Model Comparison for Semantic Grouping
This work addresses the need for better semantic similarity measures in natural language processing, but it is incremental as it builds on existing model comparison techniques.
The authors tackled the problem of quantifying semantic similarity between groups of embeddings by introducing a probabilistic framework that treats it as a model comparison task, achieving competitive results on Semantic Textual Similarity datasets.
We introduce a probabilistic framework for quantifying the semantic similarity between two groups of embeddings. We formulate the task of semantic similarity as a model comparison task in which we contrast a generative model which jointly models two sentences versus one that does not. We illustrate how this framework can be used for the Semantic Textual Similarity tasks using clear assumptions about how the embeddings of words are generated. We apply model comparison that utilises information criteria to address some of the shortcomings of Bayesian model comparison, whilst still penalising model complexity. We achieve competitive results by applying the proposed framework with an appropriate choice of likelihood on the STS datasets.