AIC Seminar Series

One of the key challenges for creating a semantic representation of a text is mapping words found in a natural language text to their meanings. This task, Word Sense Disambigutation (WSD), is confounded by the fact that words have multiple meanings, or senses, dictated by their use in a sentence and the domain. We present a method that employs random walks over the graph structure of knowledge bases giving state-of-the-art results for WSD on both general and biomedical texts. This method can also be used for computing word similarity, and we show preliminary results that show improvements on an information retrieval task.