arXiv subjects

Yang Liu

Publications and source records attributed to Yang Liu.

At least 793 records · Page 44Linked to original sources

arXiv2025

ViC-Bench: Benchmarking Visual-Interleaved Chain-of-Thought Capability in MLLMs with Free-Style Intermediate State Representations

Visual-Interleaved Chain-of-Thought (VI-CoT) enables Multi-modal Large Language Models (MLLMs) to continually update their understanding and decision space based on step-wise intermediate visual states (IVS), much like a human would, which has demonstrated impressive success in various tasks, thereby leading to emerged advancements in related downstream benchmarks. Despite promising progress, current benchmarks provide models with relatively fixed IVS, rather than free-style IVS, whch might forcibly distort the original thinking trajectories, failing to evaluate their intrinsic reasoning capabilities. More importantly, existing benchmarks neglect to systematically explore the impact factors that IVS would impart to the untamed reasoning performance. To tackle above gaps, we introduce a specialized benchmark termed ViC-Bench, consisting of four representive tasks, i.e., maze navigation, jigsaw puzzle, embodied long-horizon planning, as well as complex counting, where each task has dedicated free-style IVS generation pipeline supporting adaptive function calls. To systematically examine VI-CoT capability, we propose a thorough evaluation suite incorporating a progressive three-stage strategy with targeted new metrics. Besides, we establish Incremental Prompting Information Injection strategy to ablatively explore the prompting factors for VI-CoT. We extensively conduct evaluations for 18 advanced MLLMs, revealing key insights into their VI-CoT capability. The introduced ViC-Bench has been made publicly available at Huggingface.

cs.CV↗

arXiv2025

Non-SUSY physics and the Atiyah-Singer index theorem

The Atiyah-Singer index theorem, a cornerstone of modern mathematics, has traditionally been derived from supersymmetric (SUSY) physics. This paper demonstrates a direct derivation from non-supersymmetric quantum statistics by establishing a fundamental correspondence: the grand partition functions of non-interacting bosonic and fermionic systems are precisely the Chern characters of certain vector bundles. Furthermore, we generalize this correspondence to infinite dimensions, where we construct a novel mathematical framework of spectral-sheaf pairs. Within this framework, we formulate a generalized index theorem, identifying the topological index with a regularized spectral product. This work not only circumvents the need for supersymmetry but also provides a deeper unifying perspective, revealing quantum statistics as a sufficient foundation for topological invariants.

math-ph↗

arXiv2025

Evaluating Large Language Models for Line-Level Vulnerability Localization

Recently, Automated Vulnerability Localization (AVL) has attracted growing attention, aiming to facilitate diagnosis by pinpointing the specific lines of code responsible for vulnerabilities. Large Language Models (LLMs) have shown potential in various domains, yet their effectiveness in line-level vulnerability localization remains underexplored. In this work, we present the first comprehensive empirical evaluation of LLMs for AVL. Our study examines 19 leading LLMs suitable for code analysis, including ChatGPT and multiple open-source models, spanning encoder-only, encoder-decoder, and decoder-only architectures, with model sizes from 60M to 70B parameters. We evaluate three paradigms including few-shot prompting, discriminative fine-tuning, and generative fine-tuning with and without Low-Rank Adaptation (LoRA), on both a BigVul-derived dataset for C/C++ and a smart contract vulnerability dataset.} Our results show that discriminative fine-tuning achieves substantial performance gains over existing learning-based AVL methods when sufficient training data is available. In low-data settings, prompting advanced LLMs such as ChatGPT proves more effective. We also identify challenges related to input length and unidirectional context during fine-tuning, and propose two remedial strategies: a sliding window approach and right-forward embedding, both of which yield significant improvements. Moreover, we provide the first assessment of LLM generalizability in AVL, showing that certain models can transfer effectively across Common Weakness Enumerations (CWEs) and projects. However, performance degrades notably for newly discovered vulnerabilities containing unfamiliar lexical or structural patterns, underscoring the need for continual adaptation.

cs.SE↗

arXiv2025

Casting a SPELL: Sentence Pairing Exploration for LLM Limitation-breaking

Large language models (LLMs) have revolutionized software development through AI-assisted coding tools, enabling developers with limited programming expertise to create sophisticated applications. However, this accessibility extends to malicious actors who may exploit these powerful tools to generate harmful software. Existing jailbreaking research primarily focuses on general attack scenarios against LLMs, with limited exploration of malicious code generation as a jailbreak target. To address this gap, we propose SPELL, a comprehensive testing framework specifically designed to evaluate the weakness of security alignment in malicious code generation. Our framework employs a time-division selection strategy that systematically constructs jailbreaking prompts by intelligently combining sentences from a prior knowledge dataset, balancing exploration of novel attack patterns with exploitation of successful techniques. Extensive evaluation across three advanced code models (GPT-4.1, Claude-3.5, and Qwen2.5-Coder) demonstrates SPELL's effectiveness, achieving attack success rates of 83.75%, 19.38%, and 68.12% respectively across eight malicious code categories. The generated prompts successfully produce malicious code in real-world AI development tools such as Cursor, with outputs confirmed as malicious by state-of-the-art detection systems at rates exceeding 73%. These findings reveal significant security gaps in current LLM implementations and provide valuable insights for improving AI safety alignment in code generation applications.

cs.CR↗

arXiv2025

Hallucination Detection for LLM-based Text-to-SQL Generation via Two-Stage Metamorphic Testing

In Text-to-SQL generation, large language models (LLMs) have shown strong generalization and adaptability. However, LLMs sometimes generate hallucinations, i.e.,unrealistic or illogical content, which leads to incorrect SQL queries and negatively impacts downstream applications. Detecting these hallucinations is particularly challenging. Existing Text-to-SQL error detection methods, which are tailored for traditional deep learning models, face significant limitations when applied to LLMs. This is primarily due to the scarcity of ground-truth data. To address this challenge, we propose SQLHD, a novel hallucination detection method based on metamorphic testing (MT) that does not require standard answers. SQLHD splits the detection task into two sequentiial stages: schema-linking hallucination detection via eight structure-aware Metamorphic Relations (MRs) that perturb comparative words, entities, sentence structure or database schema, and logical-synthesis hallucination detection via nine logic-aware MRs that mutate prefix words, extremum expressions, comparison ranges or the entire database. In each stage the LLM is invoked separately to generate schema mappings or SQL artefacts; the follow-up outputs are cross-checked against their source counterparts through the corresponding MRs, and any violation is flagged as a hallucination without requiring ground-truth SQL. The experimental results demonstrate our method's superior performance in terms of the F1-score, which ranges from 69.36\% to 82.76\%. Additionally, SQLHD demonstrates superior performance over LLM Self-Evaluation methods, effectively identifying hallucinations in Text-to-SQL tasks.

cs.SE↗

arXiv2025

Detecting Non-Optimal Decisions of Embodied Agents via Diversity-Guided Metamorphic Testing

As embodied agents advance toward real-world deployment, ensuring optimal decisions becomes critical for resource-constrained applications. Current evaluation methods focus primarily on functional correctness, overlooking the non-functional optimality of generated plans. This gap can lead to significant performance degradation and resource waste. We identify and formalize the problem of Non-optimal Decisions (NoDs), where agents complete tasks successfully but inefficiently. We present NoD-DGMT, a systematic framework for detecting NoDs in embodied agent task planning via diversity-guided metamorphic testing. Our key insight is that optimal planners should exhibit invariant behavioral properties under specific transformations. We design four novel metamorphic relations capturing fundamental optimality properties: position detour suboptimality, action optimality completeness, condition refinement monotonicity, and scene perturbation invariance. To maximize detection efficiency, we introduce a diversity-guided selection strategy that actively selects test cases exploring different violation categories, avoiding redundant evaluations while ensuring comprehensive diversity coverage. Extensive experiments on the AI2-THOR simulator with four state-of-the-art planning models demonstrate that NoD-DGMT achieves violation detection rates of 31.9% on average, with our diversity-guided filter improving rates by 4.3% and diversity scores by 3.3 on average. NoD-DGMT significantly outperforms six baseline methods, with 16.8% relative improvement over the best baseline, and demonstrates consistent superiority across different model architectures and task complexities.

cs.SE↗

arXiv2025

A Pick function approach for designing energy-decay preserving schemes of the Maxwell equations in Havriliak-Negami dispersive media

This work proposes a novel approach for designing high-order energy-decaying schemes for Maxwell's equations in Havriliak-Negami dispersive media. It is shown that conventional convolution quadrature (CQ) methods, which rely directly on the generating function of linear multistep methods, cannot generate completely monotonic sequences beyond first-order accuracy. We rigorously prove that for any linear multistep method of second-or higher-order, the associated generating function $δ(ζ)$ cannot satisfy both that $-δ(ζ)$ is a Pick function and that it is analytic on $(-\infty,1)$ - a key requirement for constructing completely monotonic sequences. To overcome this fundamental limitation, we introduce a reconstruction of the generating function's structure. By strategically incorporating the theory of Pick functions, we successfully construct a second-order completely monotonic sequence. This theoretical advance leads to a discrete scheme that inherits the continuous model's energy decay property, guaranteeing unconditional stability. Numerical experiments confirm the convergence rates and energy dissipation behavior of the proposed method.

math.NA↗

arXiv2025

MCGA: Mixture of Codebooks Hyperspectral Reconstruction via Grayscale-Aware Attention

Reconstructing hyperspectral images (HSIs) from RGB inputs provides a cost-effective alternative to hyperspectral cameras, but reconstructing high-dimensional spectra from three channels is inherently ill-posed. Existing methods typically directly regress RGB-to-HSI mappings using large attention networks, which are computationally expensive and handle ill-posedness only implicitly. We propose MCGA, a Mixture-of-Codebooks with Grayscale-aware Attention framework that explicitly addresses these challenges using spectral priors and photometric consistency. MCGA first learns transferable spectral priors via a mixture-of-codebooks (MoC) from heterogeneous HSI datasets, then aligns RGB features with these priors through grayscale-aware photometric attention (GANet). Efficiency and robustness are further improved via top-K attention design and test-time adaptation (TTA). Experiments on multiple real-world benchmarks demonstrate the state-of-the-art accuracy, strong cross-dataset generalization, and 4-5x faster inference. Codes will be available once acceptance at https://github.com/Fibonaccirabbit/MCGA.

cs.CV↗

arXiv2025

Efficient Implementation of Third-Order Tensor Methods with Adaptive Regularization for Unconstrained Optimization

High-order tensor methods that employ local Taylor models of degree $p$ within adaptive regularization frameworks (AR$p$) have recently received significant attention, due to their optimal/improved global and local rates of convergence, for both convex and nonconvex optimization problems. In this paper, we showcase the numerical performance of standard second- and third-order variants ($p=2,3$) and propose novel techniques for key algorithmic aspects when $p\geq 3$. In particular, we extend the interpolation-based updating strategy for the regularization parameter introduced in [Gould, Porcelli and Toint, Comput Optim Appl (2012) 53:1--22] for $p=2$, to the case when $p \geq 3$. We identify fundamental differences between the different local minima of the regularised subproblems for $p=2$ and $p \geq 3$ and their effect on algorithm performance. For $p\geq 3$, we introduce a novel pre-rejection technique that rejects poor/unsuccessful subproblem minimizers prior to any function evaluation. Numerical studies showcase the efficiency improvements generated by our proposed modifications of the AR$3$ algorithm. We also assess numerically, the effect of different subproblem termination conditions and the choice of the initial regularization parameter on the overall algorithm performance. Finally, we benchmark our best-performing AR$3$ variants, as well as those in [Birgin et al., Optim Lett (2020) 14:815--838], against second-order ones (AR$2$). Encouraging results on standard test problems are obtained, confirming that AR$3$ variants can be made to outperform second-order variants in terms of objective evaluations, derivative evaluations, and number of subproblem solves. We provide an efficient, extensive and modular software package in MATLAB that includes many AR$2$ and AR$3$ variants, including Hessian- and tensor-free ones, allowing ease of use and experimentation for interested users.

math.OC↗

arXiv2025

Solar jet-induced perturbation propagating through coronal loops and in-loop electron beam transport indicated by type II and type N radio bursts

Solar type II radio bursts are commonly attributed to coronal shocks driven by coronal mass ejections (CMEs). However, some metric type II bursts have occasionally been reported to occur in the absence of a CME and to be associated with weak solar activities. This study aims to identify the driver of the coronal shock in this kind of type II event. We investigate a high-frequency metric type II burst with clear band splitting, observed simultaneously by the Chashan Broadband Solar radio spectrograph and the Nançay Radioheliograph. It is associated with a C3.1-class flare and a small-scale jet, but without a detectable CME in the coronagraphs. The type II burst is preceded by multiple type III bursts, one of which exhibits characteristics of a type N burst. The type II burst source is associated with the jet-induced perturbation front propagating through nearby closed loops at a speed of $\sim$880 km s$^{-1}$, rather than the much slower jet front. This suggests that the disturbance initiated by the jet can convert to a shock wave within low Alfvénic coronal loops, providing the necessary conditions for electron acceleration and subsequent radio emission. Our findings offer new insights into the formation mechanism of high-frequency type II bursts associated with weak flares and jets.

astro-ph.SR↗

arXiv2025

LVAgent: Long Video Understanding by Multi-Round Dynamical Collaboration of MLLM Agents

Existing MLLMs encounter significant challenges in modeling the temporal context within long videos. Currently, mainstream Agent-based methods use external tools to assist a single MLLM in answering long video questions. Despite such tool-based support, a solitary MLLM still offers only a partial understanding of long videos, resulting in limited performance. In order to better address long video tasks, we introduce LVAgent, the first framework enabling multi-round dynamic collaboration of MLLM agents in long video understanding. Our method consists of four key steps: 1) Selection: We pre-select appropriate agents from the model library to form optimal agent teams based on different tasks. 2) Perception: We design an effective retrieval scheme for long videos to improve the coverage of critical temporal segments while maintaining computational efficiency. 3) Action: Agents answer long video questions and exchange reasons. 4) Reflection: We evaluate each agent's performance in each round of discussion and optimize the agent team for dynamic collaboration. The agents iteratively refine their answers by multi-round dynamical collaboration of MLLM agents. LVAgent is the first agent system method that outperforms all closed-source models (like GPT-4o) and open-source models (like InternVL-2.5 and Qwen2-VL) in the long video understanding tasks. Our LVAgent achieves an accuracy of 80\% on four mainstream long video understanding tasks. Notably, LVAgent improves accuracy by 13.3\% on LongVideoBench. Code is available at https://github.com/64327069/LVAgent.

cs.CV↗

arXiv2025

Stability of inverse boundary value problem for the fourth-order Schrödinger equation

This paper is concerned with the stability of the inverse boundary value problem for the perturbed fourth-order Schrödinger equation in a bounded domain with Cauchy data. We establish stability results for the perturbed potential relying on boundary measurements. The estimates depend on various a priori information regarding the regularity and the support of the inhomogeneity. The proof primarily utilizes the complex geometric optics solution method and Fourier analysis.

math.AP↗

arXiv2025

Listening to the Mind: Earable Acoustic Sensing of Cognitive Load

Earable acoustic sensing offers a powerful and non-invasive modality for capturing fine-grained auditory and physiological signals directly from the ear canal, enabling continuous and context-aware monitoring of cognitive states. As earable devices become increasingly embedded in daily life, they provide a unique opportunity to sense mental effort and perceptual load in real time through auditory interactions. In this study, we present the first investigation of cognitive load inference through auditory perception using acoustic signals captured by off-the-shelf in-ear devices. We designed speech-based listening tasks to induce varying levels of cognitive load, while concurrently embedding acoustic stimuli to evoke Stimulus Frequency Otoacoustic Emission (SFOAEs) as a proxy for cochlear responsiveness. Statistical analysis revealed a significant association (p < 0.01) between increased cognitive load and changes in auditory sensitivity, with 63.2% of participants showing peak sensitivity at 3 kHz. Notably, sensitivity patterns also varied across demographic subgroups, suggesting opportunities for personalized sensing. Our findings demonstrate that earable acoustic sensing can support scalable, real-time cognitive load monitoring in natural settings, laying a foundation for future applications in augmented cognition, where everyday auditory technologies adapt to and support the users mental health.

cs.HC↗

arXiv2025

HUTFormer: Hierarchical U-Net Transformer for Long-Term Traffic Forecasting

Traffic forecasting, which aims to predict traffic conditions based on historical observations, has been an enduring research topic and is widely recognized as an essential component of intelligent transportation. Recent proposals on Spatial-Temporal Graph Neural Networks~(STGNNs) have made significant progress by combining sequential models with graph convolution networks. However, due to high complexity issues, STGNNs only focus on short-term traffic forecasting (e.g., 1-h ahead), while ignoring more practical long-term forecasting. In this paper, we make the first attempt to explore long-term traffic forecasting (e.g., 1-day ahead). To this end, we first reveal its unique challenges in exploiting multi-scale representations. Then, we propose a novel Hierarchical U-Net TransFormer~(HUTFormer) to address the issues of long-term traffic forecasting. HUTFormer consists of a hierarchical encoder and decoder to jointly generate and utilize multi-scale representations of traffic data. Specifically, for the encoder, we {\color{black}propose} window self-attention and segment merging to extract multi-scale representations from long-term traffic data. For the decoder, we design a cross-scale attention mechanism to effectively incorporate multi-scale representations. In addition, HUTFormer employs an efficient input embedding strategy to address the complexity issues. Extensive experiments on four traffic datasets show that the proposed HUTFormer significantly outperforms state-of-the-art traffic forecasting and long time series forecasting baselines.

cs.LG↗

arXiv2025

An Anatomy of Vision-Language-Action Models: From Modules to Milestones and Challenges

Vision-Language-Action (VLA) models are driving a revolution in robotics, enabling machines to understand instructions and interact with the physical world. This field is exploding with new models and datasets, making it both exciting and challenging to keep pace with. This survey offers a clear and structured guide to the VLA landscape. We design it to follow the natural learning path of a researcher: we start with the basic Modules of any VLA model, trace the history through key Milestones, and then dive deep into the core Challenges that define recent research frontier. Our main contribution is a detailed breakdown of the five biggest challenges in: (1) Representation, (2) Execution, (3) Generalization, (4) Safety, and (5) Dataset and Evaluation. This structure mirrors the developmental roadmap of a generalist agent: establishing the fundamental perception-action loop, scaling capabilities across diverse embodiments and environments, and finally ensuring trustworthy deployment-all supported by the essential data infrastructure. For each of them, we review existing approaches and highlight future opportunities. We position this paper as both a foundational guide for newcomers and a strategic roadmap for experienced researchers, with the dual aim of accelerating learning and inspiring new ideas in embodied intelligence. A live version of this survey, with continuous updates, is maintained on our \href{https://suyuz1.github.io/VLA-Survey-Anatomy/}{project page}.

cs.RO↗

arXiv2025

Probing Scientific General Intelligence of LLMs with Scientist-Aligned Workflows

Despite advances in scientific AI, a coherent framework for Scientific General Intelligence (SGI)-the ability to autonomously conceive, investigate, and reason across scientific domains-remains lacking. We present an operational SGI definition grounded in the Practical Inquiry Model (PIM: Deliberation, Conception, Action, Perception) and operationalize it via four scientist-aligned tasks: deep research, idea generation, dry/wet experiments, and experimental reasoning. SGI-Bench comprises over 1,000 expert-curated, cross-disciplinary samples inspired by Science's 125 Big Questions, enabling systematic evaluation of state-of-the-art LLMs. Results reveal gaps: low exact match (10--20%) in deep research despite step-level alignment; ideas lacking feasibility and detail; high code executability but low execution result accuracy in dry experiments; low sequence fidelity in wet protocols; and persistent multimodal comparative-reasoning challenges. We further introduce Test-Time Reinforcement Learning (TTRL), which optimizes retrieval-augmented novelty rewards at inference, enhancing hypothesis novelty without reference answer. Together, our PIM-grounded definition, workflow-centric benchmark, and empirical insights establish a foundation for AI systems that genuinely participate in scientific discovery.

cs.AI↗

arXiv2025

Asymptotics of Systemic Risk in a Renewal Model with Multiple Business Lines and Heterogeneous Claims

Systemic risk is receiving increasing attention in the insurance industry. In this paper, we propose a multi-dimensional Lévy process-based renewal risk model with heterogeneous insurance claims, where every dimension indicates a business line of an insurer. We use the systemic expected shortfall (SES) and marginal expected shortfall (MES) defined with a Value-at-Risk (VaR) target level as the measurement of systemic risk. Assuming that all the claim sizes are pairwise asymptotically independent (PAI), we derive asymptotic formulas for the tail probabilities of discounted aggregate claims and the total loss, which hold uniformly for all time horizons. We further obtain the asymptotics of the above systemic risk measures. The main technical issues involve the treatment of uniform convergence in the dynamic time setting. Finally, we perform a detailed Monte Carlo study to validate our asymptotics and analyze the impact and sensitivity of key parameters in the asymptotic expressions both analytically and numerically.

q-fin.RM↗

arXiv2025

VideoMem: Enhancing Ultra-Long Video Understanding via Adaptive Memory Management

Ultra long video understanding remains an open challenge, as existing vision language models (VLMs) falter on such content due to limited context length and inefficient long term memory retention. To address this, recent works have attempted to construct external knowledge bases and corresponding retrieval agumented generation (RAG) systems, yet these incur enormous storage and computational overhead. In this paper, we propose VideoMem, a novel framework that pioneers models long video understanding as a sequential generation task via adaptive memory management. Specifically, VideoMem dynamically updates a global memory buffer, which adaptively retains critical information while discarding redundant content across the video timeline. To efficiently train VLMs for such long-term tasks, VideoMem integrates the Progressive Grouped Relative Policy Optimization (PRPO) algorithm, equipped with two core modules: Progressive State Propagation (PSP) adaptively retains valid current states, propagates them to the next rollout step, and gradually narrows the model exploration space. Temporal Cascading Reward (TCR) further alleviates reward sparsity, improving sample utilization and accelerating convergence. Extensive experiments demonstrate that VideoMem significantly outperforms existing open-source models across diverse benchmarks for ultra-long video understanding tasks.

cs.CV↗