Summary of Modeling Unified Semantic Discourse Structure For High-quality Headline Generation, by Minghui Xu et al.
Modeling Unified Semantic Discourse Structure for High-quality Headline Generation
by Minghui Xu, Hao Fei, Fei Li, Shengqiong Wu, Rui Sun, Chong Teng, Donghong Ji
First submitted to arxiv on: 23 Mar 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed unified semantic discourse structure (S3) represents document semantics by combining RST trees with AMR graphs to construct S3 graphs. This hierarchical composition characterizes the semantic meaning of the overall document. A headline generation framework is developed, where S3 graphs are encoded as contextual features. To consolidate the efficacy of S3 graphs, a hierarchical structure pruning mechanism dynamically screens redundant nodes. Experimental results on two datasets demonstrate that this method outperforms existing state-of-art methods consistently. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper proposes a new way to understand and summarize long documents using something called the “unified semantic discourse structure” (S3). Think of it like building a tree-like structure that shows how words, sentences, and ideas are connected. This helps create short and catchy titles that capture the main idea of a document. The researchers develop a special system for generating headlines that uses this S3 structure. They test their method on two datasets and show that it works better than other methods. |
Keywords
» Artificial intelligence » Discourse » Pruning » Semantics
