Edge-centric optimization: a novel strategy for minimizing information loss in graph-to-text generation
Abstract Given the remarkable text generation capabilities of pre-trained language models, impressive results have been realized in graph-to-text generation. However, while learning from knowledge graphs, these language models are unable to fully grasp the structural information of the graph, leadin...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2024-12-01
|
| Series: | Complex & Intelligent Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s40747-024-01690-y |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832571144523218944 |
|---|---|
| author | Zheng Yao Jingyuan Li Jianhe Cen Shiqi Sun Dahu Yin Yuanzhuo Wang |
| author_facet | Zheng Yao Jingyuan Li Jianhe Cen Shiqi Sun Dahu Yin Yuanzhuo Wang |
| author_sort | Zheng Yao |
| collection | DOAJ |
| description | Abstract Given the remarkable text generation capabilities of pre-trained language models, impressive results have been realized in graph-to-text generation. However, while learning from knowledge graphs, these language models are unable to fully grasp the structural information of the graph, leading to logical errors and missing key information. Therefore, an important research direction is to minimize the loss of graph structural information during the model training process. We propose a framework named Edge-Optimized Multi-Level Information refinement (EMLR), which aims to maximize the retention of the graph’s structural information from an edge perspective. Based on this framework, we further propose a new graph generation model, named TriELMR, highlighting the comprehensive interactive learning relationship between the model and the graph structure, as well as the importance of edges in the graph structure. TriELMR adopts three main strategies to reduce information loss during learning: (1) Knowledge Sequence Optimization; (2) EMLR Framework; and (3) Graph Activation Function. Experimental results reveal that TriELMR exhibits exceptional performance across various benchmark tests, especially on the webnlgv2.0 and Event Narrative datasets, achieving BLEU-4 scores of $$66.5\%$$ 66.5 % and $$37.27\%$$ 37.27 % , respectively, surpassing the state-of-the-art models. These demonstrate the advantages of TriELMR in maintaining the accuracy of graph structural information. Graphical abstract |
| format | Article |
| id | doaj-art-0fcb5540ae4b4ed19eba003ecf411060 |
| institution | Kabale University |
| issn | 2199-4536 2198-6053 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Springer |
| record_format | Article |
| series | Complex & Intelligent Systems |
| spelling | doaj-art-0fcb5540ae4b4ed19eba003ecf4110602025-02-02T12:48:56ZengSpringerComplex & Intelligent Systems2199-45362198-60532024-12-0111111910.1007/s40747-024-01690-yEdge-centric optimization: a novel strategy for minimizing information loss in graph-to-text generationZheng Yao0Jingyuan Li1Jianhe Cen2Shiqi Sun3Dahu Yin4Yuanzhuo Wang5Henan Institute of Advanced Technology, Zhengzhou UniversitySchool of Computer and Artificial Intelligence, Beijing Technology and Business UniversityHenan Institute of Advanced Technology, Zhengzhou UniversityHenan Institute of Advanced Technology, Zhengzhou UniversityHenan Institute of Advanced Technology, Zhengzhou UniversityInstitute of Computing Technology, Chinese Academy of SciencesAbstract Given the remarkable text generation capabilities of pre-trained language models, impressive results have been realized in graph-to-text generation. However, while learning from knowledge graphs, these language models are unable to fully grasp the structural information of the graph, leading to logical errors and missing key information. Therefore, an important research direction is to minimize the loss of graph structural information during the model training process. We propose a framework named Edge-Optimized Multi-Level Information refinement (EMLR), which aims to maximize the retention of the graph’s structural information from an edge perspective. Based on this framework, we further propose a new graph generation model, named TriELMR, highlighting the comprehensive interactive learning relationship between the model and the graph structure, as well as the importance of edges in the graph structure. TriELMR adopts three main strategies to reduce information loss during learning: (1) Knowledge Sequence Optimization; (2) EMLR Framework; and (3) Graph Activation Function. Experimental results reveal that TriELMR exhibits exceptional performance across various benchmark tests, especially on the webnlgv2.0 and Event Narrative datasets, achieving BLEU-4 scores of $$66.5\%$$ 66.5 % and $$37.27\%$$ 37.27 % , respectively, surpassing the state-of-the-art models. These demonstrate the advantages of TriELMR in maintaining the accuracy of graph structural information. Graphical abstracthttps://doi.org/10.1007/s40747-024-01690-yGraph-generated textKnowledge representationKnowledge graphsGraph theoryEdge-aware attention |
| spellingShingle | Zheng Yao Jingyuan Li Jianhe Cen Shiqi Sun Dahu Yin Yuanzhuo Wang Edge-centric optimization: a novel strategy for minimizing information loss in graph-to-text generation Complex & Intelligent Systems Graph-generated text Knowledge representation Knowledge graphs Graph theory Edge-aware attention |
| title | Edge-centric optimization: a novel strategy for minimizing information loss in graph-to-text generation |
| title_full | Edge-centric optimization: a novel strategy for minimizing information loss in graph-to-text generation |
| title_fullStr | Edge-centric optimization: a novel strategy for minimizing information loss in graph-to-text generation |
| title_full_unstemmed | Edge-centric optimization: a novel strategy for minimizing information loss in graph-to-text generation |
| title_short | Edge-centric optimization: a novel strategy for minimizing information loss in graph-to-text generation |
| title_sort | edge centric optimization a novel strategy for minimizing information loss in graph to text generation |
| topic | Graph-generated text Knowledge representation Knowledge graphs Graph theory Edge-aware attention |
| url | https://doi.org/10.1007/s40747-024-01690-y |
| work_keys_str_mv | AT zhengyao edgecentricoptimizationanovelstrategyforminimizinginformationlossingraphtotextgeneration AT jingyuanli edgecentricoptimizationanovelstrategyforminimizinginformationlossingraphtotextgeneration AT jianhecen edgecentricoptimizationanovelstrategyforminimizinginformationlossingraphtotextgeneration AT shiqisun edgecentricoptimizationanovelstrategyforminimizinginformationlossingraphtotextgeneration AT dahuyin edgecentricoptimizationanovelstrategyforminimizinginformationlossingraphtotextgeneration AT yuanzhuowang edgecentricoptimizationanovelstrategyforminimizinginformationlossingraphtotextgeneration |