From distributional to semantic similarity

By James Richard Curran. 2004

289 citations recorded by the source corpus.

How researchers see this paper

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

In Natural Language Processing

Citations in the corpus

Zero-Shot Learning Through Cross-Modal Transfer (2013, background): “These representations have proven very effective in natural language processing tasks such as sense disambiguation [30], thesaurus extraction [24, 8] and cognitive modeling [22].”
Experimental Support for a Categorical Compositional Distributional Model of Meaning (2011, background): “The distributional models have found their way into real world applications such as thesaurus extraction (Grefenstette, 1994; Curran, 2004) or automated essay marking (Landauer, 1997), and have connections to semantically motivated information retrieval (Manning et al.”
Mathematical Foundations for a Compositional Distributional Model of Meaning (2010, background): “For example, the top 10 most similar nouns to introduction, according to the system of [11], are launch, implementation, advent, addition, adoption, arrival, absence, inclusion, creation.”
Evaluating Topic Coherence Using Distributional Semantics (2013, methodology): “Then the similarity between the words, and therefore cohesion of the topic, can be computed using the following measures (Curran, 2003; Grefenstette, 1994): • The cosine of the angles between the vectors: • The Dice coefﬁcient: • The Jaccard coefﬁcient: Each of these measures estimates the distance…”
Vector Space Models of Lexical Meaning (2015, methodology): “Curran (2004) investigates a range of context definitions for the task of automatic thesaurus extraction, focusing on the synonymy relation.”
Counter-fitting Word Vectors to Linguistic Constraints (2016, background): “This hypothesis supports unsupervised learning of meaningful word representations from large corpora (Curran, 2003; Ó Séaghdha and Korhonen, 2014; Mikolov et al., 2013; Pennington et al., 2014).”
Text: now in 2D! A framework for lexical expansion with contextual similarity (2013, methodology): “The representation as shown here is more general than representations used by e.g. Lin (1998) and Curran (2004): whereas these previous works only allow a single term to be characterized with features , we allow arbitrary splits over arbitrarily complex observations, as shown in example c).”
Distributional Models of Word Meaning (2018, background): “The most common approach to compositionality in DS assumes that the distributional representation of a complex expression is a vector, and uses linear-algebraic operations to project lexical vectors to phrase vectors. The simplest form of vector combination is vector addition (Landauer & Dumais 1997”
“Not not bad” is not “bad”: A distributional account of negation (2013, methodology): “These models, mathematically instantiated as sets of vectors in high dimensional vector spaces, have been applied to tasks such as thesaurus extraction (Grefenstette, 1994; Curran, 2004), word-sense discrimination (Schütze, 1998), automated essay marking (Landauer and Dumais, 1997), and so on.”
Category-theoretic quantitative compositional distributional models of natural language semantics (2013, methodology): “Distributional models have found their way into real world applications such as thesaurus extraction [41, 21] or automated essay marking [55], and have strong connections to semantically motivated information retrieval [59].”
Distributional techniques for philosophical enquiry (2012, background): “, vectors in a multidimensional space where each dimension corresponds to a potential context for a lexical item (Curran, 2003).”
Dependency-Based Construction of Semantic Space Models (2007, background): “Although replicating our study with Briscoe and Carroll’s parser (2002) is outside of the scope of this article, we should note that the two parsers yield comparable performances and employ a similar inventory of dependency relations (see Curran (2004) for more discussion).”
Prior Disambiguation of Word Tensors for Constructing Sentence Vectors (2013, background): “Distributional models of meaning have been proved extremely useful for a number of natural language processing tasks, ranging from thesaurus extraction (Curran, 2004) to topic modelling (Landauer and Dumais, 1997) and information retrieval (Manning et al., 2008), to name just a few.”
Composition in Distributional Models of Semantics (2010, background): “While considerable research has been directed at optimising the construction of these word level representations (e.g., Bullinaria and Levy, 2007; Weeds, 2003; Curran, 2003), less attention has been focused on the question of how to combine them.”
Learning to Follow Navigational Directions (2010, background): “Traditional accounts of learning typically rely on linguistic annotation (Zettlemoyer and Collins, 2009) or word distributions (Curran, 2003).”
Combining Symbolic and Distributional Models of Meaning (2007, methodology): “However, the recent improvements in wide-coverage parsing (Clark & Curran 2004) have led to attempts to scale up symbolic approaches, with some success (Bos 2005).”
Compositional Operators in Distributional Semantics (2014, background): “…space models in order to model and reason about human learning rates in language; Schütze (1998) performs word sense induction and disambiguation; Curran (2004) shows how distributional models can be applied to automatic thesaurus extraction; Manning et al. (2008) discuss possible applications…”
Semantic Compositionality through Recursive Matrix-Vector Spaces (2012, background): “These representations have proven very effective in sense discrimination and disambiguation (Sch¨utze, 1998), automatic thesaurus extraction (Lin, 1998; Curran, 2004) and selectional preferences.”
An Exploration of Discourse-Based Sentence Spaces for Compositional Distributional Semantics (2015, methodology): “Following the procedure outlined in Polajnar and Clark (2014), we employed t-test weighting (Curran, 2004) and context selection, and reduced our noun vectors ( n ) to K = 100 dimensions using Singular Value Decomposition (SVD).”
Multi-Prototype Vector-Space Models of Word Meaning (2010, methodology): “, 2009), and (2) using the distributional hypothesis to automatically infer a vector-space prototype of word meaning from large corpora (Agirre et al., 2009; Curran, 2004; Harris, 1954).”