arXiv subjects

Luyang Li

Publications and source records attributed to Luyang Li.

9 recordsLinked to original sources

arXiv2026

Making Agent-Mediated Contributions Governable: A Project-Level Governance Manifest for Open-Source AI Collaboration

Generative AI and coding agents are intensifying a central governance tension in open-source software (OSS): they scale contribution generation faster than maintainers can assess risk, evidence, and accountability. Existing responses improve agent-readability and traceability, but project rules must also organize contribution-specific risk, evidence, accountability, and review-gate states. We theorize this organizational arrangement as project-side governability infrastructure. A diagnostic audit of 50 GitHub repositories finds widespread general governance artifacts, observable agent-readability, and fragmented AI-governance cues, but no project-wide arrangement that coordinates shared rules, preparation obligations, verification rights, and maintainer decision authority across AI-mediated contribution workflows. We develop the Agent Governance Manifest (AGM) as a repository-hosted boundary resource and bidirectional governance contract linking contributor-side evidence preparation with maintainer-side verification. In a controlled reviewer-side evaluation with 15 participants and 75 task-level outputs, AGM-supported materials improved exact risk-label recovery (37/38 vs. 15/37) and perceived review support (6.14 vs. 3.27 on a 1-7 scale). In a contributor-side feasibility check, 15 participants completed 45 tasks; all final packages represented the core governance state correctly, and 41 passed strict structural validation. The study develops a three-layer framework of agent-readability, traceability, and governability, theorizes agent-mediated contributions as governable boundary objects, and advances compliance-enabling digital innovation governance while preserving maintainer decision authority.

cs.SE↗

arXiv2026

JW-SSD: A Multimodal Benchmark Dataset for Fine-Grained Sunspot Classification

Accurate sunspot classification is essential for assessing the eruptive potential of solar active regions and forecasting space weather. We present JW-SSD, a high-quality multimodal benchmark dataset for fine-grained magnetic-type classification of sunspots. Constructed from SDO/HMI SHARP 720s data (2010-2023, Solar Cycles 24 and 25), JW-SSD comprises 36,553 co-registered magnetogram-continuum pairs from 2,507 active regions. Unlike conventional three-class schemes, JW-SSD refines the Mount Wilson classification into five physically meaningful categories ({\alpha}, \b{eta}, \b{eta}-{\delta}, \b{eta}-{\gamma}, \b{eta}-{\gamma}-{\delta}), enabling finer characterization of magnetic complexity. Rigorous quality control-including central meridian distance restriction, saturation filtering, and sharpness screening-ensures high data validity. The dataset is provided in both FITS and PNG formats, with standard training (29,243) and test (7,310) splits. Benchmark experiments with four representative architectures (U-Net, ResNet-50, EfficientNet-B0, and ViT-Small) yield high accuracy across all models (89.43%-94.78% on the three-class task), confirming that the dataset is reliably learnable across diverse modeling paradigms. JW-SSD has further been employed to train JW-SunSpot, a multimodal large language model that achieves the highest classification accuracy, demonstrating the dataset's broad applicability to both conventional networks and large-language-model-based approaches.

astro-ph.SR↗

arXiv2025

Mass spectra and Mott transitions of neutral mesons at finite temperature and magnetic field in frame of three-flavor Polyakov-extended Nambu-Jona-Lasino model

Mass spectra and Mott transitions of neutral mesons $K_0,{\bar K}_0,\pi_0,\eta,\eta'$ at finite temperature and magnetic field are investigated in a three-flavor PNJL model. We focus on the effect of gluons, which is simulated by the Polyakov potential, and the inverse magnetic catalysis (IMC) effect, which is mimicked by using a magnetic field dependent parameter. Mass spectra show similar structure when introducing the gluon and IMC effect. The mass of $K_0\ ({\bar K}_0)$ meson $m_{K_0}=m_{{\bar K}_0}$ is controlled by chiral symmetry breaking and restoration. It increases with temperature in the low temperature region, and shows a mass jump at the Mott transition. Further increasing temperature, $m_{K_0}$ firstly decreases and then increases with temperature. $\pi_0$ meson is not only the pseudo-Goldstone boson of chiral symmetry breaking, but also influenced by the flavor mixing of $\pi_0-\eta-\eta'$. The behavior of $m_{\pi_0}$ is different from $m_{K_0}$ only at high temperature region, which decreases with temperature. $\eta,\eta'$ mesons are affected by both the $U_A(1)$ anomaly and the flavor mixing of $\pi_0-\eta-\eta'$. The mass of $\eta$ meson $m_{\eta}$ decreases with temperature in low temperature region and then shows a jump at its Mott transition. After that $m_{\eta}$ firstly decreases and later increases with temperature. $\eta'$ meson is a resonant state, and its mass $m_{\eta'}$ continuously decreases and then increases with temperature. The mass jumps of $K_0,{\bar K}_0,\pi_0,\eta$ mesons are caused by the dimension reduction of the constituent quarks under external magnetic field. In PNJL model, the Mott transition temperature of $K_0,{\bar K}_0,\pi_0$ mesons ($\eta$ meson) decreases (increases) with magnetic field. The IMC effect leads to no qualitative change to the meson Mott transition temperature but shifts them to the lower values.

hep-ph↗

arXiv2025

Implicit Bias in LLMs: A Survey

Due to the implement of guardrails by developers, Large language models (LLMs) have demonstrated exceptional performance in explicit bias tests. However, bias in LLMs may occur not only explicitly, but also implicitly, much like humans who consciously strive for impartiality yet still harbor implicit bias. The unconscious and automatic nature of implicit bias makes it particularly challenging to study. This paper provides a comprehensive review of the existing literature on implicit bias in LLMs. We begin by introducing key concepts, theories and methods related to implicit bias in psychology, extending them from humans to LLMs. Drawing on the Implicit Association Test (IAT) and other psychological frameworks, we categorize detection methods into three primary approaches: word association, task-oriented text generation and decision-making. We divide our taxonomy of evaluation metrics for implicit bias into two categories: single-value-based metrics and comparison-value-based metrics. We classify datasets into two types: sentences with masked tokens and complete sentences, incorporating datasets from various domains to reflect the broad application of LLMs. Although research on mitigating implicit bias in LLMs is still limited, we summarize existing efforts and offer insights on future challenges. We aim for this work to serve as a clear guide for researchers and inspire innovative ideas to advance exploration in this task.

cs.CL↗

arXiv2023

Inverse magnetic catalysis effect and current quark mass effect on mass spectra and Mott transitions of pions under external magnetic field

Mass spectra and Mott transition of pions $(π^0,\ π^\pm)$ at finite temperature and magnetic field are investigated in a two-flavor NJL model, and we focus on the inverse magnetic catalysis (IMC) effect and current quark mass (CQM) effect. Due to the dimension reduction of the constituent quarks, the pion masses jump at their Mott transitions, which is independent of the IMC effect and CQM effect. We consider the IMC effect by using a magnetic dependent coupling constant, which is a monotonic decreasing function of magnetic field. With IMC effect, the Mott transition temperature of $π^0$ meson $T_m^0$ is a monotonic decreasing function of magnetic field. For charged pions $π^{\pm}$, the Mott transition temperature $T_m^+$ fast increases in weak magnetic field region and then decreases with magnetic field, which are accompanied with some oscillations. Comparing with the case without IMC effect, $T_m^0$ and $T_m^+$ are lower when including IMC effect. CQM effect are considered by varying parameter $m_0$ in non-chiral limit. For $π^0$ meson, $T_m^0$ is not a monotonic function of magnetic field with low $m_0$, but it is a monotonic decreasing function with larger $m_0$. In the weak magnetic field region, $T_m^0$ is higher for larger $m_0$, but in the strong magnetic field region, it is lower for larger $m_0$. For $π^+$ meson, $T^+_m$ is only quantitatively modifies by current quark mass effect, and it becomes higher with larger $m_0$.

hep-ph↗

arXiv2023

Assessing the potential of LLM-assisted annotation for corpus-based pragmatics and discourse analysis: The case of apology

Certain forms of linguistic annotation, like part of speech and semantic tagging, can be automated with high accuracy. However, manual annotation is still necessary for complex pragmatic and discursive features that lack a direct mapping to lexical forms. This manual process is time-consuming and error-prone, limiting the scalability of function-to-form approaches in corpus linguistics. To address this, our study explores the possibility of using large language models (LLMs) to automate pragma-discursive corpus annotation. We compare GPT-3.5 (the model behind the free-to-use version of ChatGPT), GPT-4 (the model underpinning the precise mode of Bing chatbot), and a human coder in annotating apology components in English based on the local grammar framework. We find that GPT-4 outperformed GPT-3.5, with accuracy approaching that of a human coder. These results suggest that LLMs can be successfully deployed to aid pragma-discursive corpus annotation, making the process more efficient, scalable and accessible.

cs.CL↗

arXiv2022

PSNet: Fast Data Structuring for Hierarchical Deep Learning on Point Cloud

In order to retain more feature information of local areas on a point cloud, local grouping and subsampling are the necessary data structuring steps in most hierarchical deep learning models. Due to the disorder nature of the points in a point cloud, the significant time cost may be consumed when grouping and subsampling the points, which consequently results in poor scalability. This paper proposes a fast data structuring method called PSNet (Point Structuring Net). PSNet transforms the spatial features of the points and matches them to the features of local areas in a point cloud. PSNet achieves grouping and sampling at the same time while the existing methods process sampling and grouping in two separate steps (such as using FPS plus kNN). PSNet performs feature transformation pointwise while the existing methods uses the spatial relationship among the points as the reference for grouping. Thanks to these features, PSNet has two important advantages: 1) the grouping and sampling results obtained by PSNet is stable and permutation invariant; and 2) PSNet can be easily parallelized. PSNet can replace the data structuring methods in the mainstream point cloud deep learning models in a plug-and-play manner. We have conducted extensive experiments. The results show that PSNet can improve the training and inference speed significantly while maintaining the model accuracy.

cs.CV↗

arXiv2021

Light mesons around critical end points in $T-μ_B-μ_I-eB$ space

Light mesons $(σ, π^0, π^\pm)$ are investigated in $T-μ_B-μ_I-eB$ space by using a two-flavor NJL model, which are related to the chiral symmetry restoration and pion superfluid phase transition. In $T-μ_B-eB$ space, during the chiral restoration process, the mass of pseudo-Goldstone mode $π^0$ keeps increasing, together with the sudden mass jump. At the critical end point region, $π^0$ meson has a very sharp but continuous mass increase, together with a sudden mass jump at the Mott transition, and in the first order chiral phase transition region nearby, we observe twice $π^0$ mass jumps, induced by the Mott transition and quark mass jump, respectively. The mass of Higgs mode $σ$ first decreases and then increases associated with the chiral symmetry restoration, and shows a jump at the first order chiral phase transition. We plot the chiral phase diagram in terms of the change of quark mass, the Mott transition of $π^0$ and the minimum mass of $σ$. Due to the explicit breaking of chiral symmetry in physical case, the chiral restoration phase boundaries in $T-μ_B$ plane from the order parameter and meson side are different from each other. In $T-μ_I$ plane, the competition between pion superfluid phase transition and chiral symmetry restoration under magnetic fields is studied in terms of the Goldstone mode $π^+$ and the pseudo-Goldstone mode $π^0$. The separation of the two phase boundaries is enhanced by the external magnetic field. Different from the twice mass jumps of $π^0$ in the first order chiral phase transition region, the $π^+$ meson displays several mass jumps in the chiral crossover region. At the critical end point, $π^+$ also shows very sharp but continuous mass changes, together with a mass jump at the Mott transition.

hep-ph↗

arXiv2016

Truth Discovery with Memory Network

Truth discovery is to resolve conflicts and find the truth from multiple-source statements. Conventional methods mostly research based on the mutual effect between the reliability of sources and the credibility of statements, however, pay no attention to the mutual effect among the credibility of statements about the same object. We propose memory network based models to incorporate these two ideas to do the truth discovery. We use feedforward memory network and feedback memory network to learn the representation of the credibility of statements which are about the same object. Specially, we adopt memory mechanism to learn source reliability and use it through truth prediction. During learning models, we use multiple types of data (categorical data and continuous data) by assigning different weights automatically in the loss function based on their own effect on truth discovery prediction. The experiment results show that the memory network based models much outperform the state-of-the-art method and other baseline methods.

cs.CL↗