UMR Parsing Workshop
June 14, 2024
University of Colorado Boulder
Key Dates
- April 8, 2024: Submission deadline
- April 25, 2024: Notification of acceptance
- May 30, 2024: Camera-ready versions due
This workshop will focus on developing parsers for Uniform Meaning Representations. The goal is to start from raw text from real-world settings that could be in any one of many typologically different languages, even low-resource languages for which there is little or no training data. This can be achieved by exploiting a common semantic annotation standard.
This workshop has been made possible by funding for NSF Collaborative Research: Building a Broad Infrastructure for Uniform Meaning Representations (Award # 2213805), which is aimed at developing guidelines and annotation for cross-lingual Uniform Meaning Representations, based on the original Abstract Meaning Representation guidelines for English, but ensuring cross-linguistically consistent annotation and recoverability of the original raw texts. This workshop will overlap with the last day of the Colorado UMR Annotation Summer School.
The workshop is open to everyone and will cover the fundamentals of UMR annotation and the differences between AMR and UMR. In addition to the talk from our invited speaker, there will be presentations on recent successful approaches to AMR parsing and how they can be applied to UMR parsing. We welcome submissions from anyone on related topics, such as:
- AMR or UMR parsing for any language
- AMR or UMR generation for any language
- Evaluation metrics for AMR or UMR parsing
- Bootstrapping of AMRs or UMRs from related semantic representations such as Propbanks
- Projections of English AMR onto other languages
- Challenges of applying AMR annotation to languages other than English
- Challenges of accurate multi-sentence coreference as a subtask of AMR parsing
- Any other topic related to the parsing and generation of AMRs or UMRs
Submissions
Submissions should report original and unpublished research on topics of interest to the workshop. Accepted papers are expected to be presented at the workshop and will be published in the workshop proceedings. They should emphasize obtained results rather than intended work and should clearly indicate the state of completion of the reported results.
Submission is electronic, using the Workshop submission site in Easy Chair. Submissions must adhere to the two-column format of ACL venues, using the Overleaf template taken from ACL 2023. Initial submissions should be fully anonymous to ensure double-blind reviewing. Long papers must not exceed eight (8) pages of content; short papers must not exceed four (4) pages of content. References and appendices do not count against these limits.
To ensure double-blind reviews, papers must not include the authors' names and affiliations or self- references that reveal any author's identity. Papers that do not conform to these requirements will be rejected without review.