Results
TASK SUMMARY
This page lists submitted results for SemEval-2014 Task 6: Supervised Semantic Parsing of Robotic Spatial Commands. As described in the evaluation section, this task uses data from The Robot Commands Treebank. Percentage accuracy scores are shown below, with and without integrated spatial planning for additional situational context.
SUBMITTED RESULTS
| Team and Affiliation | Method (Statistical or rule-based) | Results (parsing with integrated spatial planning) | Results (parsing without integrated spatial planning) |
| Packard (University of Washington) | Hybrid (rule-based English resource grammar, with statistical Berkeley parser for back-off) | 92.50 | 90.50 |
| Jung and Stoyanchev (AT&T Labs Research) | Statistical (statistical maximum entropy parsing) | 87.35 | 60.84 |
| Evang and Bos (University of Groningen) | Statistical (combinatory categorial grammar with structured perceptron) | 86.80 | 79.21 |
| Ljunglöf (University of Gothenburg) | Rule-based (hand-crafted grammar) | 86.10 | 51.50 |
| Mattelaer, Verbeke and Nitti (Katholieke Universiteit Leuven) | Statistical (combinatory categorial grammar parsing) | 71.29 | 57.76 |
| Kate (University of Wisconsin-Milwaukee) | Statistical (Krisp parser) | - | 45.98 |