Resumen |
In this paper we describe the JUNITMZ 1 system that was developed for participation in Se-mEval 2016 Task 1: Semantic Textual Similarity. Methods for measuring the textual similarity are useful to a broad range of applications including: text mining, information retrieval, dialogue systems, machine translation and text summarization. However, many systems developed specifically for STS are complex, making them hard to incorporate as a module within a larger applied system. In this paper, we present an STS system based on three simple and robust similarity features that can be easily incorporated into more complex applied systems. The shared task results show that on most of the shared tasks evaluation sets, these signals achieve a strong (>0.70) level of correlation with human judgements. Our system's three features are: unigram overlap count, length normalized edit distance and the score computed by the METEOR machine translation metric. Features are combined to produces a similarity prediction using both a feedforward and recurrent neural network. © 2016 Association for Computational Linguistics. |