arXiv subjects

Jing Li

Publications and source records attributed to Jing Li.

At least 253 records · Page 14Linked to original sources

arXiv2025

Hunyuan3D Studio: End-to-End AI Pipeline for Game-Ready 3D Asset Generation

The creation of high-quality 3D assets, a cornerstone of modern game development, has long been characterized by labor-intensive and specialized workflows. This paper presents Hunyuan3D Studio, an end-to-end AI-powered content creation platform designed to revolutionize the game production pipeline by automating and streamlining the generation of game-ready 3D assets. At its core, Hunyuan3D Studio integrates a suite of advanced neural modules (such as Part-level 3D Generation, Polygon Generation, Semantic UV, etc.) into a cohesive and user-friendly system. This unified framework allows for the rapid transformation of a single concept image or textual description into a fully-realized, production-quality 3D model complete with optimized geometry and high-fidelity PBR textures. We demonstrate that assets generated by Hunyuan3D Studio are not only visually compelling but also adhere to the stringent technical requirements of contemporary game engines, significantly reducing iteration time and lowering the barrier to entry for 3D content creation. By providing a seamless bridge from creative intent to technical asset, Hunyuan3D Studio represents a significant leap forward for AI-assisted workflows in game development and interactive media.

cs.CV↗

arXiv2025

A Chinese Continuous Sign Language Dataset Based on Complex Environments

The current bottleneck in continuous sign language recognition (CSLR) research lies in the fact that most publicly available datasets are limited to laboratory environments or television program recordings, resulting in a single background environment with uniform lighting, which significantly deviates from the diversity and complexity found in real-life scenarios. To address this challenge, we have constructed a new, large-scale dataset for Chinese continuous sign language (CSL) based on complex environments, termed the complex environment - chinese sign language dataset (CE-CSL). This dataset encompasses 5,988 continuous CSL video clips collected from daily life scenes, featuring more than 70 different complex backgrounds to ensure representativeness and generalization capability. To tackle the impact of complex backgrounds on CSLR performance, we propose a time-frequency network (TFNet) model for continuous sign language recognition. This model extracts frame-level features and then utilizes both temporal and spectral information to separately derive sequence features before fusion, aiming to achieve efficient and accurate CSLR. Experimental results demonstrate that our approach achieves significant performance improvements on the CE-CSL, validating its effectiveness under complex background conditions. Additionally, our proposed method has also yielded highly competitive results when applied to three publicly available CSL datasets.

cs.CV↗

arXiv2025

To See a World in a Spark of Neuron: Disentangling Multi-task Interference for Training-free Model Merging

Fine-tuning pre-trained models on targeted datasets enhances task-specific performance but often comes at the expense of generalization. Model merging techniques, which integrate multiple fine-tuned models into a single multi-task model through task arithmetic, offer a promising solution. However, task interference remains a fundamental challenge, leading to performance degradation and suboptimal merged models. Existing approaches largely overlooked the fundamental roles of neurons, their connectivity, and activation, resulting in a merging process and a merged model that does not consider how neurons relay and process information. In this work, we present the first study that relies on neuronal mechanisms for model merging. Specifically, we decomposed task-specific representations into two complementary neuronal subspaces that regulate input sensitivity and task adaptability. Leveraging this decomposition, we introduced NeuroMerging, a novel merging framework developed to mitigate task interference within neuronal subspaces, enabling training-free model fusion across diverse tasks. Through extensive experiments, we demonstrated that NeuroMerging achieved superior performance compared to existing methods on multi-task benchmarks across both natural language and vision domains. Our findings highlighted the importance of aligning neuronal mechanisms in model merging, offering new insights into mitigating task interference and improving knowledge fusion. Our project is available at https://ZzzitaoFang.github.io/projects/NeuroMerging/.

cs.LG↗

arXiv2025

Towards Scalable and Structured Spatiotemporal Forecasting

In this paper, we propose a novel Spatial Balance Attention block for spatiotemporal forecasting. To strike a balance between obeying spatial proximity and capturing global correlation, we partition the spatial graph into a set of subgraphs and instantiate Intra-subgraph Attention to learn local spatial correlation within each subgraph; to capture the global spatial correlation, we further aggregate the nodes to produce subgraph representations and achieve message passing among the subgraphs via Inter-subgraph Attention. Building on the proposed Spatial Balance Attention block, we develop a multiscale spatiotemporal forecasting model by progressively increasing the subgraph scales. The resulting model is both scalable and able to produce structured spatial correlation, and meanwhile, it is easy to implement. We evaluate its efficacy and efficiency against the existing models on real-world spatiotemporal datasets from medium to large sizes. The experimental results show that it can achieve performance improvements up to 7.7% over the baseline methods at low running costs.

cs.LG↗

arXiv2025

Unveiling Galactic substructures with M Giant stars: A kinematic and chemical study based on LAMOST DR9, Gaia DR3 and APOGEE DR17

Based on the updated M giant star catalog selected from LAMOST DR9, we iden-tify substructures within the integrals-of-motion space through Friends-of-Friends cluster-ing algorithm. We obtain members belonging to several known substructures: the Sagittarius stream, Galactic Anticenter Substructure (GASS), Gaia-Enceladus-Sausage (GES), Splash, and the high-α disk. Furthermore, we also identify two groups which cannot be clearly asso-ciated with previously known substructures. Our findings confirm the existence of metal-rich constituents within the GES, representing newly formed stars that originated from the metal-enriched gas delivered during the GES merger event and subsequently evolved. Additionally, this study further expands the sample of GASS, high-α disk, and Splash stars. Analysis of these metal-rich M giant stars as members of the GES, Splash, and high-α disk compo-nents supports an evolution scenario for the early Milky Way, as proposed by previous stud-ies. In this scenario, stars initially formed in a high-α primordial disk were dynamically heated by the massive accretion event (GES). This process redistributed stellar orbits, creat-ing the Splash population, while the undisturbed portion of the primordial disk persisted as the present-day high-α disk component.

astro-ph.GA↗

arXiv2025

STARE at the Structure: Steering ICL Exemplar Selection with Structural Alignment

In-Context Learning (ICL) has become a powerful paradigm that enables LLMs to perform a wide range of tasks without task-specific fine-tuning. However, the effectiveness of ICL heavily depends on the quality of exemplar selection. In particular, for structured prediction tasks such as semantic parsing, existing ICL selection strategies often overlook structural alignment, leading to suboptimal performance and poor generalization. To address this issue, we propose a novel two-stage exemplar selection strategy that achieves a strong balance between efficiency, generalizability, and performance. First, we fine-tune a BERT-based retriever using structure-aware supervision, guiding it to select exemplars that are both semantically relevant and structurally aligned. Then, we enhance the retriever with a plug-in module, which amplifies syntactically meaningful information in the hidden representations. This plug-in is model-agnostic, requires minimal overhead, and can be seamlessly integrated into existing pipelines. Experiments on four benchmarks spanning three semantic parsing tasks demonstrate that our method consistently outperforms existing baselines with multiple recent LLMs as inference-time models.

cs.CL↗

arXiv2025

$Λ$-Type Giant Atom Mediated Controllable Single-Photon Transport in a One-Dimensional Chiral Waveguide

We investigate the single-photon scattering spectrum of a driven $Λ$-type giant atom system chirally coupled to a one-dimensional (1D) waveguide. By employing a real-space scattering approach, we obtain analytical solutions for the scattering amplitudes that remain valid in both Markovian and non-Markovian regimes. We observe that an external driving field induces a splitting of the transmission spectrum's dip into double dips, with the distance between the two dips increasing as the strength of the driving field increases. The chiral nature of the coupling allows for controlled switching between complete transmission and perfect reflection of incident photons. In the Markovian limit, we predict robust perfect transmission at specific phase values, independent of the driving field parameters.Moreover, in the non-Markovian regime, as the size of the giant atom increases, the oscillatory behavior of the scattering spectrum becomes more pronounced. Adjusting the giant atom size enables control over the number of decoupling points as well as the number of complete reflection points.

quant-ph↗

arXiv2025

Visual-RAG: Benchmarking Text-to-Image Retrieval Augmented Generation for Visual Knowledge Intensive Queries

Retrieval-augmented generation (RAG) is a paradigm that augments large language models (LLMs) with external knowledge to tackle knowledge-intensive question answering. While several benchmarks evaluate Multimodal LLMs (MLLMs) under Multimodal RAG settings, they predominantly retrieve from textual corpora and do not explicitly assess how models exploit visual evidence during generation. Consequently, there still lacks benchmark that isolates and measures the contribution of retrieved images in RAG. We introduce Visual-RAG, a question-answering benchmark that targets visually grounded, knowledge-intensive questions. Unlike prior work, Visual-RAG requires text-to-image retrieval and the integration of retrieved clue images to extract visual evidence for answer generation. With Visual-RAG, we evaluate 5 open-source and 3 proprietary MLLMs, showcasing that images provide strong evidence in augmented generation. However, even state-of-the-art models struggle to efficiently extract and utilize visual knowledge. Our results highlight the need for improved visual retrieval, grounding, and attribution in multimodal RAG systems.

cs.CL↗

arXiv2025

Mining the Social Fabric: Unveiling Communities for Fake News Detection in Short Videos

Short video platforms have become a major medium for information sharing, but their rapid content generation and algorithmic amplification also enable the widespread dissemination of fake news. Detecting misinformation in short videos is challenging due to their multi-modal nature and the limited context of individual videos. While recent methods focus on analyzing content signals-visual, textual, and audio-they often overlook implicit relationships among videos, uploaders, and events. To address this gap, we propose DugFND (Dual-community graph for fake news detection), a novel method that enhances existing video classifiers by modeling two key community patterns: (1) uploader communities, where uploaders with shared interests or similar content creation patterns group together, and (2) event-driven communities, where videos related to the same or semantically similar public events form localized clusters. We construct a heterogeneous graph connecting uploader, video, and event nodes, and design a time-aware heterogeneous graph attention network to enable effective message passing. A reconstruction-based pretraining phase further improves node representation learning. DugFND can be applied to any pre-trained classifier. Experiments on public datasets show that our method achieves significant performance gains, demonstrating the value of dual-community modeling for fake news detection in short videos.

cs.MM↗

arXiv2025

PRELUDE: A Benchmark Designed to Require Global Comprehension and Reasoning over Long Contexts

We introduce PRELUDE, a benchmark for evaluating long-context understanding through the task of determining whether a character's prequel story is consistent with the canonical narrative of the original book. Our task poses a stronger demand for global comprehension and deep reasoning than existing benchmarks -- as the prequels are not part of the original story, assessing their plausibility typically requires searching and integrating information that is only indirectly related. Empirically, 88% of instances require evidence from multiple parts of the narrative. Experimental results highlight the challenge of our task: in-context learning, RAG and in-domain training with state-of-the-art LLMs, and commercial DeepResearch services, lag behind humans by >15%. A further human study reveals that models often produce correct answers with flawed reasoning, leading to an over 30% gap in reasoning accuracy compared to humans. These findings underscore the substantial room for improvement in long-context understanding and reasoning.

cs.CL↗

arXiv2025

Current-induced spin and orbital polarization in the ferroelectric Rashba semiconductor GeTe

The Edelstein effect is a promising mechanism for generating spin and orbital polarization from charge currents in systems without inversion symmetry. In ferroelectric materials, such as Germanium Telluride (GeTe), the combination of bulk Rashba splitting and voltage-controlled ferroelectric polarization provides a pathway for electrical control of the sign of the charge-spin conversion. In this work, we investigate current-induced spin and orbital magnetization in bulk GeTe using Wannier-based tight-binding models derived from \textit{ab initio} calculations and semiclassical Boltzmann theory. Employing the modern theory of orbital magnetization, we demonstrate that the orbital Edelstein effect entirely dominates its spin counterpart. This difference is visualized through the spin and orbital textures at the Fermi surfaces, where the orbital moment surpasses the spin moment by one order of magnitude. Moreover, the orbital Edelstein effect remains largely unaffected in the absence of spin-orbit coupling, highlighting its distinct physical origin compared to the spin Edelstein effect.

cond-mat.mes-hall↗

arXiv2025

GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models

We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that supports both thinking and direct response modes. Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding (ARC) tasks, scoring 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified. With much fewer parameters than several competitors, GLM-4.5 ranks 3rd overall among all evaluated models and 2nd on agentic benchmarks. We release both GLM-4.5 (355B parameters) and a compact version, GLM-4.5-Air (106B parameters), to advance research in reasoning and agentic AI systems. Code, models, and more information are available at https://github.com/zai-org/GLM-4.5.

cs.CL↗

arXiv2025

Tight-binding photonics

Photonics, dealing with the generation, manipulation, and detection of photons in various systems, lays the foundation of many advanced technologies. A key task of photonics is to know how photons propagate in complex media such as periodic and aperiodic photonic crystals. The conventional wisdom is to numerically solve the Maxwell equations either by dedicated numerical techniques or brute-force finite-element calculations. Recently, the strict analogy between photonic crystals and theoretical tight-binding models provides an unprecedentedly convenient wayof understanding the spectra and wavefunctions of photonic systems by mapping the complicated differential equationsinto matrixed Hamiltonians that can be easily solved through the band theory and exact diagonalization. in this paper, we present a timely review of tight-binding-like photonics in various platforms, covering fundamental theories, experimental realizations, unique physical efiects, and their potential applications. We also provide a brief outlook on the future trends of this active area. Our review offers an in-depth and comprehensive picture on this rapidly developing field and may shed light on the future design on advanced tight-binding-like photonic devices.

physics.optics↗

arXiv2025

Quantum Knowledge Distillation for Large Language Models

As foundational tools in natural language processing, Large Language Models (LLMs) have immense parameter scales, which makes deployment and inference increasingly prohibitive, especially in resource-constrained devices. Therefore, knowledge distillation for LLMs, i.e., compressing the LLM to a smaller model, is meaningful. With strong parameter representation capacity, quantum computing is regarded as a promising solution. Here, we propose a Quantum knowledge Distillation model for LLMs (QD-LLM) that leverages variational quantum circuits to learn from LLMs. In classical simulation, QD-LLM outperforms several mainstream distillation methods on multiple text classification tasks in terms of both accuracy and efficiency using only 11 qubits. The results reveal an interesting phenomenon that the simulation of quantum student models may be regarded as a new class of quantum-inspired classical algorithms. Remarkably, we deploy the obtained circuits on the Baihua superconducting quantum processor via the Quafu platform to assess practical feasibility. The model maintains stable inference performance despite hardware constraints such as decoherence and finite sampling. In summary, QD-LLM marks a foundational step in connecting quantum computing with LLMs, demonstrating the feasibility of quantum-native approaches that aim to compress and deploy models of increasingly larger scales. The code of this article has been open-sourced at https://github.com/Lilingxiao-bupt/QD-LLM.

quant-ph↗

arXiv2025

Quantum Criticality by Interaction Frustration in a Square-Planar Lattice

We report experimental and theoretical investigations on ThCr$_2$Ge$_2$C, a metallic compound in which Cr$_2$C planes form a square-planar lattice. Neutron powder diffraction, magnetization, and specific heat measurements reveal no evidence of long-range magnetic order or short-range spin freezing down to 70~mK. Quantum critical behavior was indicated through logarithmic divergences in both the magnetic susceptibility and the specific heat divided by temperature. Resistivity measurements exhibit non-Fermi-liquid behavior, with a Fermi liquid recovered under magnetic fields or high pressures. First-principles calculations identify competing nearest-neighbor ($J_1$) and next-nearest-neighbor ($J_2$) exchange interactions, with $J_2/J_1 \sim -0.5$, pointing to strong magnetic frustration. The interaction frustration is reduced, and magnetically ordered phases are stabilized upon the application of negative or positive pressures. This work offers a rare example of zero-field, ambient pressure quantum criticality mainly driven by interaction frustration in a square lattice.

cond-mat.str-el↗

arXiv2025

System Report for CCL25-Eval Task 10: SRAG-MAV for Fine-Grained Chinese Hate Speech Recognition

This paper presents our system for CCL25-Eval Task 10, addressing Fine-Grained Chinese Hate Speech Recognition (FGCHSR). We propose a novel SRAG-MAV framework that synergistically integrates task reformulation(TR), Self-Retrieval-Augmented Generation (SRAG), and Multi-Round Accumulative Voting (MAV). Our method reformulates the quadruplet extraction task into triplet extraction, uses dynamic retrieval from the training set to create contextual prompts, and applies multi-round inference with voting to improve output stability and performance. Our system, based on the Qwen2.5-7B model, achieves a Hard Score of 26.66, a Soft Score of 48.35, and an Average Score of 37.505 on the STATE ToxiCN dataset, significantly outperforming baselines such as GPT-4o (Average Score 15.63) and fine-tuned Qwen2.5-7B (Average Score 35.365). The code is available at https://github.com/king-wang123/CCL25-SRAG-MAV.

cs.CL↗

arXiv2025

OASIS: Order-Augmented Strategy for Improved Code Search

Code embeddings capture the semantic representations of code and are crucial for various code-related large language model (LLM) applications, such as code search. Previous training primarily relies on optimizing the InfoNCE loss by comparing positive natural language (NL)-code pairs with in-batch negatives. However, due to the sparse nature of code contexts, training solely by comparing the major differences between positive and negative pairs may fail to capture deeper semantic nuances. To address this issue, we propose a novel order-augmented strategy for improved code search (OASIS). It leverages order-based similarity labels to train models to capture subtle differences in similarity among negative pairs. Extensive benchmark evaluations demonstrate that our OASIS model significantly outperforms previous state-of-the-art models focusing solely on major positive-negative differences. It underscores the value of exploiting subtle differences among negative pairs with order labels for effective code embedding training.

cs.CL↗

arXiv2025

Function-to-Style Guidance of LLMs for Code Translation

Large language models (LLMs) have made significant strides in code translation tasks. However, ensuring both the correctness and readability of translated code remains a challenge, limiting their effective adoption in real-world software development. In this work, we propose F2STrans, a function-to-style guiding paradigm designed to progressively improve the performance of LLMs in code translation. Our approach comprises two key stages: (1) Functional learning, which optimizes translation correctness using high-quality source-target code pairs mined from online programming platforms, and (2) Style learning, which improves translation readability by incorporating both positive and negative style examples. Additionally, we introduce a novel code translation benchmark that includes up-to-date source code, extensive test cases, and manually annotated ground-truth translations, enabling comprehensive functional and stylistic evaluations. Experiments on both our new benchmark and existing datasets demonstrate that our approach significantly improves code translation performance. Notably, our approach enables Qwen-1.5B to outperform prompt-enhanced Qwen-32B and GPT-4 on average across 20 diverse code translation scenarios.

cs.AI↗