RI: Small: RUI: AIR: Automatic Idiom Recognition

本研究项目的主要目标是开发一种与语言无关的习语自动识别方法。日常语言中有很多习语，比如“变相祝福”和“踢桶”，但它们仍然很神秘，因为人们到底是如何学习和理解这些习语的，目前还不清楚。成语没有一个统一的定义，涵盖了这一类的所有成员，但成语在语法形式和意义上往往相对固定，但在形式和意义之间的关系上相对较少可预测。此外，许多习语表达形式可以同时带有字面意思，也就是完全可以预测的解释--比方说，小女孩对着妈妈做鬼脸。(习语)vs.“小女孩用一个胡萝卜和两个按钮在雪人身上做了个鬼脸。”因此，习语对各种自然语言处理应用程序提出了巨大的挑战，包括机器翻译系统，它通常不检测习语语言。为了应对这些挑战，提出了一种既不依赖于目标成语类型、词典或大型人工标注语料库，也不通过特定类型的语言结构来限制搜索空间的算法。研究的出发点是，习语是违反衔接结构的语义离群点，尤其是在局部语境中。以下属性被量化并被结合到离群点检测算法中：1)与字面表达或其他类型的搭配相比缺乏组合性；2)破坏局部衔接关系，使得它们往往在语义上远离本地主题；3)虽然不是所有的语义离群点都是习语，但非构成性语义离群点可能是习语；4)成语不是二元属性；这项研究有助于更好地理解习语语言，有助于对此类现象的计算机处理，并通过创建高质量的、公开可用的习语注释语言资源，促进机器学习研究和大数据科学的发展。其他好处包括从大型语料库计算合成性和主题性的高效算法，关于比喻语言性质的有趣的新概括，以及培训一批本科生和研究生在一个困难的跨学科问题上进行高度实用的工作。

The main goal of this research project is to develop a language independent method for automatic idiom recognition. Idiomatic expressions, such as 'a blessing in disguise' and 'kick the bucket' are plentiful in everyday language, though they remain mysterious, as it is not clear exactly how people learn and understand them. There is no single agreed-upon definition of idiom that covers all members of this class, but idioms tend to be relatively fixed in grammatical form and meaning, but with relatively little predictability in the relation between form and meaning. Also, many idiomatic expressions can appear with both literal, i.e. fully predictable, interpretations given their form -- compare 'The little girl made a face at her mother.' (idiomatic) vs. 'The little girl made a face on the snowman using a carrot and two buttons.' (literal) As a result, idioms present great challenges for a variety of natural language processing applications, including machine translation systems, which often do not detect idiomatic language. To address these challenges, an algorithm is proposed that neither relies on target idiom types, lexicons, or large manually annotated corpora, nor limits the search space by a particular type of linguistic construction. The starting point is that idioms are semantic outliers that violate cohesive structure, especially in local contexts. The following properties are quantified and are incorporated into the outlier detection algorithm: 1) lack of compositionality comparing to literal expressions or other types of collocations; 2) violation of local cohesive ties, so that they tend to be semantically distant from the local topics; 3) while not all semantic outliers are idioms, non-compositional semantic outliers are likely to be idiomatic; 4) idiomaticity is not a binary property; rather, idioms fall on the continuum from being compositional to being partly unanalyzable to completely non-compositional.This research contributes to the better understanding of idiomatic language, to the computational treatment of such phenomena and, with the creation of high quality, publicly available linguistic resources annotated for idioms, to the facilitation of machine learning research and big data science. Additional benefits include efficient algorithms for computing compositionality and topicality from large corpora, interesting new generalizations about the nature of figurative language, and the training of a cadre of undergraduate and graduate students in highly practical work on a difficult interdisciplinary problem.