Toward Abstractive Summarization Using Semantic Representations

By Fei Liu, Jeffrey Flanigan, Sam Thomson, Norman M. Sadeh, Noah A. Smith. 2018

316 citations recorded by the source corpus.

We present a novel abstractive summarization framework that draws on the recent development of a treebank for the Abstract Meaning Representation (AMR). In this framework, the source text is parsed to a set of AMR graphs, the graphs are transformed into a summary graph, and then text is generated from the summary graph. We focus on the graph-tograph transformation that reduces the source semantic graph into a summary graph, making use of an existing AMR parser and assuming the eventual availability of an AMR-totext generator. The framework is data-driven, trainable, and not specifically designed for a particular domain. Experiments on goldstandard AMR annotations and system parses show promising results. Code is available at: https://github.com/summarization

Subfields: Artificial Intelligence, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

Citations in the corpus

Parsing Algebraic Word Problems into Equations (2015, methodology): “…discriminative models trained from a small number of word problems and their solutions – without any manual annotation; and (3) We demonstrate empirically that ALGES has broader scope than the system of Hosseini et al. (2014), and overcomes the brittleness of the method of Kushman et al. (2014).”
Multi-Reward Reinforced Summarization with Saliency and Entailment (2018, methodology): “Recent focus has shifted towards abstractive, rewriting-based summarization based on parse trees (Cheung and Penn, 2014; Wang et al., 2016), Abstract Meaning Represen-tations (Liu et al., 2015; Dohare and Karnick, 2017), and neural network models with pointer-copy mechanism and coverage (Rush et al.”
Soft Layer-Specific Multi-Task Summarization with Entailment and Question Generation (2018, background): “…on graphs and concept maps (Giannakopoulos, 2009; Ganesan et al., 2010; Falke and Gurevych, 2017) and discourse trees (Gerani et al., 2014), syntactic parse trees (Cheung and Penn, 2014; Wang et al., 2013), and Abstract Meaning Repre-sentations (AMR) (Liu et al., 2015; Dohare and Karnick, 2017).”
Concept Pointer Network for Abstractive Summarization (2019, methodology): “Abstract Meaning Representation (AMR) is used to transform a sentence into a concept graph, then merge those similar concept nodes to form a new summary graph (Liu et al., 2018).”
WikiGraphs: A Wikipedia Text - Knowledge Graph Paired Dataset (2021, background): “Example applications include generating text summaries conditioned on Abstract Meaning Representation graphs (Liu et al., 2018), generating the abstract of a scientific article given a KG and title (Koncel-Kedziorski et al., 2019) and generating text from RDF triples (Gardent et al., 2017; Jin et…”
Controlling the Amount of Verbatim Copying in Abstractive Summarization (2019, unknown): “Abstractive summarization distills the source text into its essential meanings, then performs language generation from the representation to produce an abstract (Barzilay and McKeown 2005; Liu et al. 2015; Liao, Lebanoff, and Liu 2018; Hardy and Vlachos 2018).”
Recent advances in document summarization (2017, background): “One may build sentences via merging consistent noun phrases and verb phrases [13] or linearizing graph-based semantic units derived from semantic formalisms such as abstract meaning representation (AMR) [121].”
Abstractive Dialogue Summarization with Sentence-Gated Modeling Optimized by Dialogue Acts (2018, methodology): “Extractive methods assemble the summary from the source text directly [1, 2, 3], while abstractive methods generate words to form the summary [4, 5, 6].”
Broad-coverage CCG Semantic Parsing with AMR (2015, background): “Recently, AMR (Banarescu et al., 2013) was proposed as a general-purpose meaning representation language for broad-coverage text, and work is ongoing to study its use for variety of applications such as machine translation (Jones et al., 2012) and summarization (Liu et al., 2015).”
Joint Parsing and Generation for Abstractive Summarization (2019, background): “…Berg-Kirkpatrick, and Klein 2016), rhetorical structure (Christensen et al. 2013; Yoshida et al. 2014; Li, Thadani, and Stent 2016), and abstract meaning representation (Liu et al. 2015; Liao, Lebanoff, and Liu 2018; Hardy and Vlachos 2018) to generate compressive and abstractive summaries.”
Using Local Knowledge Graph Construction to Scale Seq2Seq Models to Multi-Document Inputs (2019, methodology): “On WikiSum, we report only ROUGE-L following (Liu et al., 2018b).”
Structural Neural Encoders for AMR-to-text Generation (2019, background): “3650 2017) or tree-structured (Liu et al., 2015; Takase et al., 2016) data, while other work maintained them but did not analyze their impact on performance (e.”
Structured Neural Summarization (2018, background): “Closer to our work is the work of Liu et al. (2018), who use abstract meaning representation (AMR), in which the source document is ﬁrst parsed into AMR graphs, before a summary graph is created, which is ﬁnally rendered in natural language.”
Neural AMR: Sequence-to-Sequence Models for Parsing and Generation (2017, methodology): “, 2012), summarization (Liu et al., 2015), sentence compression (Takase et al.”
Transforming Dependency Structures to Logical Forms for Semantic Parsing (2016, background): “…to knowledge bases (Kwiatkowski et al., 2013; Reddy et al., 2014; Choi et al., 2015; Krishna-murthy and Mitchell, 2015), sentence simpliﬁcation (Narayan and Gardent, 2014), summarization (Liu et al., 2015), paraphrasing (Pavlick et al., 2015) and relation extraction (Rockt¨aschel et al., 2015).”
The Role of Discourse Units in Near-Extractive Summarization (2016, background): “…when the task at hand is more abstractive, the units are more ﬁne-grained, e.g., n-grams and phrases in abstractive summarization (Kikuchi et al., 2014; Liu et al., 2015; Bing et al., 2015), n-grams and human-annotated concept units in summarization evaluation (Lin, 2004; Hovy et al., 2006).”
Trends in Integration of Vision and Language Research: A Survey of Tasks, Datasets, and Methods (2019, background): “It can be seen that the task closely aligned to text summarization (Nallapati et al., 2016; Liu et al., 2018) mostly generating abstractive summaries.”
AMR-To-Text Generation with Graph Transformer (2020, background): “, 2019), question answering (Mitra and Baral, 2016), summarization (Liu et al., 2015; Takase et al., 2016), and event extraction (Huang et al.”
MeanSum: A Neural Model for Unsupervised Multi-Document Abstractive Summarization (2018, methodology): “Liu et al. (2015) present a framework for doing abstractive summarization in three stages.”