PURSUIT Corpus
The PURSUIT Corpus was collected at the Institute for Human and
Machine Cognition in January and February 2008. It consists of 13 audio and GPS recordings
of realtime route descriptions as they were driven in cars in the downtown Pensacola, Florida
area. The audio signals have been transcribed and separated into utterances. Additionally,
references to geospatial entities has been annotated in the corpus with the name,
address, and latitude and longitude information for each referent.
This corpus was created as part of a project for automated understanding of natural language
path descriptions. The GPS recording of the actual path taken, along with the annotated
geospatial references serve as a ``ground truth'' for understanding the path description.
It may be freely downloaded and used, although we ask that you let us know if you
are using it so we can more persuasively convince our sponsors that it is useful.
Also, please cite the following paper when referencing the corpus:
- Nate Blaylock. Semantic Annotation of Street-level Geospatial Entities. In Proceedings of the IEEE ICSC Workshop on Semantic Annotation for Computational Linguistic Resources, Palo Alto, California, September 19-21, 2011.
[ .pdf ]
Documentation
Download
Publications
- Nate Blaylock. Semantic Annotation of Street-level Geospatial Entities. In Proceedings of the IEEE ICSC Workshop on Semantic Annotation for Computational Linguistic Resources, Palo Alto, California, September 19-21, 2011.
[ .pdf ]
- Nate Blaylock, Bradley Swain, and James Allen.
Mining Geospatial Path Data from Natural Language Descriptions
In Proceedings of the 1st ACM SIGSPATIAL GIS International Workshop on Querying and Mining Uncertain Spatio-Temporal Data, Seattle, Washington. November 3, 2009.
[ .pdf ]
- Nate Blaylock, Bradley Swain, and James Allen.
TESLA: A Tool for Annotating Geospatial Language Corpora.
In Proceedings North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL HLT) 2009 Conference, Boulder, Colorado. May 31-June 5 2009.
[ .pdf ]
- Nate Blaylock and James Allen.
Real-time Path Descriptions Grounded with GPS Tracks: a preliminary
report. In Proceedings of the LREC Workshop on Methodologies and
Resources for Processing Spatial Language, Marrakech, Morocco. May
31 2008.
[ .pdf ]
Contact