Unsupervised Compound Splitting With Distributional Semantics Rivals Supervised Methods

Proceedings of NAACL-HLT 2016, pages 617–622, San Diego, California, June 12-17, 2016. c⃝2016 Association for Computational Linguistics Unsupervised Compound Splitting With Distributional Semantics Rivals Supervised Methods Martin Riedl and Chris Biemann Language Technology Computer Science Department, Technische Universit¨at Darmstadt Hochschulstrasse 10, D-64289 Darmstadt, Germany {riedl,biem}@cs.tu-darmstadt.de Abstract In this paper we present a word decom- pounding method that is based on distribu- tional semantics. Our method does not re- quire any linguistic knowledge and is initial- ized using a large monolingual corpus. The core idea of our approach is that parts of com- pounds (like “candle” and “stick”) are seman- tically similar to the entire compound, which helps to exclude spurious splits (like “candles” and “tick”). We report results for German and Dutch: For German, our unsupervised method comes on par with the performance of a rule-based and a supervised method and sig- nificantly outperforms two unsupervised base- lines. For Dutch, our method performs only slightly below a rule-based optimized com- pound splitter. 1 Introduction Germanic and agglutinative languages (e.g. German, Swedish, Finnish, Korean) have a productive mor- phology that allows the formation of not space- separated compounds in a much larger extent than e.g. in English. The task of separating such com- pounds into their corresponding single word (sub-) units is called compound splitting or decompound- ing. Decompounding showed impact in several NLP applications, e.g. ASR (Adda-Decker and Adda, 2000), MT (Koehn and Knight, 2003) or IR (Monz and de Rijke, 2001), and is generally perceived as a crucial component for the processing of respec- tive languages. However, most existing systems rely on dictionaries or are trained in a supervised fash- ion. Both approaches require substantial manual work and do not adapt to vocabulary change. In this paper we introduce an unsupervised method for de- compounding that relies on distributional semantics. For the computation of the semantic model we solely rely on a tokenized monolingual corpus and do not require any further linguistic processing. Most pre- vious research on compound splitting concentrates on the detection of lemmas that form the compound. Whereas this is important for several tasks, in this work we focus on the splitting of a compound into its word units without any base form reduction, ar- guing that lemmatization is either part of the task pipeline anyways (e.g. IR) or not required (e.g. for ASR). 2 Related Work Approaches to automatic decompounding can be classified into corpus-driven approaches and super- vised approaches. Corpus-driven approaches are usually informed by a frequency list (Koehn and Knight, 2003), by a probabilistic model (Schiller, 2005), by parallel corpora (Koehn and Knight, 2003; Macherey et al., 2011) or by the existence of pe- riphrases (i.e. reformulations) in large monolingual corpora (Holz and Biemann, 2008). As with other tasks, supervised approaches are usually superior to unsupervised approaches if sufficient training ma- terial is provided. A straightforward yet effective supervised decompounding system is contained in the ASV Toolbox (Biemann et al., 2008), which uses trie-based datastructures for recursively split- ting compounds based on learned …

Content truncated to 500 words