FORGe at SemEval-2017 Task 9: Deep sentence generation based on a sequence of graph transducers
By Simon Mille, Roberto Carlini, Alicia Burga, Leo Wanner. 2017
24 citations recorded by the source corpus.
We present the contribution of Universitat Pompeu Fabra’s NLP group to the SemEval Task 9.2 (AMR-to-English Generation). The proposed generation pipeline comprises: (i) a series of rule-based graph-transducers for the syntacticization of the input graphs and the resolution of morphological agreements, and (ii) an off-the-shelf statistical linearization component.
Subfields: Artificial Intelligence, Natural Language Processing, Natural Language Generation
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
- Claire Gardent
- Laura Perez-Beltrachini
- Yoav Goldberg
- Ido Dagan
- Anastasia Shimorina
- Shashi Narayan
- Mirella Lapata
- Mohit Bansal
In Natural Language Processing
- Claire Gardent
- Ido Dagan
- Laura Perez-Beltrachini
- Yoav Goldberg
- Anastasia Shimorina
- Shashi Narayan
- Christopher D. Manning
- Mirella Lapata
In Natural Language Generation
- Laura Perez-Beltrachini
- Claire Gardent
- Yoav Goldberg
- Anastasia Shimorina
- Amit Moryossef
- Ido Dagan
- Mihir Kale
Citations in the corpus
- The WebNLG Challenge: Generating Text from RDF Data (2017, methodology): “P IPELINE and UIT-VNU-HCM, extracted rules or templates from the training data for surface realisation, whereas the third system, UPF-FORG E , used the FORGe grammar (Mille et al., 2017).”
- Step-by-Step: Separating Planning from Realization in Neural Data-to-Text Generation (2019, background): “…Systems We compare to the best submissions in the WebNLG challenge (Gardent et al., 2017): Melbourne, an end-to-end system that scored best on all categories in the automatic evaluation, and UPF-FORGe (Mille et al., 2017), a classic grammar-based NLG system that scored best in the human evaluation.”