Chunjiang Liu

Dr. Chunjiang Liu

orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

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

Representative Publications

View All Publications
24 January 2024 in Mathematics
Verified
Journal Article
Open access
PDF file
CLChunjiang Liu
Chunjiang Liu
YHYikun Han
Yikun Han
HXHaiyun Xu
Haiyun Xu
SYShihan Yang
Shihan Yang
KWKaidi Wang
Kaidi Wang
YSYongye Su
Yongye Su
This study presents a novel approach that synergizes community detection algorithms with various Graph Neural Network (GNN) models to bolster link prediction in scientific literature networks. By integrating the Louvain community detection algorithm into our GNN frameworks, we consistently enhanced the performance across all models tested. For example, integrating the Louvain model with the GAT model resulted in an AUC score increase from 0.777 to 0.823, exemplifying the typical improvements observed. Similar gains were noted when the Louvain model was paired with other GNN architectures, confirming the robustness and effectiveness of incorporating community-level insights. This consistent increase in performance—reflected in our extensive experimentation on bipartite graphs of scientific collaborations and citations—highlights the synergistic potential of combining community detection with GNNs to overcome common link prediction challenges such as scalability and resolution limits. Our findings advocate for the integration of community structures as a significant step forward in the predictive accuracy of network science models, offering a comprehensive understanding of scientific collaboration patterns through the lens of advanced machine learning techniques.

18 October 2023 in arXiv
Verified
Preprint
Open access
PDF file
YHYikun Han
Yikun Han
CLChunjiang Liu
Chunjiang Liu
Pengfei WangPengfei Wang
Pengfei Wang
A vector database is used to store high-dimensional data that cannot be characterized by traditional DBMS. Although there are not many articles describing existing or introducing new vector database architectures, the approximate nearest neighbor search problem behind vector databases has been studied for a long time, and considerable related algorithmic articles can be found in the literature. This article attempts to comprehensively review relevant algorithms to provide a general understanding of this booming research area. The basis of our framework categorises these studies by the approach of solving ANNS problem, respectively hash-based, tree-based, graph-based and quantization-based approaches. Then we present an overview of existing challenges for vector databases. Lastly, we sketch how vector databases can be combined with large language models and provide new possibilities.

11 December 2025 in Systems
Verified
Journal Article
Open access
PDF file
XLXin Li
Xin Li
HXHaiyun Xu
Haiyun Xu
Robin HaunschildRobin Haunschild
Robin Haunschild
ZTZehua Tong
Zehua Tong
CLChunjiang Liu
Chunjiang Liu
This study integrates complex systems theory and innovation ecosystem theory to develop a unified framework encompassing the innovation environment, innovation actors, and innovation networks. Using fsQCA and NCA, it examines the impact of cross-layer interactions and the coupling of multiple factors on collaborative performance. Empirical analysis in the field of natural language processing (NLP) demonstrates that no single factor is sufficient to serve as a necessary condition for achieving high-innovation collaboration performance. Innovation actors, as endogenous evolutionary drivers, play a central and catalytic role in the collaboration process. Moreover, under specific conditions, the relationship between the innovation environment and innovation networks exhibits a substitutive effect, with certain capabilities enabling this dynamic. This study extends the theoretical understanding of collaboration pathways within innovation ecosystems and offers practical recommendations for fostering innovation cooperation across different industries and organizations. It achieves this by constructing a “situational type–configuration path” matrix, decision tree, and innovation collaboration performance realization model.

29 July 2025 in Neural Processing Letters
Verified
Journal Article
Open access
PDF file
SYShihan Yang
Shihan Yang
LLLiannan Lin
Liannan Lin
CLChunjiang Liu
Chunjiang Liu
MZMingkai Zhang
Mingkai Zhang
HYHaiyi Yang
Haiyi Yang
Relation prediction in knowledge graphs is a classic task. However, predicting unseen facts under the open-world assumption remains challenging, as conventional closure reasoning methods tend to fail. To address this, we propose an uncertain knowledge graph embedding and Bayesian inferring model, UKGEBN, which examines the capabilities of approximate inference under the open-world assumption. This model employs a large language model to encode entities and relations by treating facts as natural language sentences. It then learns knowledge confidence through a recurrent neural network and subsequently constructs Bayesian networks for the uncertain knowledge graph based on the learned confidence. The constructed Bayesian network can reasonably infer unseen knowledge facts, even if the elements of these facts-such as new entities and relationships-have never been encountered in the knowledge base before. The experimental results show that the new model has improved the MSE metric by 24.6% in relation prediction on the benchmark dataset. Case studies indicate that the model significantly enhances the revelation of latent semantic information, facilitating the prediction of implicit relationships in uncertain knowledge graphs.

29 January 2026 in Systems
Verified
Journal Article
Open access
PDF file
XHXiao yang Hu
Xiao yang Hu
HXHaiyun Xu
Haiyun Xu
Robin HaunschildRobin Haunschild
Robin Haunschild
CLChunjiang Liu
Chunjiang Liu
XTXiao Tan
Xiao Tan
Grounded in knowledge recombination theory and innovation tension theory, this study develops a novel measurement framework for scientific and technological innovation (STI) risks that captures the dynamic and systemic equilibrium between novelty and adaptation. We first analyze the endogenous mechanisms through which STI risks emerge from knowledge recombination processes, and then propose a classification framework for knowledge recombination, along with quantifiable metrics for novelty and adaptation. Next, we introduce a risk classification system for STI and corresponding quantitative evaluation metrics, facilitating dynamic monitoring of innovation risk states. Finally, we validate the framework through an empirical case study in natural language processing (NLP). Our results reveal a persistent innovation tension within the STI system between novelty and adaptation. Emerging phrases and reinforced phrases demonstrate distinct risk profiles and distribution patterns, corresponding to differentiated structural and evolutionary regimes. These differences stem from their distinct mechanisms in the novelty–adaptation interaction within a complex innovation system. Specifically, in emerging phrases, novelty shows a stable positive linear correlation with Z-score, while adaptation exhibits a significant negative linear correlation with Z-score. In reinforced phrases, novelty displays a significant bimodal association with Z-score, and adaptation demonstrates a robust inverted U-shaped relationship with Z-score. Emerging knowledge combinations show significantly higher risk scores than reinforced combinations, with high-novelty–low-adaptation combinations consistently in the highest risk quantile across stages. Moreover, the risk threshold for emerging phrases increases monotonically across developmental phases. Thus, our framework advances innovation risk assessment from static categorization to dynamic, system-level evaluation, enabling tiered risk management and optimized resource allocation for high-potential innovation pathways.

23 May 2025
Verified
Book Chapter
Closed access
PDF file
LCLingchen Cai
Lingchen Cai
CLChunjiang Liu
Chunjiang Liu
SYShirui Yu
Shirui Yu
YCYiran Cai
Yiran Cai
HXHaiyun Xu
Haiyun Xu
KWKaidi Wang
Kaidi Wang