Unsupervised Morphology Induction Using Word Embeddings

By Radu Soricut, Franz Joseph Och. 2015

153 citations recorded by the source corpus.

We present a language agnostic, unsupervised method for inducing morphological transformations between words. The method relies on certain regularities manifest in highdimensional vector spaces. We show that this method is capable of discovering a wide range of morphological rules, which in turn are used to build morphological analyzers. We evaluate this method across six different languages and nine datasets, and show significant improvements across all languages.

Subfields: Artificial Intelligence, Natural Language Processing, Computational Linguistics & Semantics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Computational Linguistics & Semantics

Citations in the corpus

Finding Function in Form: Compositional Character Models for Open Vocabulary Word Representation (2015, methodology): “This idea has been explored in prior work by composing morphemes into representations of words (Luong et al., 2013; Botha and Blunsom, 2014; Soricut and Och, 2015).”
Charagram: Embedding Words and Sentences via Character n-grams (2016, neutral): “We obtained a Spearman’s ρ of 47.1, which outperforms the 41.8 result from Soricut and Och (2015) and is competitive with the 47.8 reported in Pennington et al. (2014), despite only using PPDB for training.”
Yandex School of Data Analysis approach to English-Turkish translation at WMT16 News Translation Task (2016, methodology): “As an alternative, we trained our implementation of unsupervised morphology model, following (Soricut and Och, 2015), with a single distinctive feature: in each connected component C of the morphological graph, we select the lemma as argmax C (log f ( w ) − α · l ( w )) , where l ( w ) is word…”
Analogy-based detection of morphological and semantic relations with word embeddings: what works and what doesn’t. (2016, methodology): “Analogies have been successfully used for detecting different semantic relations, such as synonymy and antonymy (Turney, 2008), ConceptNet relations and selectional preferences (Herdadelen and Baroni, 2009), and also for inducing morphological categories from unparsed data (Soricut and Och, 2015).”
Analogical Reasoning on Chinese Morphological and Semantic Relations (2018, background): “In addition, It can be used in inducing morphological transformations (Soricut and Och, 2015), detecting semantic relations (Herdagdelen and Baroni, 2009), and translating unknown words (Langlais and Patry, 2007).”