Network-based distributional semantic models

Iosif, Elias; Ιωσήφ, Ηλίας

doi:10.12681/eadd/29846

Cited by 5 publications

(4 citation statements)

References 164 publications

(240 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Numerous metrics have been proposed for the estimation of semantic similarity between words (a more detailed analysis can be found here [33]). In this work we utilize corpus cooccurrence statistics and, specifically, the Dice coefficient D metric defined as follows…”

Section: Semantic Similarity Metricsmentioning

confidence: 99%

Lexical and affective models in early acquisition of semantics

Kolovou¹,

Iosif²,

Potamianos³

2017

6th Workshop on Child Computer Interaction (WOCCI 2017)

View full text Add to dashboard Cite

Motivated by theories of early language development in children we investigate the contribution of affective features to early acquisition of lexical semantics. For the task of semantic similarity between words, semantic and affective spaces are modeled using network-based distributed semantic models. We propose a method for constructing semantic activations from a combination of lexical and affective relations and show that affective information plays a prominent role in our lexical development model.

show abstract

Section: Semantic Similarity Metricsmentioning

confidence: 99%

Lexical and affective models in early acquisition of semantics

Kolovou¹,

Iosif²,

Potamianos³

2017

6th Workshop on Child Computer Interaction (WOCCI 2017)

View full text Add to dashboard Cite

show abstract

“…The activation layer is motivated by the phenomenon of semantic priming (McNamara, 2005), especially for highly coherent lexical units, such as unigrams and bigrams. In the framework of DSMs, activation layers were computed for the case of unigrams in (Iosif and Potamianos, 2015), and were extended to short phrases (bigrams) in (Iosif, 2013). Consider a phrase, i = (i 1 i 2 ), where i 1 and i 2 denote its first and second constituent.…”

Section: Layer 1: Activation Modelmentioning

confidence: 99%

“…Such efforts proved to be effective when computing the similarity between twoword phrases, however, their limitations were revealed for the case of longer structures (Polajnar et al, 2014), where the composition of meaning becomes more complex. Bengio and Mikolov (2003;2013) proposed an approach based on deep learning for building language models that address the prob-lem of language creativity. The models appear to constantly gain support in comparison with the traditional DSMs.…”

Section: Introductionmentioning

confidence: 99%

Proceedings of the 6th Workshop on Cognitive Modeling and Computational Linguistics

O’Donnell¹,

Schijndel²

2015

View full text Add to dashboard Cite

Syntactic priming from comprehension to production has been shown to be robust: we are more likely to repeat structures that we have previously heard. Many current models do not distinguish between comprehension and production. Here we contrast human language processing with two variants of a Bayesian belief updating model. In the first model, production-to-production priming (i.e. selfpriming) is as strong as comprehension-toproduction priming. In the second, both individuals who self-prime and those who do not are exposed to a syntactic construction via comprehension. Our results suggest that when production-to-production priming is as robust as comprehension-to-production priming, then speakers who self-prime are simultaneously less likely to be primed by input from comprehension and demonstrate different distributions of responses than speakers who do not self-prime. The computational model accords with recent results demonstrating no self-priming, and provides evidence for an account of syntactic priming that distinguishes between production and comprehension input.

show abstract

“…Concept induction is realised by estimating the semantic similarity between the terminal tokens (words) that constitute the corpus vocabulary. Word similarities can be estimated by a variety of similarity metrics [18,19]. In this work, the distributional hypothesis of meaning (i.e., "similarity of context implies similarity of meaning" [20]) is adopted and the semantic similarity between two words is estimated as the Manhattan-norm of their respective bigram probability distributions of left and right contexts [8].…”

Section: The Bottom-up Approachmentioning

confidence: 99%

Fusion of knowledge-based and data-driven approaches to grammar induction

Georgiladakis¹,

Unger²,

Iosif³

et al. 2014

Interspeech 2014

View full text Add to dashboard Cite

Using different sources of information for grammar induction results in grammars that vary in coverage and precision. Fusing such grammars with a strategy that exploits their strengths while minimizing their weaknesses is expected to produce grammars with superior performance. We focus on the fusion of grammars produced using a knowledge-based approach using lexicalized ontologies and a data-driven approach using semantic similarity clustering. We propose various algorithms for finding the mapping between the (non-terminal) rules generated by each grammar induction algorithm, followed by rule fusion. Three fusion approaches are investigated: early, mid and late fusion. Results show that late fusion provides the best relative F-measure performance improvement by 20%.

show abstract

Network-based distributional semantic models

Cited by 5 publications

References 164 publications

Lexical and affective models in early acquisition of semantics

Lexical and affective models in early acquisition of semantics

Proceedings of the 6th Workshop on Cognitive Modeling and Computational Linguistics

Fusion of knowledge-based and data-driven approaches to grammar induction

Contact Info

Product

Resources

About