arXiv subjects

Shuai Zhao

Publications and source records attributed to Shuai Zhao.

At least 37 records · Page 2Linked to original sources

arXiv2026

Heavy quark mass dependence of the $Λ_Q$ light-cone distribution amplitude in QCD

We study the heavy quark mass dependence of the leading-twist light-cone distribution amplitude (LCDA) of the $Λ_Q$ baryon in QCD. Starting from the factorization formula that relates the QCD LCDA to the boosted heavy-quark effective theory (bHQET) LCDA, we derive a first-order partial differential equation governing this mass dependence in the peak region. The equation is solved analytically, and the explicit factor connecting LCDAs at different heavy quark masses is presented. We further incorporate Borel-resummed perturbative corrections from a renormalon model into the factorization. The impact of these renormalon corrections on the mass dependence is studied, and a numerical analysis using a simple LCDA model is performed to illustrate the behavior and to assess the uncertainties arising from the corrections, thereby providing a numerical estimate of the associated power corrections to the mass dependence. Our results provide an essential tool for extrapolating lattice QCD calculations of heavy-baryon LCDAs from smaller simulated masses to the physical bottom quark mass.

hep-ph↗

arXiv2026

Accessing the HQET B-Meson Shape Function from a LaMET Quasi-Shape Function

The shape function and the light-cone distribution amplitude of heavy meson jointly characterize the nonperturbative structure of the heavy meson on the light-cone, with the former being essential for theoretical predictions of inclusive decays and the latter for exclusive decays. While first-principles lattice QCD results for the heavy meson LCDA have become available in recent years, lattice results for the shape function remain absent. In this work, we establish a two-step factorization scheme -- known as the HQLaMET framework -- for computing the $B$-meson shape function on the lattice, which fully disentangles the effects of the disparate scales $P_B^z$, $m_b$, and $Λ_{\textrm{QCD}}$. For illustration, starting from a phenomenological model for the shape function in HQET, we provide a graphical presentation of the entire procedure of this framework. The results of the current work lay the foundation for nonperturbative lattice QCD determinations of the shape function in the near future.

hep-ph↗

arXiv2026

A Comprehensive Survey of Direct Preference Optimization: Datasets, Theories, Variants, and Applications

With the rapid advancement of large language models (LLMs), aligning policy models with human preferences has become increasingly critical. Direct Preference Optimization (DPO) has emerged as a promising approach for alignment, acting as an RL-free alternative to Reinforcement Learning from Human Feedback (RLHF). Despite DPO's various advancements and inherent limitations, an in-depth review of these aspects is currently lacking in the literature. In this work, we present a comprehensive review of the challenges and opportunities in DPO, covering theoretical analyses, variants, relevant preference datasets, and applications. Specifically, we categorize recent studies on DPO based on key research questions to provide a thorough understanding of DPO's current landscape. Additionally, we propose several future research directions to offer insights on model alignment for the research community. An updated collection of relevant papers can be found on https://github.com/Mr-Loevan/DPO-Survey.

cs.AI↗

arXiv2026

ChronoPhyBench: Do MLLMs Truly Understand the World or Merely Exploit Language Priors?

Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated remarkable proficiency in open-world reasoning and understanding. However, a critical ambiguity persists: it remains unclear whether these models genuinely synthesize cross-modal information to construct physically grounded reasoning chains, or if they merely exploit strong language priors to mask single-modality reliance, thereby hallucinating advanced multimodal capabilities. Motivated by this, and to rigorously mitigate language modality bias and shortcuts, we propose a novel multimodal Chrono}logical Physical Dynamics Reasoning Benchmark ChronoPhyBench, which unifies next state prediction with Visual Question Answering (VQA) paradigms by conditioning on historical video context and textual captions to enforce models to deduce subsequent physical states through both single image selection and the inherently more complex task of multiple frame chronological sorting. Concurrently, we construct a large-scale multimodal reasoning dataset curated using the ChronoPhyBench criteria, comprising over 10,000 long-form videos paired with meticulously annotated captions, totaling 5M tokens. Our experimental evaluations reveal a stark contrast to conclusions drawn by previous benchmarks. The capacity of current open-source models to perform physically grounded multimodal reasoning remains in its infancy. Ultimately, this work seeks to systematically stress-test the reasoning capabilities of multimodal models, quantify hallucination rates, and advance the development of Physical AI, thereby providing the community with a robust and transparent evaluation framework toward Artificial General Intelligence (AGI).

cs.CV↗

arXiv2026

Understanding and Preventing Entropy Collapse in RLVR with On-Policy Entropy Flow Optimization

Reinforcement learning with verifiable rewards (RLVR) has become an effective paradigm for improving the reasoning ability of large language models. However, widely used RLVR algorithms, such as GRPO, often suffer from entropy collapse, leading to premature determinism and unstable optimization. Existing remedies, including entropy regularization and ratio-based clipping heuristics, either control entropy in a coarse-grained manner or rely on approximate on-policy training. In this paper, we revisit entropy collapse from a token-level entropy flow perspective. Our analysis reveals that entropy-decreasing tokens consistently outweigh entropy-increasing ones, resulting in a severely imbalanced entropy flow. This perspective provides a unified explanation of entropy collapse in existing RLVR algorithms and highlights the importance of balancing entropy dynamics. Motivated by this analysis, we propose On-Policy Entropy Flow Optimization (OPEFO), an adaptive entropy flow balancing mechanism that rescales entropy-increasing and entropy-decreasing updates according to their contributions to entropy change, while remaining strict on-policy. Experiments on six mathematical reasoning benchmarks demonstrate that OPEFO improves training stability and final performance. We will release the code and models upon publication.

cs.LG↗

arXiv2026

EduStory: A Unified Framework for Pedagogically-Consistent Multi-Shot STEM Instructional Video Generation

Long-horizon video generation has advanced in visual quality, yet existing methods still struggle to maintain knowledge consistency and coherent pedagogical narratives across multi-shot instructional videos, especially in STEM domains. To address these challenges, we propose EduStory, a unified framework for reliable instructional video generation. EduStory integrates pedagogical state modeling to track persistent knowledge states, script-guided structured control to organize multi-shot narratives, and learning-oriented evaluation metrics to assess knowledge fidelity and constraint satisfaction. To support rigorous evaluation, we further introduce EduVideoBench, a diagnostic benchmark with multi-granularity annotations, including pedagogical storyboards, shot-level semantics, and knowledge state transitions, together with baseline tasks for controllable instructional video generation. Extensive experiments demonstrate that domain-aware state modeling and structured control substantially reduce narrative breakdown and improve alignment with instructional intent. These results highlight the significance of domain-specific structural constraints and tailored benchmarks for advancing reliable, controllable, and also trustworthy long-horizon video generation.

cs.CV↗

arXiv2026

Self-Debias: Self-correcting for Debiasing Large Language Models

Although Large Language Models (LLMs) demonstrate remarkable reasoning capabilities, inherent social biases often cascade throughout the Chain-of-Thought (CoT) process, leading to continuous "Bias Propagation". Existing debiasing methods primarily focus on static constraints or external interventions, failing to identify and interrupt this propagation once triggered. To address this limitation, we introduce Self-Debias, a progressive framework designed to instill intrinsic self-correction capabilities. Specifically, we reformulate the debiasing process as a strategic resource redistribution problem, treating the model's output probability mass as a limited resource to be reallocated from biased heuristics to unbiased reasoning paths. Unlike standard preference optimization which applies broad penalties, Self-Debias employs a fine-grained trajectory-level objective subject to dynamic debiasing constraints. This enables the model to selectively revise biased reasoning suffixes while preserving valid contextual prefixes. Furthermore, we integrate an online self-improvement mechanism utilizing consistency filtering to autonomously synthesize supervision signals. With merely 20k annotated samples, Self-Debias activates efficient self-correction, achieving superior debiasing performance while preserving general reasoning capabilities without continuous external oversight.

cs.CL↗

arXiv2026

Continuum-Limit HQET LCDAs from Lattice QCD for Tightening B Decay Uncertainties

Heavy meson HQET light-cone distribution amplitudes (LCDAs) are critical for precision predictions of $B$ meson weak decays, but currently are one of dominant theoretical uncertainties that obscure interpretations of $B$ anomalies and CP-violating measurements. Building on the established HQLaMET framework, supplemented by lattice QCD calculations of the OPE moments, we present a precise lattice QCD calculation of HQET LCDAs by employing multi-ensemble simulations for continuum and physical pion mass extrapolation, quantifying comprehensive systematic errors, and validating results through OPE moment cross-validation. Details of the lattice calculations are provided in a companion paper \cite{HeavymesonDA_long_paper}. Our final results for key inverse moments (at $μ=1$ GeV) are $λ_B=0.340(20)$ GeV and $σ_B^{(1)}=1.685(63)$, with the total uncertainty reduced by a factor of three relative to the previous analysis. These results can greatly reduce the uncertainty in the $B \to K^*$ form factors in the large-recoil region. This work resolves the long-standing bottleneck in first-principles predictions of heavy meson LCDAs, advancing precision flavor physics to new frontiers.

hep-lat↗

arXiv2026

Determination of heavy meson light-cone distribution amplitudes: theoretical framework and lattice simulations

We present a first-principles determination of heavy meson light-cone distribution amplitudes (LCDAs) from lattice QCD in the continuum limit, improving substantially on our previous pioneering study. Within the heavy-quark large-momentum effective theory (HQLaMET) framework, supplemented by lattice QCD calculations of the OPE moments, we analyze six ensembles with lattice spacings ranging from $a=0.0519-0.1053$\,fm and pion masses from $m_π=135.5-317.2$\,MeV, thereby enabling controlled continuum, chiral, and infinite-momentum extrapolations to the physical point. Momentum-smeared sources, hypercubic-smeared Wilson lines, and optimized interpolating operators are adopted to significantly improved signals for the nonlocal correlators. Within a unified framework, we determine both QCD LCDAs and HQET LCDAs. Our resulting QCD LCDAs of $D$ meson peak at $y\approx 0.2-0.3$, with total uncertainties below $30\%$ for $0.1<y<0.9$. The leading-twist HQET LCDA is constructed using a peak-and-tail factorization, in which the nonperturbative peak region is obtained from lattice QCD and the perturbative tail is incorporated from HQET, with the two regions combined through a model-independent Laguerre-polynomial parametrization. At $μ=1$\,GeV, we obtain the inverse moment of HQET LCDA $λ_B=0.340(20)$\,GeV and first inverse-logarithmic moment $σ_B^{(1)}=1.685(63)$, consistent with experimental constraints and phenomenological determinations. Direct lattice calculations based on operator product expansion provide a nontrivial cross-check of the LaMET results. Final results and phenomenological impact of these results are presented in a companion paper~\cite{HeavymesonDA_short_paper}. Our results remove the single-lattice-spacing limitation of the previous study, and provide a robust determinations of heavy meson LCDAs in both QCD and HQET for next-generation heavy flavor physics.

hep-lat↗

arXiv2026

A Lightweight, Transferable, and Self-Adaptive Framework for Intelligent DC Arc-Fault Detection in Photovoltaic Systems

Arc-fault circuit interrupters (AFCIs) are essential for mitigating fire hazards in residential photovoltaic (PV) systems, yet achieving reliable DC arc-fault detection under real-world conditions remains challenging. Spectral interference from inverter switching, hardware heterogeneity, operating-condition drift, and environmental noise collectively compromise conventional AFCI solutions. This paper proposes a lightweight, transferable, and self-adaptive learning-driven framework (LD-framework) for intelligent DC arc-fault detection. At the device level, LD-Spec learns compact spectral representations enabling efficient on-device inference and near-perfect arc discrimination. Across heterogeneous inverter platforms, LD-Align performs cross-hardware representation alignment to ensure robust detection despite hardware-induced distribution shifts. To address long-term evolution, LD-Adapt introduces a cloud-edge collaborative self-adaptive updating mechanism that detects unseen operating regimes and performs controlled model evolution. Extensive experiments involving over 53,000 labeled samples demonstrate near-perfect detection, achieving 0.9999 accuracy and 0.9996 F1-score. Across diverse nuisance-trip-prone conditions, including inverter start-up, grid transitions, load switching, and harmonic disturbances, the method achieves a 0% false-trip rate. Cross-hardware transfer shows reliable adaptation using only 0.5%-1% labeled target data while preserving source performance. Field adaptation experiments demonstrate recovery of detection precision from 21% to 95% under previously unseen conditions. These results indicate that the LD-framework enables a scalable, deployment-oriented AFCI solution maintaining highly reliable detection across heterogeneous devices and long-term operation.

eess.SP↗

arXiv2026

Physics-Aligned Spectral Mamba: Decoupling Semantics and Dynamics for Few-Shot Hyperspectral Target Detection

Meta-learning facilitates few-shot hyperspectral target detection (HTD), but adapting deep backbones remains challenging. Full-parameter fine-tuning is inefficient and prone to overfitting, and existing methods largely ignore the frequency-domain structure and spectral band continuity of hyperspectral data, limiting spectral adaptation and cross-domain generalization.To address these challenges, we propose SpecMamba, a parameter-efficient and frequency-aware framework that decouples stable semantic representation from agile spectral adaptation. Specifically, we introduce a Discrete Cosine Transform Mamba Adapter (DCTMA) on top of frozen Transformer representations. By projecting spectral features into the frequency domain via DCT and leveraging Mamba's linear-complexity state-space recursion, DCTMA explicitly captures global spectral dependencies and band continuity while avoiding the redundancy of full fine-tuning. Furthermore, to address prototype drift caused by limited sample sizes, we design a Prior-Guided Tri-Encoder (PGTE) that allows laboratory spectral priors to guide the optimization of the learnable adapter without disrupting the stable semantic feature space. Finally, a Self-Supervised Pseudo-Label Mapping (SSPLM) strategy is developed for test-time adaptation, enabling efficient decision boundary refinement through uncertainty-aware sampling and dual-path consistency constraints. Extensive experiments on multiple public datasets demonstrate that SpecMamba consistently outperforms state-of-the-art methods in detection accuracy and cross-domain generalization.

cs.CV↗

arXiv2026

From Concept to Practice: an Automated LLM-aided UVM Machine for RTL Verification

Verification presents a major bottleneck in Integrated Circuit (IC) development, consuming nearly 70% of the total development effort. While the Universal Verification Methodology (UVM) is widely used in industry to improve verification efficiency through structured and reusable testbenches, constructing these testbenches and generating sufficient stimuli remain challenging. These challenges arise from the considerable manual coding effort required, repetitive manual execution of multiple EDA tools, and the need for in-depth domain expertise to navigate complex designs.Here, we present UVM^2, an automated verification framework that leverages Large Language Models (LLMs) to generate UVM testbenches and iteratively refine them using coverage feedback, significantly reducing manual effort while maintaining rigorous verification standards.To evaluate UVM^2, we introduce a benchmark suite comprising Register Transfer Level (RTL) designs of up to 1.6K lines of code.The results show that UVM^2 reduces testbench setup time by up to UVM^2 compared to experienced engineers, and achieve average code and function coverage of 87.44% and 89.58%, outperforming state-of-the-art solutions by 20.96% and 23.51%, respectively.

cs.AR↗

arXiv2026

Exploiting Dependency and Parallelism: Real-Time Scheduling and Analysis for GPU Tasks

With the rapid advancement of Artificial Intelligence, the Graphics Processing Unit (GPU) has become increasingly essential across a growing number of safety-critical application domains. Applying a GPU is indispensable for parallel computing; however, the complex data dependencies and resource contention across kernels within a GPU task may unpredictably delay its execution time. To address these problems, this paper presents a scheduling and analysis method for Directed Acyclic Graph (DAG)-structured GPU tasks. Given a DAG representation, the proposed scheduling scales the kernel-level parallelism and establishes inter-kernel dependencies to provide a reduced and predictable DAG response time. The corresponding timing analysis yields a safe yet nonpessimistic makespan bound without any assumption on kernel priorities. The proposed method is implemented using the standard CUDA API, requiring no additional software or hardware support. Experimental results under synthetic and real-world benchmarks demonstrate that the proposed approach effectively reduces the worst-case makespan and measured task execution time compared to the existing methods up to 32.8% and 21.3%, respectively.

cs.OS↗

arXiv2026

Experimental Phase-Matching Quantum Cryptographic Conferencing in Symmetric and Asymmetric Fiber Channels

Quantum cryptographic conferencing (QCC) allows multiple parties to establish common secure keys in quantum networks with information-theoretic security. However, the secure transmission distances of current QCC implementations are still limited to the metropolitan areas. Here, we experimentally demonstrate the three-intensity phase-matching (PM) QCC protocol considering finite-size effects by employing frequency-locking and phase-tracking techniques for three parties. The key distribution capability of the PM QCC protocol is demonstrated in the symmetric fiber channels with the distance from each party to the measurement site up to 100 km. The network adaptability of the PM QCC protocol is demonstrated in asymmetric fiber channels used to simulate fiber channel configurations in real networks. Thus, the feasibility of applying the PM QCC protocol to practical intercity quantum networks with both symmetric and asymmetric channels is verified.

quant-ph↗

arXiv2026

Global Context Compression with Interleaved Vision-Text Transformation

Recent achievements of vision-language models in end-to-end OCR point to a new avenue for low-loss compression of textual information. This motivates earlier works that render the Transformer's input into images for prefilling, which effectively reduces the number of tokens through visual encoding, thereby alleviating the quadratically increased Attention computations. However, this partial compression fails to save computational or memory costs at token-by-token inference. In this paper, we investigate global context compression, which saves tokens at both prefilling and inference stages. Consequently, we propose VIST2, a novel Transformer that interleaves input text chunks alongside their visual encoding, while depending exclusively on visual tokens in the pre-context to predict the next text token distribution. Around this idea, we render text chunks into sketch images and train VIST2 in multiple stages, starting from curriculum-scheduled pretraining for optical language modeling, followed by modal-interleaved instruction tuning. We conduct extensive experiments using VIST2 families scaled from 0.6B to 8B to explore the training recipe and hyperparameters. With a 4$\times$ compression ratio, the resulting models demonstrate significant superiority over baselines on long writing tasks, achieving, on average, a 3$\times$ speedup in first-token generation, 77% reduction in memory usage, and 74% reduction in FLOPS. Our codes and datasets will be public to support further studies.

cs.CV↗

arXiv2026

Testing a Linear Relation: Short-Range Correlations and the EMC Effect for Gluons and Quarks in Nuclei

In this work, we focus on the possible linear relation between short-range correlations (SRCs) and the EMC effect for partons in nuclei. First, we test a linear relationship pertaining to gluons in bound nuclei; it is manifested as a correlation between the slope of the reduced cross section ratio in deep inelastic scattering (DIS) and the cross section of sub-threshold $J/ψ$ photoproduction. For comparison, the results from four different global analyses groups of nuclear parton distribution functions (nPDFs) are utilized. These results show a good linear correlation between the gluons in bound nuclei and the slope of the reduced cross section ratio, consistent with the possible presence of nuclear effects in the gluon distributions. Second, we investigate the linear relationship of quarks in the proton-induced Drell-Yan process. The corresponding results for quarks show strong sensitivity to the parameterization forms adopted by the different groups. These findings enhance our understanding of the substructure in bound nuclei and provide valuable reference for future global fitting of nPDFs.

hep-ph↗

arXiv2025

Exploring Nucleon Structure through Sub- and Near-threshold $ϕ$-Meson Photoproduction

We propose to investigate short-range correlations (SRCs) in nuclei by studying sub and near-threshold photoproduction of $ϕ$-meson. The feasibility of exploring this nuclear structure in electron-positron collision experiments such as Beijing Spectrometer III (BESIII) and Super $τ$-Charm Facility (STCF) was investigated by utilizing their beryllium-based ($^{9}$Be) beam pipe. The cross sections for these processes as well as the likelihood of detection by BESIII and STCF are calculated. Due to the limited number of events, BESIII is inadequate for the proposed measurement, while the future STCF provides a higher probability of successfully measuring this process. These findings suggest that electron-positron collider could play a valuable role in elucidating the fundamental physics behind the nuclear modification of parton distribution functions. This proposed analysis of photon-nucleon interactions represents uncharted territory, promising fresh prospects for applications in both particle and nuclear physics.

hep-ph↗

arXiv2025

C2PO: Diagnosing and Disentangling Bias Shortcuts in LLMs

Bias in Large Language Models (LLMs) poses significant risks to trustworthiness, manifesting primarily as stereotypical biases (e.g., gender or racial stereotypes) and structural biases (e.g., lexical overlap or position preferences). However, prior paradigms typically address these in isolation, often mitigating one at the expense of exacerbating the other. To address this, we conduct a systematic exploration of these reasoning failures and identify a primary inducement: the latent spurious feature correlations within the input that drive these erroneous reasoning shortcuts. Driven by these findings, we introduce Causal-Contrastive Preference Optimization (C2PO), a unified alignment framework designed to tackle these specific failures by simultaneously discovering and suppressing these correlations directly within the optimization process. Specifically, C2PO leverages causal counterfactual signals to isolate bias-inducing features from valid reasoning paths, and employs a fairness-sensitive preference update mechanism to dynamically evaluate logit-level contributions and suppress shortcut features. Extensive experiments across multiple benchmarks covering stereotypical bias (BBQ, Unqover), structural bias (MNLI, HANS, Chatbot, MT-Bench), out-of-domain fairness (StereoSet, WinoBias), and general utility (MMLU, GSM8K) demonstrate that C2PO effectively mitigates stereotypical and structural biases while preserving robust general reasoning capabilities.

cs.CL↗