Abstract | ||
---|---|---|
This article reports on the novel task of spatial role labeling in natural language text. It proposes machine learning methods to extract spatial roles and their relations. This work experiments with both a step-wise approach, where spatial prepositions are found and the related trajectors, and landmarks are then extracted, and a joint learning approach, where a spatial relation and its composing indicator, trajector, and landmark are classified collectively. Context-dependent learning techniques, such as a skip-chain conditional random field, yield good results on the GUM-evaluation (Maptask) data and CLEF-IAPR TC-12 Image Benchmark. An extensive error analysis, including feature assessment, and a cross-domain evaluation pinpoint the main bottlenecks and avenues for future research. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1145/2050104.2050105 | TSLP |
Keywords | Field | DocType |
article report,spatial role,composing indicator,step-wise approach,joint learning approach,towards extraction,spatial relation,spatial preposition,natural language,clef-iapr tc-12 image,cross-domain evaluation,extensive error analysis,machine learning,conditional random field,context dependent,spatial information,spatial relations | Spatial relation,Conditional random field,Computer science,Semantic labeling,Information extraction,Natural language,Natural language processing,Artificial intelligence,Landmark,Machine learning | Journal |
Volume | Issue | ISSN |
8 | 3 | 1550-4875 |
Citations | PageRank | References |
41 | 1.52 | 33 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Parisa Kordjamshidi | 1 | 143 | 18.52 |
Martijn Van Otterlo | 2 | 174 | 9.31 |
Marie-Francine Moens | 3 | 1750 | 139.27 |