Paper: Semantic Discourse Segmentation And Labeling For Route Instructions

ACL ID P06-3006
Title Semantic Discourse Segmentation And Labeling For Route Instructions
Venue Annual Meeting of the Association of Computational Linguistics
Session Student Session
Year 2006
Authors

In order to build a simulated robot that accepts instructions in unconstrained nat- ural language, a corpus of 427 route in- structions was collected from human sub- jects in the office navigation domain. The instructions were segmented by the steps in the actual route and labeled with the action taken in each step. This flat formulation reduced the problem to an IE/Segmentation task, to which we applied Conditional Random Fields. We com- pared the performance of CRFs with a set of hand-written rules. The result showed that CRFs perform better with a 73.7% success rate.