We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:


References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: An analysis of document graph construction methods for AMR summarization

Abstract: Meaning Representation (AMR) is a graph-based semantic representation for sentences, composed of collections of concepts linked by semantic relations. AMR-based approaches have found success in a variety of applications, but a challenge to using it in tasks that require document-level context is that it only represents individual sentences. Prior work in AMR-based summarization has automatically merged the individual sentence graphs into a document graph, but the method of merging and its effects on summary content selection have not been independently evaluated. In this paper, we present a novel dataset consisting of human-annotated alignments between the nodes of paired documents and summaries which may be used to evaluate (1) merge strategies; and (2) the performance of content selection methods over nodes of a merged or unmerged AMR graph. We apply these two forms of evaluation to prior work as well as a new method for node merging and show that our new method has significantly better performance than prior work.
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2111.13993 [cs.CL]
  (or arXiv:2111.13993v1 [cs.CL] for this version)

Submission history

From: Fei-Tzin Lee [view email]
[v1] Sat, 27 Nov 2021 22:12:50 GMT (11070kb,D)

Link back to: arXiv, form interface, contact.