Zeli Guan
PulseAugur coverage of Zeli Guan — every cluster mentioning Zeli Guan across labs, papers, and developer communities, ranked by signal.
Zeli Guan to publish research on graph-based representation learning for scientific literature
Multiple recent papers highlight Zeli Guan's work on graph neural networks for scientific literature representation and patent entity alignment. This suggests a focus on leveraging graph structures to understand complex data, with a specific application to scientific texts. Future work may involve publishing a dedicated paper on this topic.
Zeli Guan to explore adaptive clustering methods in federated learning
A recent paper by Zeli Guan introduces an adaptive OPTICS clustering method for federated learning, addressing challenges with non-IID data. This indicates a potential future direction for Zeli Guan's research to further develop or apply adaptive clustering techniques within federated learning frameworks, possibly for improved model aggregation or personalization.
Zeli Guan's recent work shows a strong focus on graph representation learning across diverse domains
Zeli Guan has recently published work utilizing graph neural networks for scientific literature representation, partitioning spatio-temporal data, and patent entity alignment. This demonstrates a consistent and strong interest in applying graph-based methods to solve problems in distinct areas, suggesting this will be a recurring theme in their research output.
-
New CWUTM model excels at finding scarce topics in short texts
Researchers have developed a new topic modeling approach called CWUTM, designed to effectively identify scarce topics within unbalanced short-text datasets. This method utilizes co-occurrence word networks to capture wo…
-
New CasAug Model Enhances Relation Extraction by Reducing Triple Overlap
Researchers have developed a new model called CasAug, which aims to improve relation extraction in natural language processing by addressing the issue of triple overlap. This model enhances the semantic understanding of…
-
New method fuses graph and text for patent entity alignment
This paper introduces a novel method for aligning entities within science and technology patent knowledge graphs. The proposed approach leverages a graph convolution network combined with the BERT model to fuse structur…
-
New method improves partitioning of public safety spatio-temporal data
Researchers have developed a new method called IFL-LSTP for partitioning large-scale public safety spatio-temporal data. This approach aims to improve storage, management, and application of such data by addressing limi…
-
New method uses graph neural networks for scientific literature representation
This paper introduces a novel method for learning semantic representations of scientific literature using adaptive features and graph neural networks. The approach considers scientific literature features both globally …
-
Paper analyzes domain information mining and theme evolution in scientific research
This paper explores methods for analyzing scientific papers, focusing on how to extract domain information and track the evolution of research topics. It discusses techniques for learning semantic features, mining field…
-
New adaptive OPTICS clustering method enhances federated learning
Researchers have developed a novel method for federated learning that addresses the challenge of non-independent and identically distributed data across user terminals. This approach utilizes an adaptive OPTICS clusteri…