Hui Li

Dr. Hui Li

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
1 January 2025
Verified
Conference Paper
Open access
PDF file
HLHui Li
Hui Li
AWAnte Wang
Ante Wang
KLKunquan Li
Kunquan Li
ZWZhihao Wang
Zhihao Wang
LZLiang Zhang
Liang Zhang
JSJinsong Su
Jinsong Su
Hui Li, Ante Wang, Kunquan Li, Zhihao Wang, Liang Zhang, Delai Qiu, Qingsong Liu, Jinsong Su. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

5 November 2024 in Mathematics
Verified
Journal Article
Open access
PDF file
HLHui Li
Hui Li
LJLanlan Jiang
Lanlan Jiang
JLJun Li
Jun Li
The proliferation of the Internet and mobile devices has made it increasingly easy to propagate rumors on social media. Widespread rumors can incite public panic and have detrimental effects on individuals. In recent years, researchers have found that both the spatial structure of rumor diffusion and the temporal features of propagation can be effective in identifying rumors. However, existing methods tend to focus on either spatial structure or temporal information in isolation, and few models can effectively capture both types of information. Additionally, most existing methods treat continuously changing temporal information as static snapshots, neglecting the precise timing of propagation. Moreover, as users repost and comment, background knowledge associated with the posts also evolves dynamically, which is often ignored. To address these limitations, we propose CGNKP (Continuous-time Dynamic Graph Networks integrated with Knowledge Propagation), a model that jointly captures the spatial structure and continuous-time features of post propagation to fully understand the dynamics of background knowledge. Specifically, we introduce a novel method for encoding continuous-time dynamic graphs, modeling the propagation process through two dynamic graphs: a temporal propagation graph (for posts diffusion) and a temporal knowledge graph (for knowledge diffusion). Extensive experiments on real-world datasets demonstrate that CGNKP significantly outperforms multiple strong baselines, achieving accuracies of 0.861 on the Twitter15 dataset and 0.903 on the Twitter16 dataset.

27 May 2025 in Mathematics
Verified
Journal Article
Open access
PDF file
LJLanlan Jiang
Lanlan Jiang
CZCheng Zhang
Cheng Zhang
XQXingguo Qin
Xingguo Qin
YZYa Zhou
Ya Zhou
GHGuanglun Huang
Guanglun Huang
JLJun Li
Jun Li
In the realm of natural language processing (NLP), text classification constitutes a task of paramount significance for large language models (LLMs). Nevertheless, extant methodologies predominantly depend on the output generated by the final layer of LLMs, thereby neglecting the wealth of information encapsulated within neurons residing in intermediate layers. To surmount this shortcoming, we introduce LENS (Linear Exploration and Neuron Selection), an innovative technique designed to identify and sparsely integrate salient neurons from intermediate layers via a process of linear exploration. Subsequently, these neurons are transmitted to downstream modules dedicated to text classification. This strategy effectively mitigates noise originating from non-pertinent neurons, thereby enhancing both the accuracy and computational efficiency of the model. The detection of telecommunication fraud text represents a formidable challenge within NLP, primarily attributed to its increasingly covert nature and the inherent limitations of current detection algorithms. In an effort to tackle the challenges of data scarcity and suboptimal classification accuracy, we have developed the LENS-RMHR (Linear Exploration and Neuron Selection with RoBERTa, Multi-head Mechanism, and Residual Connections) model, which extends the LENS framework. By incorporating RoBERTa, a multi-head attention mechanism, and residual connections, the LENS-RMHR model augments the feature representation capabilities and improves training efficiency. Utilizing the CCL2023 telecommunications fraud dataset as a foundation, we have constructed an expanded dataset encompassing eight distinct categories that encapsulate a diverse array of fraud types. Furthermore, a dual-loss function has been employed to bolster the model’s performance in multi-class classification scenarios. Experimental results reveal that LENS-RMHR demonstrates superior performance across multiple benchmark datasets, underscoring its extensive potential for application in the domains of text classification and telecommunications fraud detection.

30 July 2025 in IEEE Transactions on Signal and Information Processing over Networks
Verified
Journal Article
Closed access
HLHui Li
Hui Li
LWLai Wei
Lai Wei
KLKunquan Li
Kunquan Li
GHGuimin Huang
Guimin Huang
JLJun Li
Jun Li
In today’s highly informalized society, the speed and scope of rumor dissemination pose a great threat to social stability and personal interests. Detecting rumors manually requires a lot of human effort. Therefore, automatic rumor detection has received significant attention. Recently, some researchers have focused on using propagation structural features to identify rumors. However, existing propagation structure-based methods either utilize only spatial features or only temporal features of propagation. Few models can effectively leverage both types of propagation structural features. This paper proposes a Source-Guided Temporal-Spatial joint rumor detection model (SGTS). SGTS dynamically divides the propagation process of an event into a series of temporal sub-events. Additionally, SGTS employs an information-level connection strategy that incorporates spatial structural features from previous temporal stages into the encoding of subsequent stages. In this way, SGTS can effectively capture the spatiotemporal features of propagation. Experimental results and in-depth analysis on commonly-used datasets demonstrate that SGTS achieves significant improvements over existing methods.

1 January 2026 in Computer Modeling in Engineering & Sciences
Verified
Journal Article
Open access
YFYiyang Fu
Yiyang Fu
HLHui Li
Hui Li
WWWangyu Wu
Wangyu Wu
Weakly Supervised Semantic Segmentation (WSSS), which relies only on image-level labels, has attracted significant attention for its cost-effectiveness and scalability. Existing methods mainly enhance inter-class distinctions and employ data augmentation to mitigate semantic ambiguity and reduce spurious activations. However, they often neglect the complex contextual dependencies among image patches, resulting in incomplete local representations and limited segmentation accuracy. To address these issues, we propose the Context Patch Fusion with Class Token Enhancement (CPF-CTE) framework, which exploits contextual relations among patches to enrich feature representations and improve segmentation. At its core, the Contextual-Fusion Bidirectional Long Short-Term Memory (CF-BiLSTM) module captures spatial dependencies between patches and enables bidirectional information flow, yielding a more comprehensive understanding of spatial correlations. This strengthens feature learning and segmentation robustness. Moreover, we introduce learnable class tokens that dynamically encode and refine class-specific semantics, enhancing discriminative capability. By effectively integrating spatial and semantic cues, CPF-CTE produces richer and more accurate representations of image content. Extensive experiments on PASCAL VOC 2012 and MS COCO 2014 validate that CPF-CTE consistently surpasses prior WSSS methods.

13 November 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
NDNing Ding
Ning Ding
SZSen Zhang
Sen Zhang
LWLai Wei
Lai Wei
HLHui Li
Hui Li
This article proposes the MEMS thermoelectric-photoelectric integrated generator with I-shaped thermocouple and its power management circuit. The thermoelectric-photoelectric integrated generator is composed of thermoelectric power generator and photoelectric power generator. We design the corresponding fabrication process to make the integrated generator. The generator can gather both heating and lighting energy from the surrounding area. The thermoelectric area and photoelectric area are connected in the internal of the generator. The output of integrated generator is connected with the power management circuit to increase the output voltage. Power management circuit can increase the output voltage of the integrated power generator to 3.1 V. The innovation is that the I-shaped thermocouple is used to gather thermoelectric energy. Besides, 3.1V output voltage can meet the most needs of the transceiver system. The integrated generator with the power management circuit can supply energy to the wireless sensor node.