SemEval-2017 Task 2: Multilingual and Cross-lingual Semantic Word Similarity.

This paper introduces a new task on Multilingualand Cross-lingual Semantic WordSimilarity which measures the semanticsimilarity of word pairs within and acrossfive languages: English, Farsi, German,Italian and Spanish. High quality datasetswere manually curated for the five languageswith high inter-annotator agreements(consistently in the 0.9 ballpark).These were used for semi-automatic constructionof ten cross-lingual datasets. 17teams participated in the task, submitting24 systems in subtask 1 and 14 systems insubtask 2. Results show that systems thatcombine statistical knowledge from textcorpora, in the form of word embeddings,and external knowledge from lexical resourcesare best performers in both subtasks.More information can be found onthe task website: http://alt.qcri.org/semeval2017/task2/ .