Measuring Distributional Similarity in Context

摘要: The computation of meaning similarity as operationalized by vector-based models has found widespread use in many tasks ranging from the acquisition synonyms and paraphrases to word sense disambiguation textual entailment. Vector-based are typically directed at representing words isolation thus best suited for measuring out context. In his paper we propose a probabilistic framework Central our approach is intuition that represented probability distribution over set latent senses modulated Experimental results on lexical substitution show algorithm outperforms previously proposed models.