Deguang Peng
Papers
1
Total Citations
5
H-Index
1
About
Deguang Peng is a researcher whose work lies at the intersection of natural language processing and knowledge-enhanced machine learning. His primary research focus is on semantic understanding, particularly the challenge of measuring similarity between long-form texts. In his most-cited work, "Learning Long-text Semantic Similarity with Multi-Granularity Semantic Embedding Based on Knowledge Enhancement" (2020), Peng introduced a novel framework that integrates external knowledge bases with multi-granularity embedding techniques. This approach allows models to capture both fine-grained word-level and coarse-grained sentence-level semantics, significantly improving the accuracy of long-text similarity tasks. While his citation count is still growing—with this paper garnering 5 citations—the work represents a meaningful step forward in addressing the limitations of traditional embedding methods for lengthy documents. Peng’s research is particularly relevant for applications in document retrieval, plagiarism detection, and question-answering systems. By bridging knowledge enhancement with multi-granularity analysis, he has contributed a practical methodology that continues to inform subsequent studies in semantic representation. His ongoing work promises to further refine how machines understand complex, extended texts.
Research Focus
Key Achievements
Top Papers
- 1