Xinyu Li

Dr. Xinyu Li

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 December 2025 in Sustainability
Verified
Journal Article
Open access
PDF file
HZHuiling Zhao
Huiling Zhao
YXYujie Xiang
Yujie Xiang
FGFeng Gong
Feng Gong
TXTianxiang Xu
Tianxiang Xu
YCYinghao Chen
Yinghao Chen
XLXinyu Li
Xinyu Li
Enhancing corporate green innovation has become a critical question in the context of sustainable development. Prior studies have predominantly examined the macro-level effects of regional integration while largely overlooking its micro-level impacts on enterprises. This study aims to examine the institutional effect of regional integration on corporate green innovation. Taking the Yangtze River Delta integration as a quasi-natural experiment, we utilize panel data from A-share listed companies between 2003 and 2022 and apply a multi-period difference-in-differences method. The empirical results reveal that regional integration significantly enhances corporate green innovation, with a more pronounced effect for non-state-owned firms, large firms, and those located in non-corridor cities. Mechanism analyses further reveal that regional integration promotes corporate green innovation by alleviating financing constraints and attracting foreign direct investment. By identifying regional integration as a critical driver of corporate green innovation, this study broadens the research perspective on corporate green innovation and provides policy implications for promoting sustainability through coordinated regional development strategies.

5 September 2025 in Biomedicines
Verified
Journal Article
Open access
PDF file
LWLanlan Wang
Lanlan Wang
XLXinyu Li
Xinyu Li
JYJialu Yuan
Jialu Yuan
YCYinghao Chen
Yinghao Chen
Background: Accurate and early diagnosis of autism spectrum disorder (ASD) is essential for timely intervention. Structural magnetic resonance imaging (sMRI) and functional magnetic resonance imaging (fMRI) provide complementary insights into brain structure and function. Most deep learning approaches rely on a single modality, limiting their ability to capture cross-modal relationships. Methods: We propose DAGMNet, a dual-branch attention-pruned graph neural network for ASD prediction that integrates sMRI, fMRI, and phenotypic data. The framework employs modality-specific feature extraction to preserve unique structural and functional characteristics, an attention-based cross-modal fusion module to model inter-modality complementarity, and a phenotype-pruned dynamic graph learning module with adaptive graph construction for personalized diagnosis. Results: Evaluated on the ABIDE-I dataset, DAGMNet achieves an accuracy of 91.59% and an AUC of 96.80%, outperforming several state-of-the-art baselines. To validate the method’s generalizability, we also validate it on ADNI datasets from other degenerative diseases and achieve good results. Conclusions: By effectively fusing multimodal neuroimaging and phenotypic information, DAGMNet enhances cross-modal representation learning and improves diagnostic accuracy. To further assist clinical decision making, we conduct biomarker detection analysis to provide region-level explanations of our model’s decisions.

16 February 2026 in Frontiers in Psychology
Verified
Journal Article
Open access
PDF file
XWXin Wu
Xin Wu
YZYu Zeng
Yu Zeng
ZLZhirong Li
Zhirong Li
XLXinyu Li
Xinyu Li
SCSiyu Chen
Siyu Chen
YXYoungcheng Xie
Youngcheng Xie
Introduction: Understanding how students engage with AI-driven feedback remains understudied in educational psychology. With ChatGPT’s emergence as a generative artificial intelligence tool, automated design improvement feedback (ADIF) has expanded significantly. This exploratory study investigates differential engagement patterns with ChatGPT-based ADIF across performance levels, grounded in self-regulated learning theory. Methods: A mixed-method multiple case study examined 50 design students (25 high performers, 25 low performers) during a product design session. Data included behavioral observations of prompt strategies and query patterns, lag sequential analysis of cognitive transitions, and semi-structured interviews on emotional engagement. Results: High performers employed diverse prompt strategies with iterative refinement, exhibited cyclical metacognitive transitions, and characterized interactions as exploratory and collaborative. Low performers used basic prompts with limited iterations, demonstrated linear query-to-implementation progressions, and described structured guidance-seeking interactions. Discussion: The findings extend self-regulated learning theory to human-AI contexts, revealing how metacognitive capabilities shape behavioral, cognitive, and emotional engagement with AI feedback. Results demonstrate the need for scaffolding interventions to support lower-performing students in developing metacognitive strategies for effective AI interaction. This study contributes initial insights into performance-based variations in human-AI collaboration within educational contexts.

10 October 2025 in IEEE Journal of Biomedical and Health Informatics
Verified
Journal Article
Closed access
XLXinyu Li
Xinyu Li
HKHulin Kuang
Hulin Kuang
JLJin Liu
Jin Liu
LWLanlan Wang
Lanlan Wang
PSPengcheng Shu
Pengcheng Shu
JWJianxin Wang
Jianxin Wang
Survival risk prediction is crucial for personalized treatment of gliomas. Medical image foundational models can explore complex medical features, which are critical for prognosis in gliomas. We propose SAM-Risk, which uses a customized SAM-Med3D with multi-view representation fusion and clinical knowledge-based age-grade stratified loss for glioma survival risk prediction. First, to utilize potential interactions between multiple views at an early stage, we design a 3D representation generation module that transforms 1D handcrafted radiomics and clinical features into 3D representations, which are fused with multimodal MRIs through a multi-view representation fusion module. The fused representation is fed into the customized SAM-Med3D, fine-tuned using LoRA and a disparity function to extract survival risk-related features. We design a feature refinement module to explore the inter-channel relationships among the outputs of the fine-tuned SAM-Med3D. Additionally, we propose an age-grade stratified loss based on glioma prognosis standards to make the predicted risk more consistent with clinical prior knowledge. Validated on two publicly available UCSF-PDGM and BraTS2020 datasets, SAM-Risk achieves a C-index of 75.08% and 73.67%, respectively, outperforming several survival risk prediction methods.

28 October 2026 in Sage Open
Verified
Journal Article
Open access
YZYu Zeng
Yu Zeng
YWYuhui Wang
Yuhui Wang
ZWZhe Wang
Zhe Wang
NMNan Ma
Nan Ma
XLXinyu Li
Xinyu Li
ZLZhirong Li
Zhirong Li
Despite growing interest in the potential of generative AI art tools, systematic empirical evidence remains limited regarding how non-design university students evaluate such tools after use and the psychological factors associated with these evaluations. Drawing on the Technology Acceptance Model (TAM), this study developed an extended model integrating motivational, affective, and cognitive factors and analyzed data from 317 non-design university students in China using covariance-based structural equation modeling (CB-SEM). Participants received standardized operational guidance and completed a brief image-creation task using Midjourney under a standardized study procedure. Analysis of the full sample showed that 11 of the 15 proposed hypotheses were supported (73.3%). Hedonic motivation, perceived output quality, perceived usefulness, and perceived ease of use were significantly and positively associated with post-use evaluation, whereas AI anxiety was significantly and negatively associated with post-use evaluation. Self-efficacy was indirectly associated with post-use evaluation primarily through perceived ease of use, while the structural path from perceived ease of use to perceived usefulness was not significant. These findings provide empirical support for extending TAM to the context of generative AI-assisted art learning and offer insights into how non-design students evaluate such technologies and how cross-disciplinary participation in art learning may be supported in higher education.

17 August 2026 in International Journal of Human–Computer Interaction
Verified
Journal Article
Closed access
YZYu Zeng
Yu Zeng
YWYuhui Wang
Yuhui Wang
ZGZhixuan Gong
Zhixuan Gong
XLXinyu Li
Xinyu Li
YWYumeng Wang
Yumeng Wang
ZLZhirong Li
Zhirong Li