UWM: Applying an Existing Trainable Semantic Parser to Parse Robotic Spatial Commands

Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014), pages 823–827, Dublin, Ireland, August 23-24, 2014. UWM: Applying an Existing Trainable Semantic Parser to Parse Robotic Spatial Commands Rohit J. Kate University of Wisconsin-Milwaukee Milwaukee, WI katerj@uwm.edu Abstract This paper describes Team UWM’s sys- tem for the Task 6 of SemEval 2014 for doing supervised semantic parsing of robotic spatial commands. An existing semantic parser, KRISP, was trained us- ing the provided training data of natural language robotic spatial commands paired with their meaning representations in the formal robot command language. The en- tire process required very little manual ef- fort. Without using the additional annota- tions of word-aligned semantic trees, the trained parser was able to exactly parse new commands into their meaning repre- sentations with 51.18% best F-measure at 72.67% precision and 39.49% recall. Re- sults show that the parser was particularly accurate for short sentences. 1 Introduction Semantic parsing is the task of converting natu- ral language utterances into their complete formal meaning representations which are executable for some application. Example applications of seman- tic parsing include giving natural language com- mands to robots and querying databases in natu- ral language. Some old semantic parsers were de- veloped manually to work for specific applications (Woods, 1977; Warren and Pereira, 1982). How- ever, such semantic parsers were generally brittle and building them required a lot of manual effort. In addition, these parsers could not be ported to any other application without again putting signif- icant manual effort. More recently, several semantic parsers have been developed using machine learning (Zelle and This work is licensed under a Creative Commons At- tribution 4.0 International Licence. Page numbers and pro- ceedings footer are added by the organisers. Licence details: http://creativecommons.org/licenses/by/4.0/ Mooney, 1996; Ge and Mooney, 2005; Zettle- moyer and Collins, 2005; Wong and Mooney, 2006; Kate and Mooney, 2006; Lu et al., 2008; Kwiatkowski et al., 2011). In this approach, train- ing data is first created for the domain of inter- est. Then using one of the many machine learn- ing methods and semantic parsing frameworks, a semantic parser is automatically learned from the training data (Mooney, 2007). The trained seman- tic parser is then capable of parsing new natu- ral language utterances into their meaning repre- sentations. Semantic parsers built using machine learning tend to be more robust and can be easily ported to other application domains with appropri- ate domain-specific training data. The Task 6 of SemEval 2014 provided a new ap- plication domain for semantic parsing along with training and test data. The domain involved giv- ing natural language commands to a robotic arm which would then move blocks on a board (Dukes, 2013). The domain was inspired from the classic AI system SHRDLU (Winograd, 1972). The train- ing data contained 2500 examples of sentences paired with their meaning representations in the Robot Command Language (RCL) which was de- signed for this domain (Dukes, 2013). The test data contained 909 such example pairs. We trained an …

Content truncated to 500 words