ECNU at SemEval-2018 Task 10: Evaluating Simple but Effective Features on Machine Learning Methods for Semantic Difference Detection

Proceedings of the 12th International Workshop on Semantic Evaluation (SemEval-2018), pages 999–1002 New Orleans, Louisiana, June 5–6, 2018. ©2018 Association for Computational Linguistics ECNU at SemEval-2018 Task 10: Evaluating Simple but Effective Features on Machine Learning Methods for Semantic Difference Detection Yunxiao Zhou1, Man Lan1,2∗ ,Yuanbin Wu1,2 1Department of Computer Science and Technology, East China Normal University, Shanghai, P.R.China 2Shanghai Key Laboratory of Multidimensional Information Processing 51164500061@stu.ecnu.edu.cn, {mlan, ybwu}@cs.ecnu.edu.cn Abstract This paper describes the system we submit- ted to Task 10 (Capturing Discriminative At- tributes) in SemEval 2018. Given a triple (word1, word2, attribute), this task is to predic- t whether it exemplifies a semantic difference or not. We design and investigate several word embedding features, PMI features and Word- Net features together with supervised machine learning methods to address this task. Official- ly released results show that our system ranks above average. 1 Introduction The Capturing Discriminative Attributes task (Pa- perno et al., 2018) in SemEval 2018 is to provide a standard testbed for semantic difference detec- tion, which will benefit many other application- s in Natural Language Processing (NLP), such as automatized lexicography and machine translation (Krebs and Paperno, 2016). The goal of this task is to predict whether a word is a discriminative at- tribute between two concepts. Specifically, given two concepts and an attribute, the task is to predict whether the first concept has this attribute but the second concept does not. For example, given the concepts apple and pineapple, participants are re- quired to predict whether the attribute seeds char- acterizes the first concept but not the other. In oth- er words, semantic difference detection is a binary classification task: given a triple (apple, pineap- ple, seeds), the task is to determine whether it ex- emplifies a semantic difference or not, i.e., positive or negative. Table 1 shows more data examples. If word1 has a specific attribute but word2 does not, then the correlation of attribute and word1 should be higher than that of attribute and word2. The semantic similarity is in the same way. In view of the above considerations, to address this task, we explore supervised machine learn- word1 word2 attribute label apple pineapple seeds positive candle chandelier melts positive apple coconut brine negative apple cucumber seeds negative Table 1: Examples from the training data. ing methods which use PMI features and Word- Net features. In recent years, more and more s- tudies have focused on word embeddings as an al- ternative to traditional hand-crafted features (Pen- nington et al., 2014; Tang et al., 2014). Therefore we use word embeddings to obtain the semantic similarity as word embedding features. Besides, we perform a series of experiments to explore the effectiveness of feature types and supervised ma- chine learning algorithms. 2 System Description To perform semantic difference detection of given triples, we adopt supervised learning algorithms with several features that represent semantic sim- ilarity and correlation. In the next paragraphs, we will introduce feature engineering and learning al- gorithms. 2.1 Feature Engineering In this …

Content truncated to 500 words