A Method of Semantic Redundant Information Filtering for Abstract Meaning Representation Graph

Aiming at the problem of semantic redundancy in the existing text summarization, this paper proposes a method of semantic redundant information filtering based on Abstract Meaning Representation Graph. First, it transforms the original text into the corresponding AMR graph, then filters the semantic similar redundant information based on the WordNet semantic dictionary, and then extracts the important parts of AMR graphs to generate a de-redundant summarization. The experimental results show that the semantic redundancy removal method proposed in this paper has significantly improved the ROUGE and Smatch values compared to other word-aligned eliminate redundant methods. The ROUGE-1 increased by 7.2 percentage points and The Smatch increased by 13.2 percentage points. The novelty and readability of the summarization are greatly improved.