科研成果

2025
Chen, AX; Xiang ZJSLGFMJJ. Unpacking help-seeking process through multimodal learning analytics: A comparative study of ChatGPT vs Human expert. Computers & Education [Internet]. 2025;(226). 访问链接
Yan W, Zhang X, Wang Y, Peng K, Ma Y. Unraveling the relationship between teachers’ and students’ mental health: A one-to-one matched analysis. The Journal of Experimental Education [Internet]. 2025;93(1):136-148. 访问链接Abstract
This study aims to identify the associations between teacher mental health and student mental health. Cross-sectional data were collected from 127,877 students aged 9–20 years and 2,759 teachers across 31 provinces in China. The mental health of students and teachers were assessed by well-being (life satisfaction and positive mental health), and psychological distress (depression and anxiety). Controlling for demographic variables, multilevel regression analyses suggest that higher teacher positive mental health was linked to higher student positive mental health and lower student depression; higher teacher depression were correlated with higher student depression; and teacher life satisfaction and anxiety were not correlated with any indicators of student mental health. The study highlights the significant association between teacher mental health and student mental health.
Dang Q, Li G. Unveiling trust in AI: the interplay of antecedents, consequences, and cultural dynamics. AI & SOCIETY [Internet]. 2025:1-24. 访问链接Abstract
Trust in artificial intelligence (AI) has become a central issue due to the opacity and unpredictability of AI decision-making processes. However, existing studies often produce inconsistent results and fail to provide a unified understanding of the underlying factors, making a comprehensive review necessary. To address this gap, we conducted a systematic review of 562 empirical studies to explore the antecedents and consequences of human trust in AI. The review identified key antecedents of trust, including AI capability, anthropomorphism, individual factors, and explainability, and associated consequences, such as behavioral intention, attitude, and acceptance. A cross-cultural analysis revealed significant differences in how cultural contexts influence the perception and prioritization of factors, such as capability, transparency, and anthropomorphism. These findings emphasize the need for a multidimensional approach to developing trustworthy AI systems, highlighting the importance of cultural sensitivity in AI design. The review also suggests several promising avenues for future research, including dynamic trust formation, reciprocal trust relationships, and the evolution of trust over time. Addressing these areas will enhance our understanding of trust in AI and contribute to the development of culturally adapted and ethically sound AI technologies.
Li P, Zhu R, McJeon H, Byers E, Zhou P, Ou Y. Using deep learning to generate key variables in global mitigation scenarios. Nature Climate Change [Internet]. 2025;15:760–768. [Link]
Duan R, Yan J, Meng D, Wang Y, Xu D, Zhang M, YU D, Tang K, Dong J, Han D, et al. A van der Waals ferroelectric switchable diode with ultra-high nonlinearity factor. Applied Surface Science [Internet]. 2025;693:162792. 访问链接Abstract
Two-dimensional (2D) van der Waals ferroelectric materials have emerged as promising candidates for miniaturized devices due to their atomically thin structures and unique ability to maintain ferroelectricity even at reduced dimensions. Recent research indicates that the interfacial barriers between semiconductors and ferroelectrics can be modulated by polarization charges, with ferroelectric polarization—reversible by an external electric field—playing a crucial role in the switchable diode effect. In this work, we investigate a room-temperature switchable ferroelectric diode (Fe-diode) based on a MoS2/α-In2Se3 heterojunction. The out-of-plane ferroelectric properties of the α-In2Se3 layer enable efficient modulation of the Schottky barriers at the MoS2/α-In2Se3 interface through external voltage application, thereby achieving a notable switchable diode effect with a nonlinearity of up to 934. By exploiting the inherent nonlinearity, the ferroelectric diode can effectively generate complex signal waveforms, making it highly suitable for secure communication systems. These findings make the ferroelectric diode a potential candidate for enhancing confidentiality in future communication technologies, protecting data against eavesdropping and unauthorized access.
Wu Z, SONG J, TANG X*, XU B, LUO H, YANG Y, WANG R, Xiaochen, WANG Y*. A Variation-Tolerant Continuous-Time IsingMachine with eDRAM-based Spin Interaction andLeaked Negative Feedback Annealing. IEEE Journal of Solid-State Circuit [Internet]. 2025;60(5):1793-1804. 访问链接
Wang H, Yuan B, Zhang X, Wang J, Chen X, Wang Y, Qin Y, Li X, Zhang C, Liu A. Vertical Gradient of Nitryl Chloride and Implications for Atmospheric Photochemistry in Pearl River Delta, China, during Wintertime. 2025.
Bai Y, Chen T-K, Liu J, Ma X. Wess-Zumino-Witten Interactions of Axions. Phys. Rev. Lett. 2025;134:081803.
Bai Y, Chen T-K, Liu J, Ma X. Wess-Zumino-Witten Interactions of Axions: Three-Flavor. 2025.
Shen Y, Author) YZ (C. Yuting Shen; Yanzhen Zhu; Rui Yang ; Towards Cultural Self-awareness in Research: A Study of Chinese Scholars in the Humanities and Social Sciences. Higher Education [Internet]. 2025. 访问链接Abstract
The global knowledge asymmetries are increasingly interrogated by non-Western humanities and social sciences (HSS) scholars whose research is anchored in local contexts yet must adhere to international (Western) standards. Under this circumstance, the study aims to examine how the cultural self-awareness of non-Western HSS scholars is manifested in research through a Chinese lens. Based on previous theoretical perspectives and Fei Xiaotong’s theory of cultural self-awareness, the study first constructs two analytical dimensions: academic self-reflexivity and cultural appreciation attitudes. It then performs a qualitative investigation including 28 Mainland Chinese HSS scholars through interviews and literature analyses. The findings highlight key principles for academic self-reflexivity, namely reflecting on intellectual extraversion, dichotomous thinking, and the reemphasis on Chinese culture and knowledge. The cultural appreciation attitudes are also elaborated, which are embodied in the recognition and revaluation of traditional Chinese knowledge, the continued appreciation of modern Western knowledge, and the synthesis of different cultures and knowledge in research. These findings develop Fei’s cultural self-awareness theory and add new discourses to address global knowledge imbalances, promoting a more diverse and inclusive global higher education landscape.
Huang Y, Liao X, Liang J, Quan Y, Shi B, Xu Y. Zero-Shot Low-Light Image Enhancement via Latent Diffusion Models, in Proceedings of the AAAI Conference on Artificial Intelligence (AAAI).; 2025.Abstract
Low-light image enhancement (LLIE) aims to improve visibility and signal-to-noise ratio in images captured under poor lighting conditions. Despite impressive improvement, deep learning-based LLIE approaches require extensive training data, which is often difficult and costly to obtain. In this paper, we propose a zero-shot LLIE framework leveraging pre-trained latent diffusion models for the first time, which act as powerful priors to recover latent images from low-light inputs. Our approach introduces several components to alleviate the inherent challenges in utilizing pre-trained latent diffusion models, modeling the degradation process in an image-adaptive manner, penalizing the latent outside the manifold of natural images, and balancing the strengths of the guidance from the given low-light image during the denoising process. Experimental results demonstrate that our framework outperforms existing methods, achieving superior performance across various datasets.
曲天书, 吴玺宏, 吴东航.; 2025. 一种基于交叉注意力-状态空间模型的目标声源提取方法. China patent CN 202510290916.3.
曲天书, 吴玺宏, 吴东航, 杜佳琪.; 2025. 一种基于包络估计的未知声源数量移动声源定位跟踪方法. China patent CN 202510538463.1.
曲天书, 吴玺宏, 游宇寰.; 2025. 一种基于文本辅助的视频到音频生成方法. China patent CN 202510298019.7.
唐星龙, 徐扬*. 三元空间视域下事实库数据可信交付框架探究. 情报理论与实践 [Internet]. 2025:1-11. 访问链接
谢侃侃, 粟琳. 中国与马来西亚关系50年:双边共识、多边互动与系统性制约. 南洋问题研究. 2025;总第201期(2025年3月):1-18.Abstract
自 1974 年 5 月 31 日中国与马来西亚建立正式外交关系以来,两国关系历经多个发展阶段,不断深化发展。在 50 年的演进历程中,中马关系经历几个不同历史时期,两国通过互动达成共识,构建彼此认可的身份与利益认知,进而推动双边关系的稳定发展。然而,仅从双边互动视角出发,难以全面解释复杂国际环境下中马关系的发展历程。因此,从系统层面探讨国际体系在不同时期对中马关系的制约,以及相关赋能因素的影响,是十分必要和重要的。尽管国际体系的结构性限制会影响国家行为体的互动,但行为体的能动性仍能部分突破这些限制,自主选择发展路径并塑造战略取向。这一系统视角不仅深化对中马关系本质的理解,也为探讨双边关系在不断变动的国际环境中的韧性提供一种新思路。
杨锋. 人工智能对社会科学研究影响深远. 《科技日报》 [Internet]. 2025:6月23日. 访问链接
杨锋. 国际关系测量研究中的人工智能方法. 《世界经济与政治》. 2025;(1):12-15.
王雯, 李丰. 基于分段组合VARX模型的中国出境游客数量预测. 经济管理学刊. 2025;4:255–284.Abstract
本文对结构性变化的旅游需求进行研究,基于带有外生变量的向量自回归(VARX)模型,提出了一种分段组合预测的方法。与既有研究普遍采用的基于完整数据集构建组合预测模型不同,本文创新性地将时间因素纳入组合预测考量,通过将不同时间段的变量视为独立的单元,构建出分段时间序列数据集的组合预测模型。该方法以游客的网络搜索行为作为外生变量用于预测旅游人数,并捕捉这些外生变量在不同时间节点上对旅游人数产生的差异化影响,特别是在新冠疫情等突发冲击下的动态变化。实证结果显示,VARX模型的分段组合在预测中国出境旅游人数时展现出更高的准确性,其预测精度因考虑了外生变量在不同时间段的特异性影响而得以提升。事后分析进一步显示,特别是针对2024年中国出境旅游趋势的外样本预测结果,随着新冠疫情影响的逐渐消退及全球旅游市场的逐步复苏,中国出境旅游人数将呈现积极向上的增长态势。这一结论与现有公开文献中的趋势分析相吻合,进一步印证了本文预测方法的实践应用价值。
董佶圣, 李聪, 沈子楠, 胡达齐, 沈晴霓, 吴中海. 基于国密SM9 的区块链匿名交易方案. 中国科学:信息科学. 2025;55(6):1428 - 1446.

Pages