UHD: Cross-Lingual Word Sense Disambiguation Using Multilingual Co-Occurrence Graphs

Proceedings of the 5th International Workshop on Semantic Evaluation, ACL 2010, pages 134–137, Uppsala, Sweden, 15-16 July 2010. c⃝2010 Association for Computational Linguistics UHD: Cross-Lingual Word Sense Disambiguation Using Multilingual Co-occurrence Graphs Carina Silberer and Simone Paolo Ponzetto Department of Computational Linguistics Heidelberg University {silberer,ponzetto}@cl.uni-heidelberg.de Abstract We describe the University of Heidelberg (UHD) system for the Cross-Lingual Word Sense Disambiguation SemEval-2010 task (CL-WSD). The system performs CL- WSD by applying graph algorithms pre- viously developed for monolingual Word Sense Disambiguation to multilingual co- occurrence graphs. UHD has participated in the BEST and out-of-five (OOF) eval- uations and ranked among the most com- petitive systems for this task, thus indicat- ing that graph-based approaches represent a powerful alternative for this task. 1 Introduction This paper describes a graph-based system for Cross-Lingual Word Sense Disambiguation, i.e. the task of disambiguating a word in context by providing its most appropriate translations in dif- ferent languages (Lefever and Hoste, 2010, CL- WSD henceforth). Our goal at SemEval-2010 was to assess whether graph-based approaches, which have been successfully developed for monolingual Word Sense Disambiguation, represent a valid framework for CL-WSD. These typically trans- form a knowledge resource such as WordNet (Fell- baum, 1998) into a graph and apply graph algo- rithms to perform WSD. In our work, we follow this line of research and apply graph-based meth- ods to multilingual co-occurrence graphs which are automatically created from parallel corpora. 2 Related Work Our method is heavily inspired by previous pro- posals from V´eronis (2004, Hyperlex) and Agirre et al. (2006). Hyperlex performs graph-based WSD based on co-occurrence graphs: given a monolingual corpus, for each target word a graph is built where nodes represent content words co- occurring with the target word in context, and edges connect the words which co-occur in these contexts. The second step iteratively selects the node with highest degree in the graph (root hub) and removes it along with its adjacent nodes. Each such selection corresponds to isolating a high- density component of the graph, in order to select a sense of the target word. In the last step the root hubs are linked to the target word and the Mini- mum Spanning Tree (MST) of the graph is com- puted to disambiguate the target word in context. Agirre et al. (2006) compare Hyperlex with an al- ternative method to detect the root hubs based on PageRank (Brin and Page, 1998). PageRank has the advantage of requiring less parameters than Hyperlex, whereas the authors ascertain equal per- formance of the two methods. 3 Graph-based Cross-Lingual WSD We start by building for each target word a mul- tilingual co-occurrence graph based on the target word’s aligned contexts found in parallel corpora (Sections 3.1 and 3.2). Multilingual nodes are linked by translation edges, labeled with the target word’s translations observed in the corresponding contexts. We then use an adapted PageRank al- gorithm to select the nodes which represent the target word’s different senses (Section 3.3) and, given these nodes, we compute the MST, …

Content truncated to 500 words