstrate a zero-dead-time optical clock based on two interleaved ensembles of cold $^{87}\text{Sr}$ atoms. Our system significantly suppresses this noise and achieves a fractional frequency instability at the $10^{-19}$ level between 10,000 and 20,000 seconds over repeated measurements, with a best value of $2.9 \times 10^{-19}$ at $τ= 20,000$ seconds. The estimated long-term stability based on the combined data of these measurements reaches $2.5 \times 10^{-19}$ at one day. These results represent a more than ninefold improvement over a conventional single-ensemble clock, highlighting its potential for next-generation timekeeping applications.

physics.atom-ph↗

arXiv2025

Improved systematic evaluation of a strontium optical clock with uncertainty below $1\times 10^{-18}$

We report a systematic uncertainty of $9.2\times 10^{-19}$ for the USTC Sr1 optical lattice clock, achieving accuracy at the level required for the roadmap of the redefinition of the SI second. A finite-element model with {\it in situ}-validated, spatially-resolved chamber emissivity reduced blackbody radiation shift uncertainty to $6.3\times 10^{-19}$. Concurrently, an externally mounted lattice cavity combined with a larger beam waist suppressed density shifts. Enhanced lattice depth modulation consolidated lattice light shift uncertainty to $6.3\times 10^{-19}$ by enabling simultaneous determination of key polarizabilities and magic wavelength. Magnetic shifts were resolved below $10^{-18}$ via precision characterization of the second-order Zeeman coefficient. Supported by a crystalline-coated ultra-low-expansion cavity-stabilized laser and refined temperature control suppressing BBR fluctuations, the clock also achieves a frequency stability better than $1\times10^{-18}$ at 30,000-s averaging time. These developments collectively establish a new benchmark in USTC Sr1 clock performance and pave the way for high-accuracy applications in metrology and fundamental physics.

physics.atom-ph↗

arXiv2025

BAPFL: Exploring Backdoor Attacks Against Prototype-based Federated Learning

Prototype-based federated learning (PFL) has emerged as a promising paradigm to address data heterogeneity problems in federated learning, as it leverages mean feature vectors as prototypes to enhance model generalization. However, its robustness against backdoor attacks remains largely unexplored. In this paper, we identify that PFL is inherently resistant to existing backdoor attacks due to its unique prototype learning mechanism and local data heterogeneity. To further explore the security of PFL, we propose BAPFL, the first backdoor attack method specifically designed for PFL frameworks. BAPFL integrates a prototype poisoning strategy with a trigger optimization mechanism. The prototype poisoning strategy manipulates the trajectories of global prototypes to mislead the prototype training of benign clients, pushing their local prototypes of clean samples away from the prototypes of trigger-embedded samples. Meanwhile, the trigger optimization mechanism learns a unique and stealthy trigger for each potential target label, and guides the prototypes of trigger-embedded samples to align closely with the global prototype of the target label. Experimental results across multiple datasets and PFL variants demonstrate that BAPFL achieves a 35\%-75\% improvement in attack success rate compared to traditional backdoor attacks, while preserving main task accuracy. These results highlight the effectiveness, stealthiness, and adaptability of BAPFL in PFL.

cs.LG↗

arXiv2025

WenetSpeech-Yue: A Large-scale Cantonese Speech Corpus with Multi-dimensional Annotation

The development of speech understanding and generation has been significantly accelerated by the availability of large-scale, high-quality speech datasets. Among these, ASR and TTS are regarded as the most established and fundamental tasks. However, for Cantonese (Yue Chinese), spoken by approximately 84.9 million native speakers worldwide, limited annotated resources have hindered progress and resulted in suboptimal ASR and TTS performance. To address this challenge, we propose WenetSpeech-Pipe, an integrated pipeline for building large-scale speech corpus with multi-dimensional annotation tailored for speech understanding and generation. It comprises six modules: Audio Collection, Speaker Attributes Annotation, Speech Quality Annotation, Automatic Speech Recognition, Text Postprocessing and Recognizer Output Voting, enabling rich and high-quality annotations. Based on this pipeline, we release WenetSpeech-Yue, the first large-scale Cantonese speech corpus with multi-dimensional annotation for ASR and TTS, covering 21,800 hours across 10 domains with annotations including ASR transcription, text confidence, speaker identity, age, gender, speech quality scores, among other annotations. We also release WSYue-eval, a comprehensive Cantonese benchmark with two components: WSYue-ASR-eval, a manually annotated set for evaluating ASR on short and long utterances, code-switching, and diverse acoustic conditions, and WSYue-TTS-eval, with base and coverage subsets for standard and generalization testing. Experimental results show that models trained on WenetSpeech-Yue achieve competitive results against state-of-the-art (SOTA) Cantonese ASR and TTS systems, including commercial and LLM-based models, highlighting the value of our dataset and pipeline.

cs.SD↗

arXiv2025

Multilingual Speech Recognition Using Discrete Tokens with a Two-step Training Strategy

Pre-trained models, especially self-supervised learning (SSL) models, have demonstrated impressive results in automatic speech recognition (ASR) task. While most applications of SSL models focus on leveraging continuous representations as features for training downstream tasks, the utilization of discrete units has gained increasing attention in recent years owing to its lower storage requirements and broader range of applications. In multilingual ASR tasks, representations at different layers of the model contribute differently to various languages, complicating the unification of discrete unit modeling. In this paper, we propose a two-stage training strategy to improve the discrete token performance of pre-trained models and narrow the gap with continuous representation performance. We validate our method on the XLS-R model following the settings of Interspeech2024 Speech Processing Using Discrete Speech Unit Challenge. Our method demonstrates a significant improvement on the ML-SUPERB dataset, achieving a 44% relative reduction on CER for the XLS-R model. This surpasses the previous baseline set by the WavLM model, which achieves a 26% relative reduction on CER. Furthermore, our method achieves the first place among all the single-system results on the leaderboard.

eess.AS↗

arXiv2025

DynaMind: Reconstructing Dynamic Visual Scenes from EEG by Aligning Temporal Dynamics and Multimodal Semantics to Guided Diffusion

Reconstruction dynamic visual scenes from electroencephalography (EEG) signals remains a primary challenge in brain decoding, limited by the low spatial resolution of EEG, a temporal mismatch between neural recordings and video dynamics, and the insufficient use of semantic information within brain activity. Therefore, existing methods often inadequately resolve both the dynamic coherence and the complex semantic context of the perceived visual stimuli. To overcome these limitations, we introduce DynaMind, a novel framework that reconstructs video by jointly modeling neural dynamics and semantic features via three core modules: a Regional-aware Semantic Mapper (RSM), a Temporal-aware Dynamic Aligner (TDA), and a Dual-Guidance Video Reconstructor (DGVR). The RSM first utilizes a regional-aware encoder to extract multimodal semantic features from EEG signals across distinct brain regions, aggregating them into a unified diffusion prior. In the mean time, the TDA generates a dynamic latent sequence, or blueprint, to enforce temporal consistency between the feature representations and the original neural recordings. Together, guided by the semantic diffusion prior, the DGVR translates the temporal-aware blueprint into a high-fidelity video reconstruction. On the SEED-DV dataset, DynaMind sets a new state-of-the-art (SOTA), boosting reconstructed video accuracies (video- and frame-based) by 12.5 and 10.3 percentage points, respectively. It also achieves a leap in pixel-level quality, showing exceptional visual fidelity and temporal coherence with a 9.4% SSIM improvement and a 19.7% FVMD reduction. This marks a critical advancement, bridging the gap between neural dynamics and high-fidelity visual semantics.

cs.CV↗

arXiv2025

Quantile Function-Based Models for Neuroimaging Classification Using Wasserstein Regression

We propose a novel quantile function-based approach for neuroimaging classification using Wasserstein-Fréchet regression, specifically applied to the detection of mild traumatic brain injury (mTBI) based on the MEG and MRI data. Conventional neuroimaging classification methods for mTBI detection typically extract summary statistics from brain signals across the different epochs, which may result in the loss of important distributional information, such as variance, skewness, kurtosis, etc. Our approach treats complete probability density functions of epoch space results as functional response variables within a Wasserstein-Fréchet regression framework, thereby preserving the full distributional characteristics of epoch results from $L_{1}$ minimum norm solutions. The global Wasserstein-Fréchet regression model incorporating covariates (age and gender) allows us to directly compare the distributional patterns between healthy control subjects and mTBI patients. The classification procedure computes Wasserstein distances between estimated quantile functions from control and patient groups, respectively. These distances are then used as the basis for diagnostic decisions. This framework offers a statistically principled approach to improving diagnostic accuracy in mTBI detection. In practical applications, the test accuracy on unseen data from Innovision IP's dataset achieves up to 98\%.

stat.AP↗

arXiv2025

LinguaSafe: A Comprehensive Multilingual Safety Benchmark for Large Language Models

The widespread adoption and increasing prominence of large language models (LLMs) in global technologies necessitate a rigorous focus on ensuring their safety across a diverse range of linguistic and cultural contexts. The lack of a comprehensive evaluation and diverse data in existing multilingual safety evaluations for LLMs limits their effectiveness, hindering the development of robust multilingual safety alignment. To address this critical gap, we introduce LinguaSafe, a comprehensive multilingual safety benchmark crafted with meticulous attention to linguistic authenticity. The LinguaSafe dataset comprises 45k entries in 12 languages, ranging from Hungarian to Malay. Curated using a combination of translated, transcreated, and natively-sourced data, our dataset addresses the critical need for multilingual safety evaluations of LLMs, filling the void in the safety evaluation of LLMs across diverse under-represented languages from Hungarian to Malay. LinguaSafe presents a multidimensional and fine-grained evaluation framework, with direct and indirect safety assessments, including further evaluations for oversensitivity. The results of safety and helpfulness evaluations vary significantly across different domains and different languages, even in languages with similar resource levels. Our benchmark provides a comprehensive suite of metrics for in-depth safety evaluation, underscoring the critical importance of thoroughly assessing multilingual safety in LLMs to achieve more balanced safety alignment. Our dataset and code are released to the public to facilitate further research in the field of multilingual LLM safety.

cs.CL↗

arXiv2025

Compact Attention: Exploiting Structured Spatio-Temporal Sparsity for Fast Video Generation

The computational demands of self-attention mechanisms pose a critical challenge for transformer-based video generation, particularly in synthesizing ultra-long sequences. Current approaches, such as factorized attention and fixed sparse patterns, fail to fully exploit the inherent spatio-temporal redundancies in video data. Through systematic analysis of video diffusion transformers (DiT), we uncover a key insight: Attention matrices exhibit structured, yet heterogeneous sparsity patterns, where specialized heads dynamically attend to distinct spatiotemporal regions (e.g., local pattern, cross-shaped pattern, or global pattern). Existing sparse attention methods either impose rigid constraints or introduce significant overhead, limiting their effectiveness. To address this, we propose Compact Attention, a hardware-aware acceleration framework featuring three innovations: 1) Adaptive tiling strategies that approximate diverse spatial interaction patterns via dynamic tile grouping, 2) Temporally varying windows that adjust sparsity levels based on frame proximity, and 3) An automated configuration search algorithm that optimizes sparse patterns while preserving critical attention pathways. Our method achieves 1.6~2.5x acceleration in attention computation on single-GPU setups while maintaining comparable visual quality with full-attention baselines. This work provides a principled approach to unlocking efficient long-form video generation through structured sparsity exploitation. Project Page: https://yo-ava.github.io/Compact-Attention.github.io/

cs.CV↗

arXiv2025

Magnon cat states in a cavity-magnon-qubit system via two-magnon driving and dissipation

We propose an efficient method for dissipative generation of magnonic cat states in a cavity-magnon-qubit hybrid system by exploiting a two-magnon driving and dissipation mechanism. When both the magnon and qubit are driven, a coherent nonlinear two-magnon interaction is induced, wherein the qubit and the magnon mode exchange energy through magnon pairs. The dissipation of the qubit is exploited to steer the magnon mode into a quantum superposition of distinct coherent states, where the magnon mode evolves into either an even or odd cat state, depending on the parity of the magnon initial state. For the case where the magnon initial state is a superposition state, e.g., of $\ket{0}$ and $\ket{1}$, the magnon mode can evolve into a weighted mixture of the even and odd cat states. We also find that magnon squeezed states may emerge during the short-time evolution, showcasing the capability of our mechanism in preparing diverse magnon non-classical states. Magnonic cat and squeezed states are macroscopic quantum states and find applications in macroscopic quantum studies and quantum sensing, e.g., in the dark matter search using ferromagnetic axion haloscopes.

quant-ph↗

arXiv2025

Argus Inspection: Do Multimodal Large Language Models Possess the Eye of Panoptes?

As Multimodal Large Language Models (MLLMs) continue to evolve, their cognitive and reasoning capabilities have seen remarkable progress. However, challenges in visual fine-grained perception and commonsense causal inference persist. This paper introduces Argus Inspection, a multimodal benchmark with two levels of difficulty, emphasizing detailed visual recognition while incorporating real-world commonsense understanding to evaluate causal reasoning abilities. Expanding on it, we present the Eye of Panoptes framework, which integrates a binary parametric Sigmoid metric with an indicator function, enabling a more holistic evaluation of MLLMs' responses in opinion-based reasoning tasks. Experiments conducted on 26 mainstream MLLMs reveal that the highest performance in visual fine-grained reasoning reaches only 0.46, highlighting considerable potential for enhancement. Our research offers valuable perspectives for the continued refinement of MLLMs.

cs.CV↗

arXiv2025

MemoryKT: An Integrative Memory-and-Forgetting Method for Knowledge Tracing

Knowledge Tracing (KT) is committed to capturing students' knowledge mastery from their historical interactions. Simulating students' memory states is a promising approach to enhance both the performance and interpretability of knowledge tracing models. Memory consists of three fundamental processes: encoding, storage, and retrieval. Although forgetting primarily manifests during the storage stage, most existing studies rely on a single, undifferentiated forgetting mechanism, overlooking other memory processes as well as personalized forgetting patterns. To address this, this paper proposes memoryKT, a knowledge tracing model based on a novel temporal variational autoencoder. The model simulates memory dynamics through a three-stage process: (i) Learning the distribution of students' knowledge memory features, (ii) Reconstructing their exercise feedback, while (iii) Embedding a personalized forgetting module within the temporal workflow to dynamically modulate memory storage strength. This jointly models the complete encoding-storage-retrieval cycle, significantly enhancing the model's perception capability for individual differences. Extensive experiments on four public datasets demonstrate that our proposed approach significantly outperforms state-of-the-art baselines.

cs.LG↗

arXiv2025

SafeWork-R1: Coevolving Safety and Intelligence under the AI-45$^{\circ}$ Law

We introduce SafeWork-R1, a cutting-edge multimodal reasoning model that demonstrates the coevolution of capabilities and safety. It is developed by our proposed SafeLadder framework, which incorporates large-scale, progressive, safety-oriented reinforcement learning post-training, supported by a suite of multi-principled verifiers. Unlike previous alignment methods such as RLHF that simply learn human preferences, SafeLadder enables SafeWork-R1 to develop intrinsic safety reasoning and self-reflection abilities, giving rise to safety `aha' moments. Notably, SafeWork-R1 achieves an average improvement of $46.54\%$ over its base model Qwen2.5-VL-72B on safety-related benchmarks without compromising general capabilities, and delivers state-of-the-art safety performance compared to leading proprietary models such as GPT-4.1 and Claude Opus 4. To further bolster its reliability, we implement two distinct inference-time intervention methods and a deliberative search mechanism, enforcing step-level verification. Finally, we further develop SafeWork-R1-InternVL3-78B, SafeWork-R1-DeepSeek-70B, and SafeWork-R1-Qwen2.5VL-7B. All resulting models demonstrate that safety and capability can co-evolve synergistically, highlighting the generalizability of our framework in building robust, reliable, and trustworthy general-purpose AI.

cs.AI↗

arXiv2025

Identifying Radio Active Galactic Nuclei with Machine Learning and Large-Area Surveys

Context. Active galactic nuclei (AGNs) and star forming galaxies (SFGs) are the primary sources of extragalactic radio sky. But it is difficult to distinguish the radio emission produced by AGNs from that by SFGs, especially when the radio sources are faint. Best et al. (2023) classified the radio sources in LoTSS Deep Fields DR1 through multiwavelength SED fitting. With the classification results of them, we perform a supervised machine learning to distinguish radio AGNs and radio SFGs. Aims. We aim to provide a supervised classifier to identify radio AGNs, which can get both high purity and completeness simultaneously, and can easily be applied to datasets of large-area surveys. Methods. The classifications of Best et al. (2023) are used as the true labels for supervised machine learning. With the cross-matched sample of LoTSS Deep Fields DR1, AllWISE and Gaia DR3, the features of optical and mid-infrared magnitude and colors, are applied to train the classifier. The performance of the classifier is evaluated mainly by the precission, recall and F1 score of both AGNs and non-AGNs. Results. By comparing the performance of six learning algorithms, CatBoost is chosen to construct the best classifier. The best classifier get precision = 0.974, recall = 0.865 and F1 = 0.916 for AGNs, precision = 0.936, recall = 0.988 and F1 = 0.961 for non-AGNs. After applying our classifier to the cross-matched sample of LoTSS DR2, AllWISE and Gaia DR3, we obtain a sample of 49716 AGNs and 102261 non-AGNs. The reliability of these classification results is confirmed by comparing with the spectroscopic classification of SDSS. The precission and recall of AGN sample can be as high as 94.2% and 92.3%, respectively. We also train a model to identify radio excess sources. The F1 scores are 0.610 and 0.965 for sources with and without radio excess, respectively.

astro-ph.GA↗